Re: [blfs-dev] icedtea/OpenJDK versions
On 04/13/14 05:17, Pierre Labastie wrote: Hi, While icedtea has a new version, OpenJDK is still at 1.7.0_51. My concern is that I built new binary JDK's, but that in our naming scheme, they should have the same version number as the previous ones. How should I name them? - OpenJDK-1.7.0.51-{i686,x86_64}-bin-2 - OpenJDK-1.7.0.51-2-{i686,x86_64}-bin - upload them with the same name as the prebious one? (not good, since the checksum changes) - not upload them at all Regards Pierre Hmm? 2.4.7 is 1.7 u55b14: changeset 1cffa74f89e7 in /hg/release/icedtea7-2.4 details:http://icedtea.classpath.org/hg/release/icedtea7-2.4?cmd=changeset;node=1cffa74f89e7 author: Andrew John Hughesgnu_and...@member.fsf.org date: Wed Apr 16 05:19:28 2014 +0100 Set to u55b14. 2014-04-15 Andrew John Hughesgnu_and...@member.fsf.org * Makefile.am: (JDK_UPDATE_VERSION): Bump to 55 to match upstream. (BUILD_VERSION): Bump to b14 to match upstream. Something odd with your packaging? --DJ -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] icedtea/OpenJDK versions
On 04/17/14 19:11, DJ Lucas wrote: On 04/13/14 05:17, Pierre Labastie wrote: Hi, While icedtea has a new version, OpenJDK is still at 1.7.0_51. My concern is that I built new binary JDK's, but that in our naming scheme, they should have the same version number as the previous ones. How should I name them? - OpenJDK-1.7.0.51-{i686,x86_64}-bin-2 - OpenJDK-1.7.0.51-2-{i686,x86_64}-bin - upload them with the same name as the prebious one? (not good, since the checksum changes) - not upload them at all Regards Pierre Hmm? 2.4.7 is 1.7 u55b14: Ugh. Sorry. Noticed the date about 3 seconds too late. FYI, we should see u60 and 2.5.0 hopefully right around the middle of next month. --DJ -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] CA Certificates
On 03/06/14 11:15, Bruce Dubbs wrote: Henrik /KaarPoSoft wrote: Dear all, On http://www.linuxfromscratch.org/blfs/view/svn/postlfs/cacerts.html you indicate to download CA Certificates from: http://mxr.mozilla.org/mozilla/source/security/nss/lib/ckfw/builtins/certdata.txt?raw=1 However, on the mxr frontpage http://mxr.mozilla.org/ the branch Mozilla CVS http://mxr.mozilla.org/mozilla/ is described as follows: QUOTE This contains the entire current CVS repository. For Gecko, XULRunner, and Firefox, CVS trunk is no longer the trunk, and is instead used for Gecko 1.9 / Firefox 3 and the 1.9.0.* / 3.0.* security releases. UNQUOTE So I would like to suggest that alternative sources may be described a well. See e.g. http://kaarpux.kaarposoft.dk/packages/c/certdata.html#certificates_from_mozilla (You are more that welcome to link to this page, if you find it appropriate). We are not the only ones struggling to figure out which branch to use. See e.g. the thread started here: http://curl.haxx.se/mail/archive-2013-12/0033.html The integrity of the certdata.txt file is essential, so I would also like to suggest that 1) you download from https://hg.mozilla.org/... 2) you include a sha256 checksum for the file. It would seem that https://hg.mozilla.org/releases/mozilla-release/raw-file/058ed8ee9adf/security/nss/lib/ckfw/builtins/certdata.txt is correct right now, but I don't see a way to specify 'current' or 'latest' for the raw file that we need. We could write a script to download the html and then parse the raw file URL, but that would require downloading a 5M file just to get the url of a 1.5M files. :( I don't see how we can give a checksum if the file is changing. We need to let users decide which version they need. I'd be interested in other ideas. -- Bruce Couple of possible suggestions. First, and easiest, leave it alone. I know that the file in that repo was updated at least fairly recently. I'd imagine it will continue unless they are killing off maintenance on 1.9. Second, look at the url in the comments of the perl script which was taken from Fedora. It has a link to their package, just follow their lead. A third possible solution is to add a comment to the each of the 4 Mozilla packages to update a copy of the cacerts.txt on Anduin from whichever is the latest package at the time of update. Personally, the third is my favorite, but it adds editor work. HTH --DJ -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] BLFS Systemd Additions
On 02/21/14 13:20, Bruce Dubbs wrote: Armin K. wrote: I know what you meant. I have no problems with shipping package specific notes for now. Maintaining seperate BLFS is something I can't do alone and maintaining same instructions in one branch, then generating two different books is something I don't know a) to do b) if it's possible. a. Can be cured. b. Yes, it's possible. Though I've never messed with it, it really doesn't look too difficult. Lookup use of xsl:when. Since this direction interests me, I'll take a look maybe next week. --DJ -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] BLFS Systemd Additions
On 02/22/14 11:29, DJ Lucas wrote: Profiling is really what we are looking for, and we already have the functionality for two pass profiling in our Makefile for BLFS. See the attached patch. Oh, I forgot to mention, I didn't check to see if bind includes a unit file, it was just used as an example. --DJ -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] possibly uneeded IcedTea patch (Was: [lfs-patches] r2759 - trunk/icedtea)
Author: fernando Date: Thu Nov 7 03:35:00 2013 New Revision: 2759 Log: Bump patches versions for icedtea-2.4.3. Added: trunk/icedtea/icedtea-2.4.3-add_cacerts-1.patch (contents, props changed) Bruce added the script into the BLFS cacerts package, so no longer need to use the patch. Just generate the cacerts before running the tests and all is good. --DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] BLFS Version 7.4 is released
Bruce Dubbs bruce.du...@gmail.com wrote: After five years, The BLFS Team is happy to present version 7.4 of Beyond Linux From Scratch. Wow! It's been a long time since I've heard even mention of a BLFS release. I can barely believe that it is possible with the number of packages now days. Congrats! Excellent work y'all!!! --DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] libdrm
Ken Moffat zarniwh...@ntlworld.com wrote: On Sun, Sep 01, 2013 at 04:45:29PM +0200, Igor Živković wrote: Is there any reason why libdrm is in the general libraries section? I'd propose to move it to X installation before MesaLib or at very least to X libraries section since it depends on X libraries. What do you think? I think it's there because of history. I can see it in the 6.2.0 version of the book in the museum. We still had traditional X (i.e. monolithic) in both Xorg and XFree86 flavours, as well as Modular X (7.1). At that time, Mesa was listed among the X libraries (because it required libraries from Xorg. No idea why libdrm went into general libraries. Personally, I think moving it to chapter 24 makes a lot of sense - the only reason I can see for installing libdrm is to build Xorg. History is right, despite the seemingly logical order, Xorg packages are from X.org and the libdrm package is from freedesktop.org. Like Mesa, it isn't part of core X. We already make a few exceptions to the core for xclock, twm, and xinit (which could probably go at this point), but those are all formerly a part of the distribution. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Security: OpenJDK, mkcacerts.sh - zero, none, no certificate installed
Fernando de Oliveira fam...@yahoo.com.br wrote: Em 20-07-2013 10:54, Bruce Dubbs escreveu: On Fri, Jul 19, 2013 at 9:38 PM, Fernando de Oliveira Problem: no root certificates generated for some non-english locales. In this message, first, I describe the problem, then the solution. Many users may have the bug, so a test is necessary. This test is described below. I will need to spend some time studying this. OK. The most important point was informing about the problem, so, if someone has a problem, we can send to the thread, and user will not have to rebuild everything. This is useful even after the book being fixed. ... $ echo $LANG pt_BR.utf8 That's probably why I don't see that. Today, failed for many other locales. ... First, correct the patch...: sed 's/echo yes | /echo yes | env LC_ALL=C /' \ icedtea-2.4.1-add_cacerts-1.patch icedtea-2.4.1-add_cacerts-2.patch OK, so we fix the patch. Thanks. This is almost all. Unfortunately, the binaries do not have the certificates, almost sure they need to be corrected. Shall I do it or wait for a future release? The rest of the message was for others to be informed about the problem and have a solution. 5. Difference Between Original And New Scripts $ diff -Naur mkcacerts-blfs-1.0.7.40-2.4.1.sh mkcacerts-new-1.0.7.40-2.4.1.sh ... I'm not sure where this change goes. Is it not the same as the patch? Yes, it was just for cross-checking if the previous part was correct. Cat command and diffs are all broken by the mail client. I hope that those who read it understand the necessity of editing (here, I am not referring to you, Bruce). One command was wrong. When editing, I deleted the certificates filename: sh mkcacerts.sh -d /etc/ssl/certs/ -k bin/keytool -s /usr/bin/openssl \ -o ../ It should read: sh mkcacerts.sh -d /etc/ssl/certs/ -k bin/keytool \ -s /usr/bin/openssl -o ../certificates output (-o ../) was being sent to a directory. Please, do not bother about allowing a big message through. It would be useful for other, not you, but readers can edit the mail client broken lines, of original one, I think... Good catch. One request though, since this will not likely get upstreamed, why not just install mkcacerts.sh in the jre bin directory and add a blurb about it in the cacerts page? Could even add a conditional for it during update. --DJ -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] OpenJDK-1.7.0.40-2.4.1
Fernando de Oliveira fam...@yahoo.com.br wrote: Instructions in BLFS for OpenJDK were very robust. Originally, by DJ, I think last update was from Bruce, cannot remember if Armin or Ken touched it. Unfortunately, since icedtea-2.4.0, problems started. For all versions, I have been using book's patches, just renaming according to the icedtea's version. First, the icedtea-2.4.0-add_cacerts-1.patch was first to finally fail a chunk (hope this sentence is written in an understandable language). Even though, it seemed to partially work, the three sites where I have to use java understood everything from icedtea-web-1.4. Now, with icedtea-2.4.1, java seems to be working in most pages, but in the test pages and in the three sites where I have to use java, firefox freezes and in one of the latter, I get an error message to install java plugin. Another symptom of problems is that from some older version and up to and including icedtea-2.4.0, first time a site wanted to execute (or install, cannot remember) an applet, a popup informed and requested confirmation, and if this reply was permanent. Now, with 2.4.1, it freezes firefox, no popup, perhaps this is exactly what is causing the freezing. So, as this is a *critical security issue*, at least for me, I am asking help. *Please, someone, update the book.* Book's version: Latest OpenJDK1.7.0.9 1.7.0.40 (at ArchLinux) IcedTea2.3.3 2.4.1 (we already have a ticket) IcedTea went up to 2.3.9, before turning 2.4.0, so it has been a long time since last update. Run Firefox from a terminal and see if this is the save writ as this one: http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=1495 --DJ -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] OpenJDK-1.7.0.40-2.4.1
DJ Lucas d...@linuxfromscratch.org wrote: Fernando de Oliveira fam...@yahoo.com.br wrote: Instructions in BLFS for OpenJDK were very robust. Originally, by DJ, I think last update was from Bruce, cannot remember if Armin or Ken touched it. Unfortunately, since icedtea-2.4.0, problems started. Snip First, the icedtea-2.4.0-add_cacerts-1.patch was first to finally fail a chunk (hope this sentence is written in an understandable language). Even though, it seemed to partially work, the three sites where I have to use java understood everything from icedtea-web-1.4. This seems relevant as well: http://icedtea.classpath.org/hg/icedtea-web/rev/2469bedc6d63 --DJ -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] known to build on 7.3
On 03/13/2013 05:02 AM, Pierre Labastie wrote: Le 13/03/2013 07:52, Thomas Trepl a écrit : Hiho, in another thread there was the info that comments about packages built against LFS-7.3 would be of interest. I have so far: which bc pcre popt gpm ed sudo openssl openssh curl libidn wget ntp dhcpcd libtirpc attr acl libcap rpcbind nfsutils-client traceroute nettools libgpgerror libgcrypt vpnc(not in book) unzip zip lynx java(only binary installation) tcl sqlite db expat pth libffi python2 gamin glib2 ant libxml2 libxslt cyrus-sasl openldap-client postfix mailx sgml-common sgml-dtd docbook-xml docbook-xsl docbook-dsssl fop tidy xmlto fcron alsa-lib alsa-utils bind-client cvs rcs apr neon apr-util subversion libusb libusb-compat pciutils usbutils wirelesstools -- Thomas You can add gcc, tested on both 64 bit and 32 bit. I also built openjdk starting from the gcj jvm, using upstream ant and ecj binaries. If anybody is interested, I can share the instructions. Pierre Hmm. I haven't done that in quite a while. I'd be interested to see what is required now days. I always just bootstrap from my previous incarnation. I guess it shouldn't be necessary to provide anything more than CLASSPATH now, but am unsure. Did you do a system install or keep it isolated in /opt (like in BLFS)? -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xulrunner-18.0.1 installed files ownership
Randy McMurchy ra...@linuxfromscratch.org wrote: If you did that as the root user, then the package is broken. They use their own install script as opposed to the system's install. I ran across this a long time ago, probably when system XUL patch went in. Replacing it broke stuff. --DJ -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Samba4
On 01/31/2013 12:39 PM, Thomas Trepl wrote: Hi, On 01/28/2013 06:48 PM, Thomas Trepl wrote (in another thread): Well, dropping heimdal was for sure not without a reason. At least at the moment, for me its not quite a problem to continue without as Samba4 seems to support mit-krb5 well (which means so far, it compiles well against it). For the moment, I think there's no real pressure to get heimdal back - except you what it. I'll continue my survey of Samba4 for the moment with mit-krb5. They have prepared enough challenges to master, one is their waf build system, the dns stuff and such. Maybe, if interests are there, we could rethink that when Samba4 manages it to get into the book. It turns out that this is no longer true. Seems so that AD-controller functionality is indeed only available when having Heimdal around. Even Samba4 compiles well against mit-krb5, it disables building AD-DC functionality (even though the build log says that AD support gets compiled in. Maybe true for becoming an AD member, but the controller part isn't built/installed. You can check that simply whether samba-tool gets installed or not). Also the Arch developers came across this issue and build their Samba4 packet with the bundled heimdal package and there is/was a discussion on the Samba ML in order to get over that issue. Currently, there seems to be no way to get AD-DC functionality with using any krb package except (the bundled?) Heimdal. I put the bundled in parenthesis as I currently do not have a standalone Heimdal installation (and will do a fresh build of {,B}LFS next time to have a clean environment without mit-krb5 leavings around. Unfortunatly I missed to take a snapshot of the VM before installing mit-krb5 and such). I'm just testing with LINKFLAGS=-ltirpc ./configure \ --prefix=/usr \ --sysconfdir=/etc \ --localstatedir=/var\ --with-piddir=/run \ --enable-fhs\ --enable-nss-wrapper\ --enable-socket-wrapper \ --disable-rpath-install \ --dns-backend=SAMBA_INTERNAL --with-dnsupdate \ --without-pam \ --with-ads --with-ldap --with-swat --with-winbind --enable-gnutls Just in case someone is interested in building an AD controller... Hmm...I forget exactly how it works technically, but an NTLM ticket needs to be attached (for lack of a better word) to the kerberos ticket. I remember reading about it a few years ago, but this was the part that Microsoft did not give back to MIT when they used their kerberos implementation to design Windows 2000. It was a bit of a sore spot for MIT for a while. This is not the same article that I had read so long ago, but it gives a brief overview. http://www.networkworld.com/news/2000/0511kerberos.html Notice the date, a lot has changed since then, and Microsoft has given quite a bit of help to the Samba 4 devs (I'm not sure how that arrangement came about). At any rate, I guess Heimdal must have some feature not present in the MIT reference implementation. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Samba4
On 01/31/2013 02:40 PM, Randy McMurchy wrote: Thomas Trepl wrote these words on 01/31/13 12:39 CST: It turns out that this is no longer true. Seems so that AD-controller functionality is indeed only available when having Heimdal around. [snip] Currently, there seems to be no way to get AD-DC functionality with using any krb package except (the bundled?) Heimdal. I am just finishing up a Heimdal build. I will put it back into the book and maintain it going forward. Thanks for the great observations, Thomas. While glad to have it back, I'm not sure that fixes the original problem. Looks like the in-tree version might have some NTDS specific hacks in both the kdc and libs. Heimdal git looks much closer than 1.5.3, though much further along, but ATM, doesn't look like current stable will work with Samba4. I may be confused, but home page says that 1.5.2 is current stable, so is 1.5.3 is a development version? Hopefully 1.5.4 is sooner, rather than later. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] krb5 + tcl,python
On 01/27/2013 03:22 AM, Thomas Trepl wrote: Hi all, I tried to compile krb5 (as prerequisite for Samba4) but the make process failed with following error: Which implementation of krb5? I am not positive, but I think the Samba devs went with Heimdal in-tree over MIT. While either may be usable as a system krb5, if Heimdal, it likely will be getting more attention nowadays WRT to Samba. I have always preferred (and used) Heimdal over MIT due only to the license. Last time I looked, configuration of both was nearly identical. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] A good example why Command Explanations are important
On 01/25/2013 10:40 PM, Bruce Dubbs wrote: Randy McMurchy wrote: 3) The --libexecdir= switch points to a non-standard location. libexecdir location has always been /usr/lib/packagename, Yes, historically it was, however, it should now be where the package maintainer chooses for it to be unless it violates the FHS. yet in the Obexd instructions, the location varies from this. Is this because other packages expect it to be where the libexecdir switch points to it, or simply because the Editor who originally put the package in BLFS decided to go against the standards that we have created? I completely forgot about the --libexecdir blfs standard. This was just discussed in April of 2012 where it was decided that {,B}LFS scrap that rule in favor of the standard /usr/libexec, as per the draft of FHS 3.0. Unfortunately, the document was never finalized due to the attack on kernel.org. Looks like Andy, Bruce, Ken, and myself agreed on the new standard. Armin did not at the time, but I think that his disagreement was with the separation of libexec items, not with the spec. Randy wasn't back into editing mode yet for that one. Here is a link to that discussion: http://comments.gmane.org/gmane.linux.lfs.beyond.devel/21616 It might be interesting to get each person's take on this in light of the continued delay of FHS-3.0. Even though I'm not editing, FWIW, seeing as I've already adopted it (along with the /usr merge and systemd (still yuck, but standardized and works well enough to be tolerable)), my thoughts on the topic remain unchanged. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] icedtea-web on seamonkey problem
On 01/11/2013 11:39 PM, JZB2 wrote: On Friday, January 11, 2013 04:38:10 PM Fernando de Oliveira wrote: icedtea-2.3.3 (openjdk-1.7.0_09, built from source, after binary, per BLFS) seamonkey-2.13.2 icedtea-web-1.3 I my case, $ seamonkey --version Mozilla SeaMonkey 2.15 Thanks for the feedback. That's my next step, to try a newer monkey... $ xulrunner -version Mozilla XULRunner 18.0 - 20130109221022 That's a good point. I did not specifically install xulrunner. I simply install seamonkey as described in BLFS. The instructions for installing icedtea-web reference xulrunner as a prerequisite, but if one installs firefox, it seems that xulrunner is not needed per se. I made the logical inference then that since seamonkey is a superset of firefox, seamonkey alone should have sufficed. I was encoured in this belief by being able to build icedtea-web by merely adding the proper MOZILLA_CFLAGS and MOZILLA_LIBS env. vars., and installing seamonkey alone, no firefox and no xulrunner. I therefore assumed that no further dependence on xulrunner is needed. Do you see any runtime depencencies that this assumption does not cover? Yes. If you run ldd on the IcedTeaPlugin.so, the lack of not found errors should make it fairly obvious. If you just build XULRunner by the book, it very well might work without rebuilding the plugin (assuming your modifications didn't break anything else), but I think I'd still rebuild the plugin after XUL. If you really want to get around XUL, adding LD_LIBRARY_PATH to whatever the Seamonkey equivalent is of the run-mozilla.sh script might be sufficient (I honestly don't know), but you should be aware that it is not tested...anywhere. Also, please don't make that variable or the libs global either. Experimentation is fun, but I think I'd stick to the book on this one (been there, done that) unless you are just really cramped for space. Also, for future reference, in general it is best practice to consult the support team of the distro unless you are absolutely sure about a bug. It's not an ego thing, it's just that it can look bad on the distro if too many invalid bugs are opened from its users (so I guess it is an ego thing). :-) When I was actually active on distro-pkg-dev, I tackled more than a few off list. Please do take the time to add a comment to the icedtea bug stating that this was user error and can be closed invalid (or, if you have permissions, just close it yourself). As evidenced by the number of bugs (a majority of which will eventually be tracked down to developers depending on undocumented features or deprecated code in the closed plugin), those guys have enough on their plate. :-) -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] My status (Was: Happy new year)
On 12/31/2012 04:04 PM, Bruce Dubbs wrote: I've updated the dates for the book to 2013 and archived the 2012 changelog. You won't see the update in -book since it is over 200K and I deleted the posting. The on-line book will be updated in a few hours. Here are some statistics: 1469 Total Changes 127 abenton 247 bdubbs 4 Chris 37 dj 191 ken 719 krejzi 8 randy 135 rthomsen 1 wblaszcz That's a lot of work! Thanks for all who contributed and those who commented on the mailing lists. -- Bruce Wow! Armin takes the prize! Good work all! The book is, and continues to be, in the best shape it has been in a very long time. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] [blfs-book] r10911 - in trunk/BOOK: . introduction/welcome x/installing
On 12/30/2012 04:17 PM, kre...@linuxfromscratch.org wrote: Modified: trunk/BOOK/x/installing/installing.xml === --- trunk/BOOK/x/installing/installing.xml2012-12-30 22:09:54 UTC (rev 10910) +++ trunk/BOOK/x/installing/installing.xml2012-12-30 22:17:40 UTC (rev 10911) @@ -67,10 +67,10 @@ xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=xcursor-themes.xml/ xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=x7font.xml/ xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=xkeyboard-config.xml/ + xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=xorg-server.xml/ + xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=x7driver.xml/ xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=printproto.xml/ xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=libXp.xml/ - xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=xorg-server.xml/ - xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=x7driver.xml/ xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=twm.xml/ xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=xterm.xml/ xi:include xmlns:xi=http://www.w3.org/2001/XInclude; href=xclock.xml/ FYI, printproto and libXP are not official and haven't been updated in a long time. They should both be removed at this point as the only remaining consumer was OpenJDK and that went away after Bruce's last update to it. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xorg-7.7 MANDATORY paths
On 11/03/2012 11:28 AM, Fernando de Oliveira wrote: I have noticed some modifications in the Introduction to Xorg-7.7 page, particularly: sed 's@/usr/X11R6@PREFIX@g' -i /etc/man_db.conf In my host LFS7.1, I have not done that, there is the symlink /usr/X11R6 - /usr and just in case, I have added /usr/X11 - /usr Now in the new LFS7.2, Neither of those modifications should be needed, but the instructions referenced below should not be run if building in /usr. If the book says to run them unconditionally, then I've made a mistake when reorganizing the section a while back. I'm not in a position to look at it for the moment. # grep X11 /etc/man_db.conf MANPATH_MAP /usr/X11R6/bin /usr/X11R6/man MANPATH_MAP/usr/bin/X11 /usr/X11R6/man MANDB_MAP/usr/X11R6/man /var/cache/man/X11R6 I used that sed, however: # grep X11 /etc/man_db.conf MANPATH_MAP/usr/bin/X11 PREFIX/man MANDB_MAP PREFIX/man /var/cache/man/X11R6 1. Is it correct having PREFIX/man? 2. Should not /var/cache/man/X11R6 be replaced too, perhaps by /var/cache/man/X11 (s@/var/cache/man/X11R6@/var/cache/man/X11@g)? The second question is not so important, I believe, but it seems necessary, just for consistency reason (not using X11R6, but X11). Thanks. []s, Fernando I think for builds with a prefix other than /usr, it would be better to just add all three of the paths to the end and force a cache update in the configuration section (using 'cat /etc/mandb.conf EOF' syntax). It should be tested by somebody who uses something other than /usr as the installation prefix. I'm still of the mind to completely drop the alternate prefix business for everything (the wiki is the place for differing opinions), but others seem to still want and use it, so it should stay as long as somebody is keeping an eye on it for breakage. I generally do keep it in mind, but only to a minimum as I do not actually test in an alternate prefix. Also, does mandb.conf support a source/include directive? Best would be to use the mandb.d/* method if supported. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] iced tea
On 10/20/2012 02:38 PM, Bruce Dubbs wrote: DJ Lucas wrote: BTW, what do we call this version. It's icedtea-2.3.3, but the build says: java version 1.7.0_0 OpenJDK Runtime Environment (build 1.7.0-b36) OpenJDK 64-Bit Server VM (build 23.2-b09, mixed mode) Right now we are calling the current version OpenJDK-1.7.0.5, but the most recent version does not seem to fall into that naming convention. Odd...it should be pulling that value from the OpenJDK version (which should already match the closed version in git). Upstream must have broken that somehow. Nothing on distro-pkg-dev list yet. The earlier post was before the final executable was done. Now I have: $ openjdk.build/j2sdk-image/bin/javac -version javac 1.7.0_09 $ openjdk.build/j2sdk-image/bin/java -version java version 1.7.0_09 OpenJDK Runtime Environment (IcedTea7 2.3.3) (Linux From Scratch build 1.7.0_09-b30) OpenJDK 64-Bit Server VM (build 23.2-b09, mixed mode) The question still remains, what should we call the destination directory. What we have in the book is /opt/OpenJDK-1.7.0.5: $ /opt/OpenJDK-1.7.0.5-bin/bin/javac -version javac 1.7.0_04 $ /opt/OpenJDK-1.7.0.5-bin/bin/java -version java version 1.7.0_04-icedtea OpenJDK Runtime Environment (IcedTea7 2.2) (linux-gnu build 1.7.0_04-icedtea-b21) OpenJDK 64-Bit Server VM (build 23.0-b21, mixed mode) So do we just call it 'OpenJDK-1.7.0.9' ? -- Bruce That was the convention I had used before. I'm not sure why the binary u5 is showing u4 though. root [ ~ ]# java -version java version 1.7.0_05-icedtea OpenJDK Runtime Environment (IcedTea7 2.2.1) (linux-gnu build 1.7.0_05-icedtea-b21) OpenJDK 64-Bit Server VM (build 23.0-b21, mixed mode) root [ ~ ]# ls -l $JAVA_HOME lrwxrwxrwx 1 root root 19 Oct 4 22:46 /opt/OpenJDK - OpenJDK-1.7.0.5-bin Either way, u4 should be fine for a bootstrap. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] iced tea
On 10/19/2012 12:37 PM, Bruce Dubbs wrote: I've been looking at updating iced tea. rant I'm into problems because the make files have hard coded paths into make files. /usr/bin/head (which is in /bin) and /bin/touch (which is in /usr/bin/touch). What are these guys thinking? why can't they use the PATH? /rant Yes, the problem and the current fix have been there since Tushar put the JDK in the book back at 1.4.2 IIRC. This will not likely be fixed in OpenJDK-7 but is already done in 8 IIRC. Thanks for doing this Bruce. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] OpenJDK-1.7.0.9 (was: iced tea)
On 10/19/2012 05:54 PM, Bruce Dubbs wrote: Fernando de Oliveira wrote: Previous attachment had a line out of order and with a typo. []s, Fernando Thanks. That's helpful. Right now I'm running tests, but that takes quite a while. I think we should change the book to do a 'make download' to get the aux files. I also think the patch for fixing paths can be done with a sed and that makes is a little more obvious about what it is doing. It was a little difficult to figure out what was going on here, but I think I've got it now. Simplifying the build instructions should help others. Right now I'm up to 139 failures in the tests (3536 Passed). I'll examine it a little closer when it finally finishes. -- Bruce A couple more things... I did put reasons in the patch for the test failures. If they are the same, then that particular test can go. Also, the certificates patch could be dropped and done manually if desired. It was written for upstream and rejected as they did not want to use a bash script for UTF8 issues that I was unable to reproduce. It might be nice to make it a part of the ca-certificates section at some point. I still plan to upstream that patch again at some point in the future. I have a java version written by SUSE that I've modified to replace the bash script, but I haven't gotten around to writing a verify function (required in the case of crypto not built with NSS). -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] iced tea
On 10/20/2012 10:55 AM, Bruce Dubbs wrote: DJ Lucas wrote: On 10/19/2012 12:37 PM, Bruce Dubbs wrote: I've been looking at updating iced tea. rant I'm into problems because the make files have hard coded paths into make files. /usr/bin/head (which is in /bin) and /bin/touch (which is in /usr/bin/touch). What are these guys thinking? why can't they use the PATH? /rant Yes, the problem and the current fix have been there since Tushar put the JDK in the book back at 1.4.2 IIRC. This will not likely be fixed in OpenJDK-7 but is already done in 8 IIRC. I'm getting there. I found that I can't do a sed on the paths because the files are not exposed until after make extracts the downloaded tarballs. The patch still works though. The build is OK, but there are still lots of test failures. I'm doing a rebuild now in a different screen. BTW, what do we call this version. It's icedtea-2.3.3, but the build says: java version 1.7.0_0 OpenJDK Runtime Environment (build 1.7.0-b36) OpenJDK 64-Bit Server VM (build 23.2-b09, mixed mode) Right now we are calling the current version OpenJDK-1.7.0.5, but the most recent version does not seem to fall into that naming convention. -- Bruce Odd...it should be pulling that value from the OpenJDK version (which should already match the closed version in git). Upstream must have broken that somehow. Nothing on distro-pkg-dev list yet. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] [lfs-book] [LFS Trac] #3162: Glibc 2.16.0 Suggestions
On 08/20/2012 04:11 PM, Randy McMurchy wrote: Welcome back! LFS Trac wrote these words on 08/20/12 15:37 CST: Comment(by Krejzi): I just want to say that I personaly won't make use of /usr/libexec anywhere where I make the changes untill it is official. I was even thinking about reverting that for Postfix and vte-0.28 untill it's official. Are those the only two packages in BLFS currently using /usr/libexec? I realize that the issue right now is with LFS's Glibc, so just for the record I prefer individually named directories. But I can go either way, it is rather small. I know I will continue in my personal builds to create individual directories for each package the needs a libexec, but that's just me. Whatever we decide for the books is fine with me. I cc'd BLFS-Dev for anyone wishing to discuss this issue as it pertains to BLFS. It used to cause minor headaches with Gnome-2.26-32 IIRC, and a fair bit of patching was required for a few packages, but upstream was accepting of patches (except for GDM which eventually got a proper fix as suggested by Dan (ck-* files)), but I'm not sure now days. I build everything in /usr and let /usr/libexec be, seeing as it's now a standard path in the v3 draft. But, official FHS-3.0 may take some time. The repo was lost a while ago (when kernel.org got hit I think). I don't know if they ever got it back. /usr/libexec will eventually be standard, so it's probably good to continue in that direction. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] Xorg introduction page
Guys, I'm about to commit the md5sums and loops into the book instructions themselves (as per Armin's work a couple of weeks ago). On a later commit, I also plan to roll the introduction into the chapter header page as there aren't 3 X Window Systems in the book any longer and to get rid of the x7- page names and entities as part of a more general cleanup. Even though it is pretty strait forward, the latter will affect a lot of pages, does anyone have objections? -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] [blfs-book] r10477 - in trunk/BOOK: . introduction/welcome x/installing
On 07/29/2012 05:48 PM, Bruce Dubbs wrote: Armin K. wrote: On 07/29/2012 10:09 PM, d...@linuxfromscratch.org wrote: Author: dj Date: 2012-07-29 14:09:42 -0600 (Sun, 29 Jul 2012) New Revision: 10477 Modified: trunk/BOOK/general.ent trunk/BOOK/introduction/welcome/changelog.xml trunk/BOOK/x/installing/installing.xml trunk/BOOK/x/installing/x7app.xml trunk/BOOK/x/installing/x7font.xml trunk/BOOK/x/installing/x7lib.xml trunk/BOOK/x/installing/x7proto.xml trunk/BOOK/x/installing/xorg-config.xml trunk/BOOK/x/installing/xorg7.xml Log: Removed external Xorg wget and md5sums files, and included for-in-do loops. Um, not bad ... But I think we still should split the packages into seperate ones (not including Xorg Applications) and add loop as an option for unpatient ones. Eventually that will happen, piece by piece. This only gets us passed the added version number and gets all of the current instructions into the book. I'm not sure that I understand why you want to make a special case for xorg applications though. Also, I dislike the use of sudo by default there. It isn't even listed in dependencies, Sure it is...it is listed as recommended in both the introduction to Xorg (as was wget) and for now in xorg-proto (the first group of packages that use it), but I think we'll use Bruce's suggestion below to make it completely optional. and not everyone installs sudo. I have it installed, but I don't like it and I rarely use it. I think we could use some kind of variables to check wether: User would like to compile and install everything as root (with a BIG warning), wether to use sudo (with explanation that user should make his user not to ask for password to speed up things) or even to use su -c make install (Yes, it asks for password every time). Personally I do like sudo but I also understand Armin's point. I also prefer using pushd $packagedir ... popd to the cd constructs. For the install, we might want to consider something like: lt;as rootgt; make install With a note/entity like: !ENTITY as_root noteparaWhen installing multiple packages in a script, the installation needs to be done as root. There are three general options that can be used to do this:para orderedlist listitemRun the entire script as the root user. (not recommended) /listitem listitemUse the applicationsudo/application program (ulink linkend='sudo') with or without methods for avoiding requests for the root password./listitem listitemUse commandsu -c make install/command which will ask for the root passwords for every iteration of the loop./listitem /orderedlist/para This could be reused elsewhere should the need arise again (though orderedlist is giving me a fit in an entity). Just prior to the build commands. That is: as_root; screenuserinputfor ... packagedir=${package%.tar.bz2} tar -xf $package pushd $packagedir ./configure $XORG_CONFIG make lt;as rootgt; make install popd rm -r $packagedir done/userinput/screen -- Bruce I think that should meet everyone's expectations. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xorg Applications
On 07/12/2012 01:35 AM, Guy Dalziel wrote: On Wed, Jul 11, 2012 at 06:18:38PM -0400, Baho Utot wrote: On 07/11/2012 01:54 AM, Guy Dalziel wrote: [putolin] Can anyone recall why we don't build as root in the first place? That's a good way to trash your system. As discussed before, I'm well aware of that; I was looking for an explanation beyond the obvious. I'm still convinced that compiling as root has issues with permissions of the compiled files. Not that I'm aware of. In fact, I'm pretty sure we'd hear about it if it did. I imagine that a lot of user run build scripts as root. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xorg Applications
On 07/10/2012 10:52 AM, Armin K. wrote: Hi there. There was some talk about splitting Xorg Packages into seperate ones. I went with Xorg Applications first. I did not make instructions for every package, since there are too many of them, and nearly all of them are required at runtime of Xorg Server, so it's dependency list would become big. Anyways, I took the way of generating the wget and md5sum lists manualy instead of putting them to server. These lists are generated using xml entities of each package. You can see how that looks here: http://linuxfromscratch.org/~krejzi/xorg/x/x7app.html Also, there is a patch for that at http://linuxfromscratch.org/~krejzi/xorg.patch.gz if desired. I won't commit anything if someone does not like that way. Drivers splitup is next on my TODO list. Comments? Excellent! Thanks Armin. Looks good. What I was thinking earlier is what you have, plus the loop to build the packages, using sudo as mentioned elsewhere in this thread. I have a list of seds that you can use as instructions to automate adding the packages and md5sums to the xml on -support in the thread titled Worlds worst sedder. There were a couple of other things I wanted to do with the page, but what you have now can be committed any time as that immediately fixes the problem of individual package upgrades. The other things I wanted to do were to get rid of the wget file and just use the md5 file for both - just run the grep on the md5sums file and then pipe to awk '{print $2}'. I had also planned to add some abuse of bridgehead and segmented list for package descriptions for those who wish to build the packages separately. The first of those two is easy, and the other will be time consuming, plus I haven't gotten output I like yet, but both of those can wait. If you have the time and want to do what you have above, please do so as time and interest permit. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] [blfs-book] r10408 - in trunk/BOOK: introduction/important kde4/add kde4/core postlfs/filesystems postlfs/virtualization
On 07/11/2012 08:11 AM, Ragnar Thomsen wrote: On Tue, Jul 10, 2012 at 9:02 PM, Bruce Dubbs bruce.du...@gmail.com wrote: We've had discussions here before about this type of use. Generally we want to use 'an' to prevent two adjacent vowel sounds. In this case I think of lvm as 'ell vee em', so the 'an' is appropriate to to avoid the 'a ell' combination. Ok, I will revert it. -Ragnar- My favorite place to look up English grammar quirks is the the online writing lab at Purdue University. It is a great resource to have handy. http://owl.english.purdue.edu/ -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xorg Applications
On 07/11/2012 06:53 PM, Andrew Benton wrote: On Wed, 11 Jul 2012 12:10:01 +0100 Ken Moffat zarniwh...@ntlworld.com wrote: I drop a different set, e.g. I drop bdftopcf (I don't use the traditional fonts) and luit (I don't use xterm itself : the book does, so I think luit is required in the book, depending on what languages / locales you need - I might be way off base). Xterm doesn't require luit, it's an optional dep. I think you're right, xterm uses luit for internationalisation Conversely, I use xgamma so that photo prints bear *some* relationship to how the paper looks - saves me a lot of ink, and on my latest machines I also use it to reduce the worst colour-tints from my integrated graphics (must find more time to try my colorhug). But I agree that some apps can be dropped by some people :) Another difference - I rely on xmodmap to be able to easily type certain obscure glyphs. It's all about choice. And there's the rub; how can people make a choice if they don't have the info about what depends on what? If we just present all the apps in a long take it or leave it list people will have to work out for themselves which ones they need. It's the kind of information that should be in the book. Documenting dependencies is what the book is good at. Yep, I plan to do it, but I put it on the back burner because I got frustrated as I couldn't get it to render the way I wanted it. See my previous response to the OP for what I had in mind to cover both scripted build and those of us who package. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xorg Applications
On 07/10/2012 10:52 AM, Armin K. wrote: Hi there. There was some talk about splitting Xorg Packages into seperate ones. I went with Xorg Applications first. I did not make instructions for every package, since there are too many of them, and nearly all of them are required at runtime of Xorg Server, so it's dependency list would become big. Anyways, I took the way of generating the wget and md5sum lists manualy instead of putting them to server. These lists are generated using xml entities of each package. You can see how that looks here: http://linuxfromscratch.org/~krejzi/xorg/x/x7app.html Also, there is a patch for that at http://linuxfromscratch.org/~krejzi/xorg.patch.gz if desired. I won't commit anything if someone does not like that way. Drivers splitup is next on my TODO list. Comments? LOL..too funny. I just looked at your diff. I didn't want the package versions in general.ent either. We had almost identical idea of how to go about it: dj [ installing ]$ diff -au x7app.xml applications.xml --- x7app.xml2012-06-11 21:15:46.0 -0500 +++ applications.xml2012-06-26 00:21:21.0 -0500 @@ -4,28 +4,102 @@ !ENTITY % general-entities SYSTEM ../../general.ent %general-entities; - !ENTITY x7apps-download-http http://xorg.freedesktop.org/releases/individual/app/; - !ENTITY x7apps-download-ftp ftp://ftp.x.org/pub/individual/app/; - !ENTITY x7apps-wget files-anduin;/xorg/app-xorg7-release;.wget - !ENTITY x7apps-md5sum files-anduin;/xorg/app-xorg7-release;.md5 - !ENTITY x7apps-size 4.8 MB - !ENTITY x7apps-buildsize 39.2 MB - !ENTITY x7apps-time 1.5 SBU + !ENTITY applications-download-http http://xorg.freedesktop.org/releases/individual/app/; + !ENTITY applications-download-ftp ftp://ftp.x.org/pub/individual/app/; + !ENTITY applications-size 4.8 MB + !ENTITY applications-buildsize 39.2 MB + !ENTITY applications-time 1.5 SBU + + !-- versions and md5sums -- + !ENTITY bdftopcf-version 1.0.3 + !ENTITY bdftopcf-md5sum 4a7a4a848c43c42f7d499b60666434a4 + !ENTITY iceauth-version 1.0.5 + !ENTITY iceauth-md5sum 08e3f6b523da8b0af179f22f339508b2 + !ENTITY luit-version 1.1.1 + !ENTITY luit-md5sum c4a3664e08e5a47c120ff9263ee2f20c + !ENTITY mkfontdir-version1.0.7 + !ENTITY mkfontdir-md5sum 18c429148c96c2079edda922a2b67632 ... + !ENTITY xwud-version 1.0.4 + !ENTITY xwud-md5sum 3025b152b4f13fdffd0c46d0be587be6 ] -sect1 id=xorg7-app xreflabel=Xorg Applications - ?dbhtml filename=x7app.html? +sect1 id=x-applications xreflabel=Xorg Applications + ?dbhtml filename=x-applications.html? ... + paraFirst, create the md5sums file./para + +screenuserinputcat gt; applications.md5 lt;lt; EOF +# Xorg-7.7-1 application package md5sums +bdftopcf-md5sum; bdftopcf-bdftopcf-version;.tar.bz2 +iceauth-md5sum; iceauth-iceauth-version;.tar.bz2 ... +EOF/userinput/screen + + paraThen download the needed files:/para + +screenuserinputmkdir applications amp;amp; +cd applications amp;amp; +grep -v '^#' ../applications.md5 | sed 's@^[0-9a-z]\{32\} @@' | wget -i- -c \ -B http://xorg.freedesktop.org/releases/individual/app/ amp;amp; -md5sum -c ../app-xorg7-release;.md5/userinput/screen +md5sum -c ../applications.md5/userinput/screen ... Adding the spacing like you did is much easier to look at, and just ignore that sed...awk is better in this case. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Shared library file locations
On 07/08/2012 03:43 AM, Andrew Benton wrote: On Sun, 08 Jul 2012 01:14:43 +0100 Bruce Dubbs bruce.du...@gmail.com wrote: The question was a bit more subtle. Is it OK to put things only needed for building in /lib? Historically, that location was only for the minimum needed to get critical programs running at boot time before /usr was mounted. Huge disks are cheap now. The time when /usr needed to be on a separate partition has long since passed (I've never seen a system where /usr was a separate partition). I don't see the need for /usr at all now. Why don't we install everything into /? It simplifies configuring many packages as the don't need to be told --sysconfdir=/etc or --localstatedir=/var. There are other reasons to have /usr on a separate partition, but I question how often they are actually used. A School computer lab is a prime example, but I'm seeing more and more OS X and OpenDirectory in that space. Just did a Lion migration last week in fact. Unfortunately, we are at the point where /usr cannot be on a separate partition and still meet FHS-2.3 without an initrd (well, we could ditch udev and make BLFS a lot more difficult). Lots of changes coming though. With the loss of a few servers (in May IIRC), FHS-3.0 was put on hold. At this point, as much as I hate it, I can't suggest any way to handle it properly. We can suggest that users move forward with an initrd that can mount /usr but that still doesn't address the spec non-compliance at present. Thus far, FHS draft has done nothing to address the /usr debate as it can be _partially_ pushed off by providing an initrd that is capable of mounting /usr. RedHat is moving forward in F17 with the symlinked directories for /lib, /lib64, /bin, and /sbin, and Debian will continue to push forward with their sane approach to multi-lib instead of the abuse for binaries to the FHS spec that had been adopted by everyone but Debian (libqual for primary arch). Who knows which way things are going at this point? There is going to have to be some give though, because the spec dependencies aren't quite compatible (POSIX-FHS-LSB). The only downside is that for themes to work the environment variable XDG_DATA_DIRS=/share needs to be set. That could be fixed in the freedesktop source, or easier, symlink /share - /usr/share just like /lib, /libqual, /bin, and /sbin in the RedHat configuration. Do you mean to ditch the /usr directory completely? -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Shared library file locations
On 07/08/2012 11:58 AM, Bruce Dubbs wrote: DJ Lucas wrote: On 07/08/2012 03:43 AM, Andrew Benton wrote: The only downside is that for themes to work the environment variable XDG_DATA_DIRS=/share needs to be set. That could be fixed in the freedesktop source, or easier, symlink /share - /usr/share just like /lib, /libqual, /bin, and /sbin in the RedHat configuration. Do you mean to ditch the /usr directory completely? I thought the direction was the reverse: ditch /bin, /sbin, /lib and put everything in /usr. It is, I believe Andy was shooting for the reverse, which is why I asked. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] GDM (very minor) issue and a suggestion
Okay, so first of all, in the default X configuration, the gdm user should be added to the video group so that Gnome doesn't fallback, and also to the audio group so that sound works properly. This can be done when creating the user (there are other methods, but they are not in the book and I've left them out intentionally). The other issue is the gdm bootscript lives on. I've always disliked using a bootscript for gdm. We use wait lines in inittab, so the TTYs are never brought up because rc never finishes. Of course, it could be argued that they never should be brought up, but that leaves one without a failsafe should gdm get broken or get stuck in a loop (for which it will gracefully kill itself eventually IIRC). I would much rather drop the bootscript and start directly from inittab. Setting initdefault to 5 and adding a seventh line 7:5:respawn:/usr/sbin/gdm -nodaemon line is sufficient (and as an aside, it should get rid of the 7th tty patch). In the event of gdm fouling up, Ctrl+Alt+F{1..6} will still get you to a shell login prompt where you can 'telinit 3' to gracefully ditch gdm until fixed. Comments? # Begin /etc/inittab id:5:initdefault: si::sysinit:/etc/rc.d/init.d/rc S l0:0:wait:/etc/rc.d/init.d/rc 0 l1:S1:wait:/etc/rc.d/init.d/rc 1 l2:2:wait:/etc/rc.d/init.d/rc 2 l3:3:wait:/etc/rc.d/init.d/rc 3 l4:4:wait:/etc/rc.d/init.d/rc 4 l5:5:wait:/etc/rc.d/init.d/rc 5 l6:6:wait:/etc/rc.d/init.d/rc 6 ca:12345:ctrlaltdel:/sbin/shutdown -t1 -a -r now su:S016:once:/sbin/sulogin 1:2345:respawn:/sbin/agetty --noclear tty1 9600 2:2345:respawn:/sbin/agetty tty2 9600 3:2345:respawn:/sbin/agetty tty3 9600 4:2345:respawn:/sbin/agetty tty4 9600 5:2345:respawn:/sbin/agetty tty5 9600 6:2345:respawn:/sbin/agetty tty6 9600 7:5:respawn:/usr/sbin/gdm -nodaemon # End /etc/inittab -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Proposed changes
On 07/07/2012 08:50 PM, Ken Moffat wrote: Back in January last year, I made a proposal to add some discussion of vulnerabilities in the security chapter, and how to look at what the distros are doing to fix their builds (for vulnerabilities but also for when packages no longer build). There was some interest from users, but no comment from editors. At that time, I *thought* several editors were still active, but I was probably mistaken. My normal attitude is to not break things, and to follow the party line even if that means doing nothing. Je suis anarchiste bourgeois. [1] I then went off in a huff, but had to come back because changes in LFS-7.0 (glibc and bootscripts) broke nfs. Since then, we've got a lot more active editors, but I think those proposed changes would still be useful - it's not as if we are on top of known vulnerabilites. Some of this proposal is updated and expanded. Sounds good. Additionally, I think that now is a suitable time to replace gnome-media with gvolwheel, and to add another lightweight window manager (icewm). Cool cool. I don't use either, but I gather that others will appreciate them. When I started preparing this, my current BLFS version was 2012-06-23. You can find a rendered version of the changes, and a gzipped diff of the xml, at http://www.linuxfromscratch.org/~ken/tmp/ This contains the following changes: 1. new page in chapter 2 to explain static/shared libraries : in the book we now often use --disable-static but I don't think we've explained why. Sounds good. I don't know if this is still covered in LFS. It used to be before PureLFS. 2. add pointers to some distros in Beyond BLFS, change freshmeat.net to freecode.com, and add a link to a rpm2cpio script. Basically, as in last year's suggestion (the change to google/linux is no longer appropriate, that has become google/webhp and is not linux specific). Any particular reason for rpm2cpio specifically? CLFS has an rpm2targz that we could add. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Firefox/Thunderbird MimeType
On 07/05/2012 07:11 AM, Andrew Benton wrote: On Thu, 05 Jul 2012 07:15:20 +0100 DJ Lucas d...@linuxfromscratch.org wrote: We need to add the MimeType to the .desktop files for Firefox and Thunderbird. These are required to make them the default handlers in Gnome and KDE. I believe there are other ways in both, but I haven't looked in KDE, and haven't been able to figure it out in Gnome yet, but this is the way the spec wants it. I'm not sure about LXDE and XFCE...do they follow freedesktop.org standard? I don't know. With XFCE the firefox.desktop file that's in the book shows up in the menu. If I click on a html file in Thunar a dialog pops up that asks me to pick an application. If I add a MimeType line to the firefox.desktop it makes no difference, it still doesn't suggest to use Firefox, I still have to choose it from all the others. Once I've chosen it I can make it the default so it's not something that needs to be done every time. It only needs to be done once for each mime type. Also, the --enable-calendar switch is not included in the Thunderbird mozconfig that is in the book. I've never used this calendar but if it's useful to some people that's fine. Maybe we could put the option in the mozconfig with a line that says: # If you want to compile the Mozilla Calendar, uncomment this line: # ac_add_options --enable-calendar There is actually more to it than that. I didn't notice it missing until after I built it and didn't have a calendar sidebar. I just installed the binary lightening plugin for now. Will revisit all of my previous comments/commitments as soon as I have a fully functional system again. What I have for mime types is... Firefox: MimeType=text/html;text/xml;application/xhtml+xml;x-scheme-handler/http;x-scheme-handler/https; Thunderbird: MimeType=text/calendar;text/x-vcard;text/directory;application/mbox;message/rfc822;x-scheme-handler/mailto;x-scheme-handler/news;x-scheme-handler/snews;x-scheme-handler/nntp;x-scheme-handler/feed Fair enough. Those line make no difference that I'm aware of to LXPanel or to XFCE but I don't think it'll do any harm and if it helps someone then it should go in the book. I'm worried about the long lines in the book. What do you think of putting only the mime types in by default? Defaults would be: Firefox: MimeType=text/html;text/xml;application/xhtml+xml; Thunderbird: MimeType=text/x-vcard;text/directory;application/mbox;message/rfc822 And then providing a list in the book with the available additional types (only text/calendar is added for thunderbird) and handlers (all above) and instructions to modify as appropriate for the system it is going on? Again, I'll be happy to take care of it in a couple of days. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Firefox/Thunderbird MimeType
On 07/07/2012 11:23 AM, Bruce Dubbs wrote: DJ Lucas wrote: What I have for mime types is... Firefox: MimeType=text/html;text/xml;application/xhtml+xml;x-scheme-handler/http; x-scheme-handler/https; Thunderbird: MimeType=text/calendar;text/x-vcard;text/directory;application/mbox; message/rfc822;x-scheme-handler/mailto;x-scheme-handler/news; x-scheme-handler/snews;x-scheme-handler/nntp;x-scheme-handler/feed I'm worried about the long lines in the book. What do you think of putting only the mime types in by default? Defaults would be: Firefox: MimeType=text/html;text/xml;application/xhtml+xml; Thunderbird: MimeType=text/x-vcard;text/directory;application/mbox;message/rfc822 Sometimes long lines are a problem. I've noticed that the new Seamonkey breaks lines at dashes for me, but I'm sure some browsers don't. I'm no longer concerned about a pdf version of BLFS. We haven't created a version in years and I don't recall anyone saying they'd like one. I've seen some conversation about an epub version, but my real concern regarding Firefox (and long lines in general) is that i imagine quite a few people are still using links or lynx at that point. One of them puts spaces at the beginning of the new line. I actually forgot about SeaMonkey as I don't use it, but it too is affected and should have the same MimeType lines combined and added, and of course, the length is probably even more important there regarding links/lynx. I'm not sure of the additional mime-types for the irc client. Does SeaMonkey include .desktop file(s) by default? For MimeType we could just break it manually and add a note saying the line breaks should not be present. Alternatively, if we are going to use echo, we could do something like: m1=text/calendar;text/x-vcard;text/directory; m2=application/mbox;message/rfc822;x-scheme-handler/mailto; m3=x-scheme-handler/news;x-scheme-handler/snews; m4=x-scheme-handler/nntp;x-scheme-handler/feed echo MimeType=$m1$m2$m3$m4 some-file We might even get clever and write something like: m5=`echo x-scheme-handler/{mailto,news,snews,nntp,feed}|sed '/ /;/g'` m6=`echo text/{calendar,x-vcard,directory} |sed '/ /;/g'` echo MimeType=$m6;application/mbox;message/rfc822;$m5 some-file :) :) Yes, that works well for all. Sounds good. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] Firefox/Thunderbird MimeType
We need to add the MimeType to the .desktop files for Firefox and Thunderbird. These are required to make them the default handlers in Gnome and KDE. I believe there are other ways in both, but I haven't looked in KDE, and haven't been able to figure it out in Gnome yet, but this is the way the spec wants it. I'm not sure about LXDE and XFCE...do they follow freedesktop.org standard? Also, the --enable-calendar switch is not included in the Thunderbird mozconfig that is in the book. What I have for mime types is... Firefox: MimeType=text/html;text/xml;application/xhtml+xml;x-scheme-handler/http;x-scheme-handler/https; Thunderbird: MimeType=text/calendar;text/x-vcard;text/directory;application/mbox;message/rfc822;x-scheme-handler/mailto;x-scheme-handler/news;x-scheme-handler/snews;x-scheme-handler/nntp;x-scheme-handler/feed -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] pam_ck_connector and pam_loginuid
On 07/02/2012 01:47 AM, Armin K. wrote: It is not my fault that sudo is broken when it comes to pam. Everything else works but it and I don't want to sacrifice everything else for some stuff I don't care about. Just don't use system-session in sudo in the first place like I do. Well, that is the problem, sudo isn't broken, it is just doing what it was told to do. I'm going to disagree with you about sudo including session defaults (see below), but I'm going to follow your example nonetheless. I don't particularly like it as it was not what I had intended when I wrote those files, but it looks like you and Ubuntu do agree on it. They have added a common-session-noninteractive to handle this particular use case since I last visited their configuration (for which I based a good portion of BLFS's PAM configuration, though minimalist). While I dislike it, seeing as I did base it from theirs, I'm going to continue to follow their lead and do similar. ck_connector and loginuid will require no changes in your instructions this way, and the new can be used for cron and samba later on (as in Ubuntu, so this might even be expected by some users). As far as your sudo configuration, for what reason do you not follow the book? Only the above, or do you go well beyond the minimal defaults? If so, do you have any other suggestions? I wasn't aware that any other editors actually used it. While I'm browsing through it, I see a few other wrinkles, for instance, session limits should probably be added to system-session as well--while no limits are configured by default, it is probably surprising to an end user if they make changes and they don't see them immediately. I'm going to pick through it a little more as our defaults are getting a little long in the tooth (about 2 years old now). I'd like to keep pam_unix as a session module in system-session for logging though. In the case of sudo, it is an easy way to catch abuse cases of 'sudo su' or 'sudo bash' or similar. Do you have any other suggestions for the default PAM configuration? -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] pam_ck_connector and pam_loginuid
Shall I try this again pam_ck_connector and pam_loginuid should not be placed into system-session, but rather directly in login, {g,k,x}dm, sshd, etc. as session optional. These modules are only intended for login sessions. When using sudo, pam_ck_connector causes gnome-shell to segfault. I don't completely understand the interaction just yet, but I suspect it has something to do with root now owning the session cookie. :-) I think the best solution would be to create empty login-* configurations and append to those (include them by default in login configuration), and then include that into login utilities pam configuration. I'll try and get a look at it Wed night or so (maybe even tonight...we'll see). -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] xorg problems
On 06/27/2012 09:54 PM, Bruce Dubbs wrote: Ken Moffat wrote: On Wed, Jun 27, 2012 at 05:43:48PM -0500, Bruce Dubbs wrote: Ken Moffat wrote: On Wed, Jun 27, 2012 at 04:22:39PM -0500, Bruce Dubbs wrote: Good spot. I do not have CONFIG_INPUT_EVDEV set. LOL - I looked through my notes, which was why I suggested it : on both of my old x86_64 machines I was trying out evdev in 2009, when still building both i686 and x86_64 kernels, and had a similar problem: EVDEV set on the 64-bit kernels but not on the 32-bit. I rebuilt the kernel with CONFIG_INPUT_EVDEV and now evdev works fine. Now I'm having a problem with getting my two monitors working properly. It seems that anything I do, I get the same display reflected on both monitors instead of stretching across both. Dang, it's been a really long time since I messed with it. The man page suggests 'Option Dualhead true' needs Virtual. Maybe two monitor sections and only one screen section with 'Virtual 3840 x 1200', (or 1920 x 2400 for top and bottom) or ditch the Dualhead option if using Xinerama extension, but I don't honestly know if you can use both outputs with the nv driver that way. Playing with an Nvidia card is something I've been meaning to do for a couple of years now and I have just never gotten around to it. Whatever the final solution is, please replace the example in the book, or post it so I can. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Release Plan (Was: Separate Xorg packages)
On 06/26/2012 08:25 PM, Bruce Dubbs wrote: Ken Moffat wrote: On Tue, Jun 26, 2012 at 04:02:05PM -0500, Bruce Dubbs wrote: The only caveat is to consider the release plan. Right now, I'd like to release a coordinated LFS/BLFS on 1 September. That means a package freeze and an -rc1 for both about 1 August or slightly over one month from now. Someone might also want to think about some of the items at the bottom of longindex.html, Some of what is in section g ('Others') is reasonable, some doesn't necessarily belong there. Not sure if everything in the preceding Bootscripts section belongs there - maybe it all does! Certainly the subsequent 'L' and 'O' headings look as if their contents are in the wrong places. Indexing is easily broken ;) Yes, the index entries are tedious, especially for a new package with lots of items. That's one reason I think a whole month is needed. Even that may not be enough. I'm afraid a month probably won't be enough time. It has been several years since a BLFS release. I'm not even sure what year that was now, but it has been overwritten by years of only basic proof reading by people perusing the commit logs and the instruction authors themselves. The entire book needs to be reviewed for spelling, grammar, and consistency, which is a painful job that few of us will _want_ to do. In addition to a spell checker and manual reading (out loud), a decent screen reader works wonders for finding transposed words and odd phrases (and it gives those of us without disabilities a good reason to test at least part of the accessibility stack). There's a lot of things that could be done at that time like moving packages to different sections, reviewing especially the Preface and the first three chapters which have a lot of text, etc. If I'm around and not doing anything else after the package freeze, you could ping me - I suspect most editors don't really care about longindex. I tend to use it a lot, and have just been looking at it for some things I plan to (re)propose shortly. Beyond that, are we supposed to start tagging packages for 7.2 once the freeze starts ? If so, what happens if anything in LFS has to be revised during the -rc ? If we freeze LFS, then I think we can mark BLFS with an entity like lfs_72_checked; and have it display -rc1 while we are in freeze and then at release just change the definition. Actually, the original plan was to make it empty at release if we are going for a matching release version. I don't know if this holds true today. Also, in the old days (it was sooo long ago), we had waited for LFS final, then freeze. BLFS would lag behind LFS by at least a couple of months for final commits and cleanup. I don't know if we should follow the old way, but figured I'd throw it out there. I suspect that packages that must be changed during freeze would only be due to bugs or security issues, not enhancements. Generally, if a package is major.minor.patch and only the patch level changes, updating does not affect other packages. Personally, I'd suggest branching at package freeze time, and back porting changes from head. I realize that is extra work for whoever the RM is, but the one thing I really do not like to see is somebody sitting on a big commit for a month or better because of a package freeze for release. That is really frustrating from a peon's POV! We seem to have a pretty good editorial team now, maybe a little rough around the edges, but for the most part we are tolerable of each other and work pretty well together. I'd like to see some of the (not so) new additions stick around for a while. :-) One thing to note is that my release proposal is just that, a proposal. I'd like to get some agreement this is the right thing to do and the approach is reasonable. See minor points above. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Separate Xorg packages (Was: Xorg-7.7-1 upgrade)
On 06/26/2012 04:02 PM, Bruce Dubbs wrote: DJ Lucas wrote: On 06/19/2012 08:18 PM, Fernando de Oliveira wrote: Embedding the md5 and wget files in the book seems like an OK idea to me. The only caveat is to consider the release plan. Right now, I'd like to release a coordinated LFS/BLFS on 1 September. That means a package freeze and an -rc1 for both about 1 August or slightly over one month from now. Would such a change be reasonable by then? With the existing pages, it'd take about 10 minutes. What I want to do, however, is to separate the packages but leave the group instructions (or at very least see what I can make it look like). I've obviously stalled on that for a little bit, but I'll get back to it over the weekend. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Separate Xorg packages (Was: Xorg-7.7-1 upgrade)
On 06/22/2012 06:08 PM, Armin K. wrote: On 06/23/2012 12:46 AM, Fernando de Oliveira wrote: Yuck! I don't believe that world will end in December. OT Warning! LOL! Sorry, just being silly. But yeah, I don't really get the supposed Mayan calendar correlation. I figure the translators just screwed up somewhere along the lines and it actually ends in another 140 years or so, at a logical end point for a long calendar given the belief system. I like the idea. So, an individual package being upgraded, the cat .. EOF. sounds good, easier for book upgrade patches, if I understand correctly. I am looking forward to see the proto page. Yes, that is the point of putting the wget lists in the book directly. Great. I've said that I'm going to split Drivers into individual packages in Python/Perl Modules or Additional Xorg Drivers style (All of them on one page, but still seperated - since not all of them are really needed for one machine). Looking at this thread, I could also make instructions for generating wget and md5sum lists at the beginning of the instructions. I plan to start that in the first week of July since I have some exams to do right now and I don't have any time for that. I am willing to volunteer to what you think I could do. If I take a page from xml, make changes and post a diff this is what you would need? Cheers, Yeah, just clone the repo, edit xml and post svn diff output as an email attachment. Check if it validates before sending it. Someone will look at your changes and possibly merge them if they seem apropriate. See http://www.linuxfromscratch.org/blfs/edguide/ for some basic information regarding book editing. Actually, in this particular case, I'd suggest creating new pages as almost nothing of the old will be reused. I'm going to play around with a few ideas over the weekend and see if I can come up with something aesthetically pleasing...something a little more compact than say the various modules pages as there will be quite a few more packages per page. Maybe abuse bridgehead-renderas a little more. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Separate Xorg packages (Was: Xorg-7.7-1 upgrade)
On 06/22/2012 06:08 PM, Armin K. wrote: Great. I've said that I'm going to split Drivers into individual packages in Python/Perl Modules or Additional Xorg Drivers style (All of them on one page, but still seperated - since not all of them are really needed for one machine). Looking at this thread, I could also make instructions for generating wget and md5sum lists at the beginning of the instructions. I plan to start that in the first week of July since I have some exams to do right now and I don't have any time for that. Forgot to add that the drivers might be fine without the wget lists, as most people will want only the 5 drivers, but maybe better to keep it for consistency's sake. IDK, guess we'll just see how it turns out. Of course, without a wget list for drivers, the additional drivers page can be merged with the main page. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] Separate Xorg packages (Was: Xorg-7.7-1 upgrade)
On 06/19/2012 08:18 PM, Fernando de Oliveira wrote: Since we had a discussion about upgrading Xorg, I stopped doing some upgrades. Other than the few new splits, I only had to upgrade 8 packages: -libX11-1.4.99.901.tar.bz2 +libX11-1.5.0.tar.bz2 -libXaw-1.0.10.tar.bz2 +libXaw-1.0.11.tar.bz2 -libXft-2.3.0.tar.bz2 +libXft-2.3.1.tar.bz2 -libXi-1.6.0.tar.bz2+libXi-1.6.1.tar.bz2 -xbitmaps-1.1.0.tar.bz2 +xbitmaps-1.1.1.tar.bz2 -xinput-1.5.99.1.tar.bz2+xinput-1.6.0.tar.bz2 -xmodmap-1.0.6.tar.bz2 +xmodmap-1.0.7.tar.bz2 -xorg-server-1.12.1 +xorg-server-1.12.2 With much respect, IMHO, as Xorg is not anymore monolithic, i.e., is modular, it would be good keeping doing upgrades, for example, once a month, as several distros do upgrades continuously. I'm not sure about once per month, but it should happen periodically...but I'm going to take it a bit further (see below). If we stick with the format in the book now, I'd suggest the usual security, feature, and bugfix updates as they always have been done (when an editor (usually me) deems it necessary), in addition to a scheduled quarterly update. This would give us ~6-8 scheduled releases between katamari releases, but, on to the new subject line... Unlike other packages in BLFS, I've unintentionally limited Xorg updates over the past several years to release updates (katamari), security updates (few), and the occasional feature group updates (as above) due to the incrementing release number that is tacked on to the wget lists. This is usually when somebody chimes in with the separate the packages argument, and where I shoot it down citing the amount of work it would take and the user experience. I always give the thumbs up for educational and packaging value, but down it goes when it comes to a user's experience as we go from something like 35 sets of instructions, to around 270 sets of instructions from start to finish (not to mention the editor workload involved). Well, I'm not going to shoot it down tonight, but rather suggest a compromise, one that I've hinted at before. What I envision for that compromise is that we continue with the logical groups of packages on one page (I really do like the additional drivers page that was added, but it could be simplified even more for the group of packages pages), with instructions to create the wget and md5sums lists with the cat file EOF syntax, instructions to edit those files based on the information below (package information and purpose), and the completed loop instructions on the page. This makes individual package updates easier as we are not assigning an arbitrary build number which requires new wget and md5 file lists to be generated, each package's purpose (or lack thereof) can be explained, and we cater to the packagers without giving up our simplified build of ~35 sets of instructions (which will continue to be necessary for the average user). With the above goals stated, if we are going to do it, then it needs to be done right the first time. We can stick with what we have now and take our time to attack the interdependent explanations of the packages. I plan to do a -2 release about a month after Meas/libdrm is completed to give the upstream packagers time to incorporate the changes in their respective packages after general availability. I'm open to suggestions on the explanations here, but I suspect it will come in a totally foreign dependency order (actually reversed for the most part) due to the way that Xorg was separated (Ex: This package contains protocol headers for libXabc, libXxyz, and the server components X, Y, and Z.). Comments or suggestions would really be appreciated here. The pages could be written initially without the package descriptions/explanations to get most of the grunt work done by whoever might be interested in helping. Have any of you already started on documenting any of this on the side? Is there a dependency matrix someplace that I'm unaware of? I'll plan to do a quick mockup of the proto page (hopefully this weekend, following any info gleaned in this thread) as an example and a template of sorts, but I definitely need volunteers. I assure you that if left to do it alone, you can expect completion sometime well after the end of the world in December! ;-) -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xorg-7.7-1 Update
On 06/12/2012 04:26 AM, Andrew Benton wrote: On Tue, 12 Jun 2012 03:40:08 +0100 DJ Lucasd...@linuxfromscratch.org wrote: Update is complete, save for the index entries in the app, driver, font, lib, and proto sections. I ran out of time for them (I may still get it tonight, if not, Friday, but not much has changed). Otherwise, it's done. Let me know if you find any problems. Hey DJ, what's happening with the Xorg section? It looks like you've added separate pages for some things (like libXp) but not others (like libX11)? It seems a bit random at the moment, is there a plan? Personally I'd like to see all the xorg packages have their own pages (possibly organised into subsections like proto, app, lib, driver and util). It seems as though you've started to break some of the xorg packages out onto their own pages. Is this something that will continue? Andy No, the only separate packages that are those that are not part of the katamari, not from x.org, or those few from the x.org folks that must be separated for proper build order to still use the groups of packages. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] openjdk test report
On 06/11/2012 01:49 PM, Bruce Dubbs wrote: Fernando de Oliveira wrote: On 08-06-2012 03:21, Bruce Dubbs wrote: I was able to build and test openjdk. The results are OK, but I'm disappointed in the number of issues that are not addressed upstream. I used: #!/bin/bash export DISPLAY=:20 Xvfb :20 -screen 0 1x1x24 -ac echo $! Xvfb.pid make jtregcheck -k 2jdktest-errs |tee jdktest kill -9 `cat Xvfb.pid` unset DISPLAY rm -f Xvfb.pid Don't you and/or DJ think these lines could be in the page? Yes, I think they are useful, but I was waiting for DJ to have some time to discuss. Yeah, but 1x1 session is not adequate for testing things like borders, window decorations, and transparent backgrounds, so I left it out for now. There were several lines in jdktest-errs like WARNING: Path does not exist as file or directory: Makefile:205: Extraneous text after `ifeq' directive xxx uses or overrides a deprecated API Note: Some input files use unchecked or unsafe operations. I just don't think a package as important as java should have those types of errors/warnings in their test suite. Completely agree with this and the disappointment you mentioned in the first line above. Since this was orifinally done by Sun, the tests probably assume a shell program different from bash. That may explain some of the problems. Personally, I wouldn't release anything that causes any warnings. If other environments cause warnings, then that is worth checking out and either fixing the code or fixing the test. With the patch I've just added, it stabilizes it a bit more, but there are still unexplained test failures if not in an absolutely pristine environment. I unfortunately don't have a better way to explain it in the book. All of the crypto and SSL failures are due to tests that have not been updated for newer NSS (padding failures). The JDK code has long ago been updated, but these tests are from a very old version of the full JTReg (as explained in the book). A 1x1 Xvfb session doesn't cut it either (see above), hence why I only added a mention of it instead of the actual instructions. As it is now, I am passing all remaining tests on i686 in TWM with the xset commands Bruce posted earlier. As I understand it, hacking up JTReg to work in the limited environment expected (no junit - and alot in there to assist) was a chore to begin with, and even more so with a newer version, so we simply have to wait or fix them ourselves and send upstream. IcedTea bugs #913 and #914 are to fix the NSS related failures, which will eliminate ~40 of the 80 (currently expected) failures, and are still slotted for the 2.2 release cycle (presumably icedtea-2.2.1). -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] openjdk test report
On 06/08/2012 01:21 AM, Bruce Dubbs wrote: I was able to build and test openjdk. The results are OK, but I'm disappointed in the number of issues that are not addressed upstream. I used: #!/bin/bash export DISPLAY=:20 Xvfb :20 -screen 0 1x1x24 -ac echo $! Xvfb.pid make jtregcheck -k 2jdktest-errs |tee jdktest kill -9 `cat Xvfb.pid` unset DISPLAY rm -f Xvfb.pid There were several lines in jdktest-errs like WARNING: Path does not exist as file or directory: Makefile:205: Extraneous text after `ifeq' directive xxx uses or overrides a deprecated API Note: Some input files use unchecked or unsafe operations. I just don't think a package as important as java should have those types of errors/warnings in their test suite. The tests themselves were: --- jtreg console summary for hotspot --- Test results: passed: 154 --- jtreg console summary for jdk --- Test results: passed: 4,052; failed: 33; error: 2 --- jtreg console summary for langtools --- Test results: passed: 1,938 So here is 35 errors out of about 7000 tests. Workable, but not impressive. Tests like com/oracle/security/ucrypto/TestAES.java should not be screen or disk sensitive. It doesn't really say why it failed, but just that it threw an exception. Sorry, I thought I had responded earlier to this one Unfortunately, JTReg is a monster. You'll have to edit icedtea-2.2/test/jtreg/com/sun/javatest/TestResult.java and set DEFAULT_MAX_OUTPUT_SIZE above 10 to get the full output for that test (I just added a zero). For that particular one, it is a padding error (known issue). -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] Xorg-7.7-1 Update
Update is complete, save for the index entries in the app, driver, font, lib, and proto sections. I ran out of time for them (I may still get it tonight, if not, Friday, but not much has changed). Otherwise, it's done. Let me know if you find any problems. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] Old files
After running across the SVN server page, I gave the server a minor workout tonight... cd ~/LFS/BLFS/BOOK for file in `find . -name *.xml | sed '/\.svn/d' | sed '/stylesheets/d'` do svn info $file svninfolist.txt done for year in 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 do echo ## Last updated in $year old-files.txt grep -B 11 -A 1 Last Changed Date: $year svninfolist.txt | grep ^Path | sed 's@Path: @@' old-files.txt echo old-files.txt done The results are interesting, very good, not unexpected! I've omitted the *795* pages that have been updated over the past 5 months! Great work guys! dj [ BOOK ]$ cat old-files.txt ## Last updated in 2002 ## Last updated in 2003 ## Last updated in 2004 ## Last updated in 2005 postlfs/security/syslog.xml postlfs/security/nessus.xml ## Last updated in 2006 ## Last updated in 2007 networking/netprogs/othernetprogs.xml appendices/creat-comm.xml book/dedication.xml book/errata.xml archive/obsolete/nautilus-media.xml archive/obsolete/gal.xml xincludes/X11R6_symlink.xml xincludes/scrollkeeper-dir.xml xincludes/kde-apidocs.xml xincludes/kde-sysconfdir.xml xincludes/lib-config.xml xincludes/use-unzip.xml postlfs/filesystems/ext3.xml postlfs/config/skel.xml postlfs/config/logon.xml postlfs/config/random.xml postlfs/config/vimrc.xml postlfs/config/etcshells.xml postlfs/config/inputrc.xml introduction/welcome/acknowledgments.xml introduction/welcome/mirrors.xml introduction/welcome/wiki.xml introduction/welcome/maillists.xml introduction/welcome/conventions.xml introduction/welcome/newsserver.xml introduction/welcome/packages.xml introduction/important/bootscripts.xml introduction/important/position.xml introduction/important/pkgmgt.xml introduction/important/patches.xml ## Last updated in 2008 appendices/mit-lic.xml postlfs/config/compressdoc.xml introduction/welcome/version.xml introduction/important/beyond.xml ## Last updated in 2009 networking/netprogs/openssh-client.xml general/prog/other-tools.xml book/whoread.xml archive/kde/devel/kdewebdev.xml archive/kde/devel/kdesdk.xml postlfs/config/profile.xml introduction/welcome/askhelp.xml introduction/important/building-notes.xml ## Last updated in 2010 gnome/gnome-intro.xml networking/textweb/textweb.xml general/general.xml archive/kde/devel/devel.xml archive/obsolete/gail.xml archive/obsolete/gnome-keyring-manager.xml archive/obsolete/ggv.xml archive/obsolete/gnome-vfs-monikers.xml archive/obsolete/eel.xml archive/obsolete/gnome-audio.xml archive/obsolete/gnopernicus.xml xincludes/gtk-doc-rebuild.xml x/x.xml pst/pst.xml pst/scanning/scanning.xml pst/sgml/sgml.xml introduction/important/important.xml introduction/introduction.xml server/databases/databases.xml server/server.xml ## Last updated in 2011 networking/netlibs/libpcap.xml networking/netprogs/rpcbind.xml networking/netprogs/netfs.xml networking/netutils/nmap.xml networking/netutils/traceroute.xml networking/connect/ppp.xml networking/connect/connect.xml networking/textweb/links.xml networking/networking.xml appendices/glossary.xml general/prog/svnserver.xml general/prog/gc.xml general/prog/librep.xml general/prog/slang.xml general/genlib/talloc.xml general/sysutils/sysstat.xml general/sysutils/gpm.xml general/sysutils/eject.xml general/sysutils/fcron.xml general/graphlib/aalib.xml general/graphlib/jasper.xml general/genutils/rxvt-unicode.xml general/genutils/rep-gtk.xml archive/kde/add/kdegames.xml archive/kde/add/kdeadmin.xml archive/kde/add/kdemultimedia.xml archive/kde/kde-intro.xml archive/kde/core/core.xml archive/kde/core/config.xml archive/kde/core/pre-install-config.xml archive/obsolete/gpdf.xml postlfs/config/bootdisk.xml postlfs/editors/joe.xml postlfs/security/tripwire.xml x/installing/xorg7.xml x/wm/sawfish.xml x/wm/fluxbox.xml pst/sgml/openjade.xml pst/printing/lprng.xml ## Last updated in 2012 Snip 795 udpated pages -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Fwd: [ANNOUNCE] X11R7.7
On 06/07/2012 03:16 AM, Armin K. wrote: Yay, finally! Excellent! Want to update the book? :-) If not, I'll try and get to it on Monday. The only changes I had planned for it are adding a note about katamari and why our version number is what it is for the full distribution and then separating printproto, libXp, and xinit from the wget lists since they are not part of the katamari. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] OpenJDK Test failures
On 06/02/2012 09:32 PM, DJ Lucas wrote: Okay, so digging into the test results on the new OpenJDK/IcedTea, the following jdk test failures can be ignored while testing on icedtea-2.2: Just append the following to icedtea-2.2/test/jtreg/excludelist.jdk.jtx to actually skip the tests Snip In addition to the above, I've found that RedHat is using Xvfb to run the display tests... export DISPLAY=:20 http://pkgs.fedoraproject.org/gitweb/?p=java-1.7.0-openjdk.git;a=blob;f=java-1.7.0-openjdk.spec;h=9bfd2949fb14ac75590148a3eb064b90166081a3;hb=f2f1a7cb522137f2078b9a75388efa7c53fce8df#l918Xvfb :20 -screen 0 1x1x24 -ac echo $! Xvfb.pid make jtregcheck -k http://pkgs.fedoraproject.org/gitweb/?p=java-1.7.0-openjdk.git;a=blob;f=java-1.7.0-openjdk.spec;h=9bfd2949fb14ac75590148a3eb064b90166081a3;hb=f2f1a7cb522137f2078b9a75388efa7c53fce8df#l921kill -9 `cat Xvfb.pid` http://pkgs.fedoraproject.org/gitweb/?p=java-1.7.0-openjdk.git;a=blob;f=java-1.7.0-openjdk.spec;h=9bfd2949fb14ac75590148a3eb064b90166081a3;hb=f2f1a7cb522137f2078b9a75388efa7c53fce8df#l922unset DISPLAY http://pkgs.fedoraproject.org/gitweb/?p=java-1.7.0-openjdk.git;a=blob;f=java-1.7.0-openjdk.spec;h=9bfd2949fb14ac75590148a3eb064b90166081a3;hb=f2f1a7cb522137f2078b9a75388efa7c53fce8df#l923rm -f Xvfb.pid ...that handles the screensaver and screen blank settings, and gets us away from any WM discrepancies and user interaction. I'm not so sure about 1x1 but if it works, great. I'm going to give that a shot with the tests removed for WM compat put back and see where we are. Hopefully I'll have the book updated later this afternoon with whatever results I have. I have come to the conclusion that the test suite is just too fragile to be run on a system that is doing anything else, though it is much improved over the ~180 failures in the IT6 version. My plan for the book is for any test that I've restored that still results in failure to be removed again, as well as the NSS dependent tests that seem to still be broken, and leave the current note with an estimated 60 failures since my results are fairly consistent on both supported arches. With the other 17 removed, I'm down to zero failures on both if I do nothing, but I'm also in a very limited, and dedicated environment. Test time outs occur while copying the resultant j2sdk out of the source tree for instance. While Xvfb does give some of that back, I'm not convinced that the suite itself is ready for a perfect run by an end user, there are just too many environmental factors to consider (nor would I expect a user to adhere to such a limiting set of expectations). -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] OpenJDK Test failures
On 06/06/2012 11:09 AM, DJ Lucas wrote: On 06/02/2012 09:32 PM, DJ Lucas wrote: Okay, so digging into the test results on the new OpenJDK/IcedTea, the following jdk test failures can be ignored while testing on icedtea-2.2: Just append the following to icedtea-2.2/test/jtreg/excludelist.jdk.jtx to actually skip the tests Snip In addition to the above, I've found that RedHat is using Xvfb to run the display tests... export DISPLAY=:20 http://pkgs.fedoraproject.org/gitweb/?p=java-1.7.0-openjdk.git;a=blob;f=java-1.7.0-openjdk.spec;h=9bfd2949fb14ac75590148a3eb064b90166081a3;hb=f2f1a7cb522137f2078b9a75388efa7c53fce8df#l918Xvfb :20 -screen 0 1x1x24 -ac echo $! Xvfb.pid make jtregcheck -k http://pkgs.fedoraproject.org/gitweb/?p=java-1.7.0-openjdk.git;a=blob;f=java-1.7.0-openjdk.spec;h=9bfd2949fb14ac75590148a3eb064b90166081a3;hb=f2f1a7cb522137f2078b9a75388efa7c53fce8df#l921kill -9 `cat Xvfb.pid` http://pkgs.fedoraproject.org/gitweb/?p=java-1.7.0-openjdk.git;a=blob;f=java-1.7.0-openjdk.spec;h=9bfd2949fb14ac75590148a3eb064b90166081a3;hb=f2f1a7cb522137f2078b9a75388efa7c53fce8df#l922unset DISPLAY http://pkgs.fedoraproject.org/gitweb/?p=java-1.7.0-openjdk.git;a=blob;f=java-1.7.0-openjdk.spec;h=9bfd2949fb14ac75590148a3eb064b90166081a3;hb=f2f1a7cb522137f2078b9a75388efa7c53fce8df#l923rm -f Xvfb.pid Oops... That should read export DISPLAY=:20 Xvfb :20 -screen 0 1x1x24 -ac echo $! Xvfb.pid make jtregcheck -k kill -9 `cat Xvfb.pid` unset DISPLAY rm -f Xvfb.pid -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] OpenJade requires Perl4::Corelibs
FYI, getopts is gone, so we may see a bit more of this. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] OpenJade requires Perl4::Corelibs
On 06/05/2012 04:33 AM, Armin K. wrote: On 06/05/2012 11:20 AM, DJ Lucas wrote: FYI, getopts is gone, so we may see a bit more of this. -- DJ Lucas I have used this patch to build OpenJade with perl 5.16.0 ... And since the perl script is only used for some locale related files generation, I think it should be enough. There is also patch from BLFS combined with this one. LOL. In my defense, It is 4:40 AM here! But yeah, I guess that putting back one fairly simple function is a little easier than installing several of them globally. :-) -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] OpenJDK and libreoffice-3.5.4.2 failures (Was: OpenJDK Test failures)
On 06/04/2012 01:15 PM, Fernando de Oliveira wrote: 1.7 is not a supported Java bytecode version! Looks like this has been fixed upstream, but may require a backport. When was the current version of LibreOffice released? You can try to get rid of the bytecode check in configure.ac, and then do ./autogen.sh and see if it works, but if it was released before February, then you will need the hsqldb changes also mentioned here. http://comments.gmane.org/gmane.comp.documentfoundation.libreoffice.devel/24011 -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Fwd: [ANNOUNCE] libX11 1.5.0
On 06/02/2012 04:49 AM, Armin K. wrote: At last. You can start upgrading Xorg section now. -- Not quite. There are still two outstanding bugs https://bugs.freedesktop.org/showdependencytree.cgi?id=47255hide_resolved=0 -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] OpenJDK Test failures
javax/swing/JTableHeader/6889007/bug6889007.java javax/swing/JTextArea/4697612/bug4697612.java javax/swing/JTree/6263446/bug6263446.java javax/swing/JTree/6505523/bug6505523.java javax/swing/JViewport/7107099/bug7107099.java javax/swing/plaf/basic/BasicHTML/4251579/bug4251579.java javax/swing/PopupFactory/6276087/NonOpaquePopupMenuTest.java javax/swing/Security/6657138/ComponentTest.java javax/swing/SwingUtilities/4917669/bug4917669.java javax/swing/text/CSSBorder/6796710/bug6796710.java javax/swing/text/DefaultEditorKit/4278839/bug4278839.java # Still using LCMS1 (replacement LCMS2 tests need backports from java 8 tree) sun/java2d/cmm/ColorConvertOp/MTSafetyTest.java sun/java2d/cmm/ProfileOp/ReadWriteProfileTest.java # Invalid test case of initializer (constant) sun/net/idn/TestStringPrep.java # Fails to compile sun/security/ec/TestEC.java ### Likewise, the one langtools test failure can be ignored: Append to icedtea-2.2/test/jtreg/excludelist.langtools.jtx to exclude it. ### # Fails to compile tools/javac/processing/6499119/ClassProcessor.java ### The above gets me down to only a few failures, 17 actually. Obviously, some failures I could not simply explain away, but the mass majority of them do seem to be easily explained (invalid tests for one reason or another). I am unsure of the reason for the NSS failures. These same tests used to fail when building without NSS, but with NSS they shouldn't (this may also prove yet to be a classpath error). Perhaps they simply haven't caught up. I did see something about a wrap failure that might explain the ucrypto tests, and possibly some other fixes for some of the remaining failures. If these upstream fixes check out after testing, they'll be included in a yet to be determined patch (individual tests have to be patched in the source tree after expansion of the project tarballs). x86_64 bin package will be up in a little bit, and I'll try and get an i686 build done on Monday or Tuesday (and book updated at that time). http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK-1.7.0.4/ -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] icedtea notes
On 05/23/2012 12:02 PM, Bruce Dubbs wrote: Bruce Dubbs wrote: The second time through the build/test not going through ssh seems to have a lot fewer errors but I've probably still got a couple of hours to go. I'll give a final result tomorrow. OK, here are some results: Sources directory 195M Build directory: 6.8G (at one time grew to at least 7.2G) /opt/icedtea-bin 423M /opt/OpenJDK-1.7.0.3 422M rhino1_7R3 17M /usr/share/java 2.5M Build time (minutes) 74:35 ( 44 SBU) Test time (minutes) 216:59 (129 SBU) Internally in the log I found: -- Build times -- Target all_product_build Start 2012-05-22 21:16:55 End 2012-05-22 22:09:15 00:06:56 corba 00:07:29 hotspot 00:00:18 jaxp 00:01:06 jaxws 00:35:52 jdk 00:00:39 langtools 00:52:20 TOTAL -- Build times -- Target all_product_build Start 2012-05-22 22:09:22 End 2012-05-22 22:30:12 00:01:38 corba 00:07:10 hotspot 00:00:20 jaxp 00:00:24 jaxws 00:10:41 jdk 00:00:37 langtools 00:20:50 TOTAL -- Build times -- Target all_product_build Start 2012-05-22 22:30:51 End 2012-05-22 22:33:01 00:00:02 corba 00:00:07 hotspot 00:00:03 jaxp 00:00:07 jaxws 00:01:46 jdk 00:00:05 langtools 00:02:10 TOTAL -- Build times -- Target all_product_build Start 2012-05-22 22:33:03 End 2012-05-22 22:35:24 00:00:04 corba 00:00:11 hotspot 00:00:03 jaxp 00:00:05 jaxws 00:01:54 jdk 00:00:04 langtools 00:02:21 TOTAL Test results: passed: 144 Report written to test/hotspot/JTreport/html/report.html Test results: passed: 3,959; failed: 136; error: 10 Report written to test/jdk/JTreport/html/report.html Test results: passed: 1,920; failed: 1 Report written to test/langtools/JTreport/html/report.html === The number of hotspot and langtools failures are the same as DJ's, but I got 136 failures in the jdk where DJ only got 78. In particular, I got 28 failures in com/sun/jdi/ where DJ got none. Hmm...JDI is the debug interface. I don't have an immediate explanation for that. Do you have traditional debugging tools available in the environment? They should not have an effect, but at the time I tested, I did not have them available. I also gor 29 failures in java/awt/ where DJ got 12. This could be due to the window manager. I always sit and watch it through the tests. There were a couple where my login window was to blame when it was trying to iconify and restore windows. Also, root or unprivileged user? My results were obtained as an unprivileged user, using TWM on a fairly minimal system (just enough to meet required and recommended deps). On a side note, 2.2 is scheduled for release next Wednesday (5/30). I don't anticipate any changes in the options or dependencies. 2.2 will basically be the same as the Oracle 7u4 with some of the security updates that are slotted for 7u5, along with the typical build fixes for newer system software and the closed parts of the software replaced by free software (easily arguable as _better_ replacements (pulse for the old sgi audio for example)). Basically, just disregard the upstream fixes patch and build as before. Existing 2.1 builds should work fine as a bootstrap compiler, however, we'll be pulling the downloads from git again (make download if you have wget installed, the all target will also get them for you in one step). I don't know if these differences is due to different HW (e.g. cheap video), installed packages, or are even due to kernel configuration. For instance I got an error in NoUpdateUponShow where DJ did not: JavaTest Message: Test threw exception: sun.awt.SunToolkit$OperationTimedOut: 10001 I did take some pains to ensure the screensaver did not kick in. == The open question is how we should address the test failures in the book. Unfortunately, until I hear back from the other maintainers on their test suite results, I really have nothing to go on. I know one of the core ITW developers had mentioned that he was working on a public repo for test results in private message, but as I understand it, he has about as much free time as I do lately. :-) Perhaps I should just install Fedora or Debian and build from their respective sources to make a meaningful comparison. In the grand scheme of things, the numbers are small in comparison to the number of tests, and not that different from what is in the book now which was deemed OK previously (by me using comparison with the other distros)...at least once the mystery is solved about the JDI tests. Maybe we should consider dropping the -samevm flag as one failure could conceivably cause a whole group of tests to fail if the running environment gets trashed. The downside to that is that a new VM is created for each test, which has a rather obvious effect on the wall clock. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed
Re: [blfs-dev] Observations about firefox and xulrunner
On 05/19/2012 04:20 AM, Andrew Benton wrote: On Sat, 19 May 2012 01:01:28 +0100 Bruce Dubbsbruce.du...@gmail.com wrote: For FF, the command tar -xvf firefox-build-dir/dist/firefox-12.0.en-US.linux-$(uname -m).tar.bz2 -C /usr/lib/firefox-12.0 --strip-components=1 hung on me. I already had a /usr/lib/firefox-12.0 directory with items installed from previous builds, but deleting all the files there allowed the instruction ot complete. We may want to do something like: rm -rf /usr/lib/firefox-12.0 mkdir /usr/lib/firefox-12.0 Fair enough, it never occurred to me that people may be reinstalling Firefox. The instructions on the firefox page says: If you have linked against an already installed Xulrunner, as the root user: make -C firefox-build-dir install ... My problem was that I didn't have a firefox-build-dir directory Then you must have altered your mozconfig or make -f client.mk failed. Yes, despite the lack of --enable-application=firefox in the mozconfig file, I still had this directory and the instructions worked as expected. I believe it is defined in one of the deeper config.mk files...it's been a while since I played in the combined source tree. I think we may need to clarify that make -f client.mk is still required but the options # ac_add_options --with-system-libxul # ac_add_options --with-libxul-sdk=/usr/lib/xulrunner-devel-12.0 will bypass most of the long build. Yes, maybe simply provide the full build SBU and a note that this will reduce the build to about 0.3 SBU. The page says Compile Firefox by issuing the following commands: (which include make -f client.mk) and in the mozconfig it says: # The rest of these options have no effect if you're # building against an already installed xulrunner: I don't see why anyone would want to install xulrunner. I don't think the Firefox page should mention xulrunner. It is complicated enough to describe the options without xulrunner, adding a useless xulrunner into the mix doesn't help. Just because you don't see a use for it, doesn't mean the rest of us don't. I am perfectly fine with only Firefox providing our dev tools (In fact, I'd even prefer it at this point), but it does require a fair bit of work to make a *complete* Firefox build (something the book hasn't done since around 2008, maybe earlier) and even then it was a bit messy (see the old OpenOffice system mozilla patches for an example). If you want XUL gone from the book, good luck, but it is possible to replace it's individual functionality with Firefox, and that includes the development libs and the XUL pc files, or at least some method for building plug-ins (so we can hack up the few dependent packages to use firefox-plugin.pc or whatever is needed). Alternately, we just continue to build Firefox against XUL with the option to build stand-alone for those who don't need it. Again, if you can get an --enable-application=firefox build to produce everything that we need, then great! I'm all for it, and may be able to help out after I finish up Java. The only test cases left in the book that I am sure of are OpenJDK and IcedTea-Web, but I believe LibreOffice can still use it too. It would be a shame if we lost mail-merge functionality with thunerbird/seamonkey (and the browser plug-in was actually nice to have on occasion). One popular external example is gecco-media-player, and of course, there are literally thousands of others that I personally happen to find useless, most of which provide working binaries. Also, what is the status of the merged projects now, IOW is TB/SM still building a hacked up XUL for the MozAB or is it all in one again? It would be really nice if we could get down to a ~60MB XUL install and two tiny installs for Firefox and Thunderbird (or somewhere there abouts). FYI, my Firefox build size is 420MB and my install size is 3MB. If TB can build against an installed XUL, then maybe SM could too...any arguments for removing XUL would certainly be killed if that were the case. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] [blfs-book] r10221 - trunk/BOOK/xsoft/graphweb
Bruce Dubbs bruce.du...@gmail.com wrote: a...@linuxfromscratch.org wrote: Author: andy Date: 2012-05-19 16:21:51 -0600 (Sat, 19 May 2012) New Revision: 10221 Modified: trunk/BOOK/xsoft/graphweb/firefox.xml Log: Neither Seamonkey or Thunderbird can use xulrunner Really? I haven't tried but Wikipedia says: All XUL-based applications like Mozilla Firefox, Mozilla Thunderbird, Songbird, Flickr Uploadr, SeaMonkey, Conkeror, Sunbird, Miro, Joost, and TomTom Home 2.0 run on XULRunner. I haven't validated that, but I'll investigate some more. -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page -- This message has been scanned for viruses and dangerous content, and is believed to be clean. TB and SM used to require many changes to xpcom for mozab, not sure this is the case any more, I haven't tried since 4.0 or so. It was supposed to be one code base and they were getting close at that time. -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Observations about firefox and xulrunner
On 05/18/2012 06:41 PM, Bruce Dubbs wrote: When I started FF, I looked at Help-About. Since I was using twm, I couldn't close that window! There is no close button for 'About' and twm doesn't have a window close function that I could find. This isn't a BLFS issue, but I thought it was worthy of comment. Yes, the old Ctrl+Shift+W went away it seems. Keyboard shortcuts for special windows such as downloads and bookmarks still toggle correctly though. Wonder why they did that. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] OpenJDK downloads
On 05/16/2012 10:52 PM, Fernando de Oliveira wrote: On 16-05-2012 22:22, Fernando de Oliveira wrote: On 14-05-2012 20:26, DJ Lucas wrote: On 05/14/2012 10:38 AM, Fernando de Oliveira wrote: --- Em seg, 14/5/12, DJ Lucas escreveu: De: DJ Lucas Assunto: [blfs-dev] OpenJDK downloads Para: BLFS Development List Data: Segunda-feira, 14 de Maio de 2012, 2:16 The makefile will download the needed files as part of the build process (alternately you can use the download make target). I will try one of these in my next build. Think It sould be this way in the book, like LibreOffice, so, in the next version, book's instructions could be used just changing the version. I believe that the versions in Anduin will be useless then. I have not built on i686 yet. Help would be appreciated here, but I'll get to it in a few days if nobody steps up. I've included a tar.xz extract of the Fedora i686 RPM to use as an initial bootstrap compiler. If anybody would like to create the i686 binary, please do a bootstrap build using the Fedora JDK, followed by a bootstrap using the previously created one, and then tar it up with the format of the x86_64 file, after having run the test suite against the final. I'm currently looking into the few remaining test suite failures, but I suspect a number of them are related to running in twm (windows not automatically anchored in awt tests). I had a fatal failure in make, due to the fact that oracle jdk was installed and found by configure. Moved out the /opt/jdk* (link and versioned dir), and still failed. Next, linked /opt/jdk to icedtea-bin, and next failure: no cpio. so, cpio is a requirement. Had noticed this requirement months ago also in book's version. Phase1: using Fedora-bin Phase2: using New-bin I have done phase 1 START - Building OpenJDK-1.7.0.3-i686-bin - Qua Mai 16 14:41:02 BRT 2012 END - Building OpenJDK-1.7.0.3-i686-bin - Qua Mai 16 18:32:12 BRT 2012 real231m10.016s user156m43.051s sys 36m36.799s For make -j1: 66m22.809s For make -k jtregcheck: 164m15.324s For phase2, after make, the tests are running, now. Please, keep on reading below. Files are available here: Bin builds: http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/OpenJDK-1.7.0.3-x86_64-bin.tar.xz or http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/java-1.7.0-openjdk-1.7.0.3-i686-fedora-18.tar.xz Source files: http://icedtea.classpath.org/download/source/icedtea-2.1.tar.gz http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/corba.tar.gz http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/hotspot.tar.gz http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/jaxp.tar.gz http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/jaxws.tar.gz http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/jdk.tar.gz http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/langtools.tar.gz http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/openjdk.tar.gz ftp://ftp.mozilla.org/pub/mozilla.org/js/rhino1_7R3.zip Patches: http://www.linuxfromscratch.org/patches/downloads/icedtea/icedtea-2.1-upstream_fixes-2.patch http://www.linuxfromscratch.org/patches/downloads/icedtea/icedtea-2.1-fixed_paths-1.patch http://www.linuxfromscratch.org/patches/downloads/icedtea/icedtea-2.1-add_cacerts-1.patch Instructions: Extract the binary installation, and move it to /opt/icedtea-bin Extract Rhino and put the js.jar and js-14.jar files into /usr/share/java Extract the icedtea-2.1 tarball as would normally be done. link the other tar.gz downloads into the source tree: ln -s ../openjdk.tar.gz . ln -s corba.tar.gz . ln -s jaxp.tar.gz . ln -s jaxws.tar.gz . ln -s jdk.tar.gz . ln -s langtools.tar.gz . ln -s hotspot.tar.gz . Of course, ln -s ../ everywhere, above, right? Apply the patches and prep for build: patch -Np1 -i ../icedtea-2.1-upstream_fixes-2.patch patch -Np1 -i ../icedtea-2.1-fixed_paths-1.patch patch -Np1 -i ../icedtea-2.1-add_cacerts-1.patch ./autogen.sh Finally, build and install the thing: ./configure --with-jdk-home=/opt/icedtea-bin \ --enable-nss \ --enable-pulse-java Here, failure, as I do not have pulse installed, so, another requirement, and --enable-pulse-java was dropped. Having read troubles from others with puse, which I do no need untill now, after finished current build, will make a copy of this machine to install pulse and rebuild. This will be done after I install xulrunner and build IcedTea plugin. make Here, race failure, with make -j4. Used -j1. Same behavior in phase 1 and 2. make -k jtregcheck For phase1: $ grep Test results /home/fernando/Downloads/blfs/OpenJDK-1.7.0.3-i686-bin-with-Fedora-2012.05.16.log Test results: passed: 144 Test results: passed: 3,643; failed: 452; error: 10 Test results: passed: 1,920; failed: 1 Test results: passed: 144 Test results: passed: 3,643; failed: 452; error: 10 Test results: passed: 1,920
Re: [blfs-dev] OpenJDK downloads
Pierre Labastie pierre.labas...@neuf.fr wrote: Le 15/05/2012 01:26, DJ Lucas a écrit : Okay, well this is gonna be ugly compared to typical book instructions, but: x86_64 production bin file is in place. It no longer includes IcedTea-Web. I have not built on i686 yet. Help would be appreciated here, but I'll get to it in a few days if nobody steps up. I've included a tar.xz extract of the Fedora i686 RPM to use as an initial bootstrap compiler. If anybody would like to create the i686 binary, please do a bootstrap build using the Fedora JDK, followed by a bootstrap using the previously created one, and then tar it up with the format of the x86_64 file, after having run the test suite against the final. I'm currently looking into the few remaining test suite failures, but I suspect a number of them are related to running in twm (windows not automatically anchored in awt tests). Files are available here: Bin builds: http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/OpenJDK-1.7.0.3-x86_64-bin.tar.xz or http://anduin.linuxfromscratch.org/files/BLFS/OpenJDK/java-1.7.0-openjdk-1.7.0.3-i686-fedora-18.tar.xz [...files and installation instructions] Thanks for the instructions. Will give it a try, but what about the deps? Is xulrunner really needed? Pierre -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page -- This message has been scanned for viruses and dangerous content, and is believed to be clean. Not any more as I don't include icedtea web...but nss should be recommended s well as pulse. Giflib, cups, and libXp are required. -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] iced tea web
On 05/14/2012 08:43 PM, DJ Lucas wrote: On 05/12/2012 12:10 AM, Bruce Dubbs wrote: Looking at iced tea web, it looks like xulrunner is a required dependency. It's looking for mozilla-plugin.pc. When I search the web, it looks like that's the only thing that installs it. Not sure though. -- Bruce We should probably backport the --with-gtk={3,2} patch that came across on 5/11, else if built against gtk-2 and used with Midori, Epiphany, or other webkit based browsers built against gtk+-3, it will fail to work. -- DJ Lucas Tested, checks OK. I'll add it when I do the new OpenJDK instructions if not done before then. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Recent build notes
On 05/15/2012 06:18 PM, Ken Moffat wrote: On Tue, May 15, 2012 at 05:44:14PM -0500, DJ Lucas wrote: Couple of things I noticed on my latest build. The DRM backend is broken in cairo with Xorg-7.7, no idea if this is the result of the export symbols patch or latest xorg. Anybody built it? (probably unused) I think my recent build was still on cairo-1.12.0 and libdrm-2.4.33. Does the backend fail to build, or fail at runtime ? Do you have to enable it specifically in the build, and is there an easy test to prove it works ? ĸen You have to specifically enable it, it fails at build time, and nothing uses it yet AFAIKhence why I didn't dig in further. -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Recent build notes
On 05/15/2012 06:57 PM, Andrew Benton wrote: On Tue, 15 May 2012 23:54:06 +0100 DJ Lucasd...@linuxfromscratch.org wrote: Couple of things I noticed on my latest build. The DRM backend is broken in cairo with Xorg-7.7, no idea if this is the result of the export symbols patch or latest xorg. Anybody built it? (probably unused) I spent an evening working on trying to build with --enable-drm and gave up. I could get past each error but after fixing about 15 or 20 bugs I gave up. It seems to be unmaintained code. I never bothered to build it before. I started down the same path, but gave up after the second mismatch figuring there would be more to fix. Ghostscript lists lcms as a dependency, it should be lcms2. It lists lcms2 as recommended and lcms1 as optional, or do you mean we should be recommending lcms1? Indeed it does. I guess I must have been looking for little-cms-2 because it is right there in front of me. Udev 182 didn't rebuild with pci-utils without a decompressed pci.ids file. Because of build order (a very minimal target), the firefox.desktop file got installed as /usr/share/applications (need to check first, and create the directory if necessary). Doh! Good point. I'll fix that tomorrow. Thanks Yeah, I suffered a major space cadet moment on that one. Funny now, but I was like Permission denied? Huh? It's 755! WTH?!?! 10 minutes of the same cycle later the light bulb comes on. Oh...Yep I'm a dumb a**! LOL -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] OpenJDK downloads
On 05/14/2012 07:21 PM, Ken Moffat wrote: On Mon, May 14, 2012 at 06:26:56PM -0500, DJ Lucas wrote: I suspect a number of them are related to running in twm (windows not automatically anchored in awt tests). I know we include twm for a minimal build of xorg, but I didn't think anybody used it for a completed system. Obviously, I'm missing something - for me, fluxbox is an almost-usable wm, twm is too painful. Normally, I use icewm-1.3 (which needs gtk+-2) - feel free to call me a wimp, but I really don't understand how you can use twm comfortably! ĸen Heh, no I'd shoot myself if subjected to it for too long. Just meeting the most minimal requirements for a Java rebuild, and for that, a decent WM was not high prioritybut will be in about 10 minutes. That's still 190 packages in (-LFS packages included but not individual xorg packages)...probably just over the halfway point. Now I have heard that somebody is actually taking the time to modernize twm, but the only evidence I've seen of it is that it still builds against Xorg-7.7. Years ago I had seen some pretty elaborate twmrc setups, but I've never taken the time. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] iced tea web
On 05/12/2012 12:10 AM, Bruce Dubbs wrote: Looking at iced tea web, it looks like xulrunner is a required dependency. It's looking for mozilla-plugin.pc. When I search the web, it looks like that's the only thing that installs it. Not sure though. -- Bruce We should probably backport the --with-gtk={3,2} patch that came across on 5/11, else if built against gtk-2 and used with Midori, Epiphany, or other webkit based browsers built against gtk+-3, it will fail to work. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] OT: twm [Was: Re: OpenJDK downloads]
On 05/14/2012 08:50 PM, Bruce Dubbs wrote: On 5/14/12, DJ Lucasd...@linuxfromscratch.org wrote: Heh, no I'd shoot myself if subjected to it for too long. Just meeting the most minimal requirements for a Java rebuild, and for that, a decent WM was not high prioritybut will be in about 10 minutes. That's still 190 packages in (-LFS packages included but not individual xorg packages)...probably just over the halfway point. Now I have heard that somebody is actually taking the time to modernize twm, but the only evidence I've seen of it is that it still builds against Xorg-7.7. Years ago I had seen some pretty elaborate twmrc setups, but I've never taken the time. I agree that twm is not a wm that you want to spend a lot of time with, but it's extreme simplicity makes it a good choice for testing after building X. -- Bruce twm *can* be made to act modern-ish! http://www.over-yonder.net/~fullermd/projects/twmrc 6 work spaces, a full applications menu (right click like flux/black), sane windowing behavior, etc:-) -- DJ -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] XUL Runner and Firefox
On 05/13/2012 01:49 PM, bdu...@linuxfromscratch.org wrote: Author: bdubbs Date: 2012-05-13 12:49:25 -0600 (Sun, 13 May 2012) New Revision: 10201 Modified: trunk/BOOK/general.ent trunk/BOOK/introduction/welcome/changelog.xml trunk/BOOK/x/lib/lib.xml trunk/BOOK/x/lib/xulrunner.xml Log: Restore xulrunner Actually, I've been thinking a bit more on this one. As a compromise, we could probably still get rid of the XUL Runner page and use only the Firefox page to first build XUL Runner, and then Firefox in a separate build directory from only the Firefox source tarball. This is the way Firefox should actually be built anyway for our purposes. Without the dev libs, a user cannot build browser extensions from source (still ignoring the fact that the build tree needs to kept around). The FF portion of the build should take 0.1 SBU. The reason it is not done that way by default is that the Mozilla devs choose not to cater to developers, but rather end users...understandable give the huge Windows target, but a minor PITA for us and a source of mild disagreement. Anybody think this is a bad compromise? -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] XUL Runner and Firefox
On 05/13/2012 02:55 PM, Andrew Benton wrote: On Sun, 13 May 2012 20:13:02 +0100 DJ Lucasd...@linuxfromscratch.org wrote: Actually, I've been thinking a bit more on this one. As a compromise, we could probably still get rid of the XUL Runner page and use only the Firefox page to first build XUL Runner, and then Firefox in a separate build directory from only the Firefox source tarball. This is the way Firefox should actually be built anyway for our purposes. Without the dev libs, a user cannot build browser extensions from source (still ignoring the fact that the build tree needs to kept around). The FF portion of the build should take 0.1 SBU. The reason it is not done that way by default is that the Mozilla devs choose not to cater to developers, but rather end users...understandable give the huge Windows target, but a minor PITA for us and a source of mild disagreement. Anybody think this is a bad compromise? Yes, I think it's stupid to make people install half a gigabyte of files just to get a browser. Firefox works fine with 31 megabytes of files. Only people who want to make the java plugin will benefit from the extra half a gigabyte. It is years since I've installed java. Andy Understood. There is an easier workaround someplace. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] XUL Runner and Firefox
On 05/13/2012 03:01 PM, Armin K. wrote: On 05/13/2012 09:55 PM, Andrew Benton wrote: On Sun, 13 May 2012 20:13:02 +0100 DJ Lucasd...@linuxfromscratch.org wrote: Actually, I've been thinking a bit more on this one. As a compromise, we could probably still get rid of the XUL Runner page and use only the Firefox page to first build XUL Runner, and then Firefox in a separate build directory from only the Firefox source tarball. This is the way Firefox should actually be built anyway for our purposes. Without the dev libs, a user cannot build browser extensions from source (still ignoring the fact that the build tree needs to kept around). The FF portion of the build should take 0.1 SBU. The reason it is not done that way by default is that the Mozilla devs choose not to cater to developers, but rather end users...understandable give the huge Windows target, but a minor PITA for us and a source of mild disagreement. Anybody think this is a bad compromise? Yes, I think it's stupid to make people install half a gigabyte of files just to get a browser. Firefox works fine with 31 megabytes of files. Only people who want to make the java plugin will benefit from the extra half a gigabyte. It is years since I've installed java. Andy Half a gigabyte? 29M /usr/lib/xulrunner 3.6M /usr/lib/xulrunner-devel 24M /usr/include/xulrunner 6.8M /usr/share/idl/xulrunner 6.8M /usr/lib/firefox About 70 MB when using Firefox with system Xulrunner. Think he is referring to the build size...actually...IDK. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] XUL Runner and Firefox
On 05/13/2012 03:16 PM, Bruce Dubbs wrote: Armin K. wrote: On 05/13/2012 09:55 PM, Andrew Benton wrote: On Sun, 13 May 2012 20:13:02 +0100 DJ Lucasd...@linuxfromscratch.org wrote: Actually, I've been thinking a bit more on this one. As a compromise, we could probably still get rid of the XUL Runner page and use only the Firefox page to first build XUL Runner, and then Firefox in a separate build directory from only the Firefox source tarball. This is the way Firefox should actually be built anyway for our purposes. Without the dev libs, a user cannot build browser extensions from source (still ignoring the fact that the build tree needs to kept around). The FF portion of the build should take 0.1 SBU. The reason it is not done that way by default is that the Mozilla devs choose not to cater to developers, but rather end users...understandable give the huge Windows target, but a minor PITA for us and a source of mild disagreement. Anybody think this is a bad compromise? Yes, I think it's stupid to make people install half a gigabyte of files just to get a browser. Firefox works fine with 31 megabytes of files. Only people who want to make the java plugin will benefit from the extra half a gigabyte. It is years since I've installed java. Andy Half a gigabyte? 29M /usr/lib/xulrunner 3.6M /usr/lib/xulrunner-devel 24M /usr/include/xulrunner 6.8M /usr/share/idl/xulrunner 6.8M /usr/lib/firefox About 70 MB when using Firefox with system Xulrunner. The build size is pretty big and /usr/lib/xulrunner-devel-12.0 is pretty big for me. $ find /usr -type d -name xulrunner\* -exec du -sh {} \; 2.4G/usr/src/firefox/mozilla-release/xulrunner-build-dir 32M /usr/src/firefox/mozilla-release/xulrunner-build-dir/dist/xulrunner 1.4M/usr/src/firefox/mozilla-release/xulrunner-build-dir/xulrunner 2.2M/usr/src/firefox/mozilla-release/xulrunner 45M /usr/src/xulrunner 6.8M/usr/share/idl/xulrunner-12.0 27M /usr/include/xulrunner-12.0 475M/usr/lib/xulrunner-devel-12.0 32M /usr/lib/xulrunner-12.0 Most of this is in /usr/lib/xulrunner-devel-12.0/sdk/lib/libxul.so -rwxrwxr-x 1 root root 463M May 13 13:45 libxul.so, but stripping it gets it down to 24M. Oh. Nix the -g flag in the XUL runner build and install everything. Firefox should be no different than any other package in BLFS. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] OpenJDK downloads
I'm getting prepared to commit OpenJDK (IcedTea-2.1) to the book, however, I wanted to ask for opinions on how to do the downloads. The IcedTea tarball only contains the build environment. The makefile will download the needed files as part of the build process (alternately you can use the download make target). Unfortunately, the files are named with git sums and renamed to something useful in the source tree. In all other BLFS instructions, packages can be downloaded prior to building. So options are, download 7 files with names such as 7ceda3124828.tar.gz, host on anduin with the files already renamed, or just let the makefile handle it with a note saying that if you want to build offline, you'll need to download these files before hand? -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] xulrunner is required for icedtea source build
Pierre Labastie pierre.labas...@neuf.fr wrote: Le 08/05/2012 22:22, Pierre Labastie a écrit : The Firefox and Seamonkey pages have a paragraph that mentions how you can install all the development libs. If they need changing to accommodate icedtea please let me know. Andy Thanks for pointing me to that. I'll test tomorrow. Pierre I am sorry to say that one is hard. After building d-bus, cups, giflib and firefox (and all deps), I obtain this error: -- /usr/bin/make -f /sources/icedtea6/icedtea6-1.9.7/openjdk-ecj/hotspot/make/linux /Makefile checks make[7]: Entering directory `/sources/icedtea6/icedtea6-1.9.7/openjdk.build-ecj/hotspot/outputdir' 2 echo *** This OS is not supported: `uname -a`; exit 1; *** This OS is not supported: Linux turbolivirt 3.3.4-lfs-1 #1 SMP Tue May 8 00:34:11 CEST 2012 x86_64 GNU/Linux make[7]: *** [check_os_version] Error 1 make[7]: Leaving directory `/sources/icedtea6/icedtea6-1.9.7/openjdk.build-ecj/hotspot/outputdir' make[6]: *** [linux_amd64_compiler2/debug] Error 2 make[6]: Leaving directory `/sources/icedtea6/icedtea6-1.9.7/openjdk.build-ecj/hotspot/outputdir' make[5]: *** [generic_build2] Error 2 make[5]: Leaving directory `/sources/icedtea6/icedtea6-1.9.7/openjdk-ecj/hotspot/make' make[4]: *** [product] Error 2 make[4]: Leaving directory `/sources/icedtea6/icedtea6-1.9.7/openjdk-ecj/hotspot/make' make[3]: *** [hotspot-build] Error 2 make[3]: Leaving directory `/sources/icedtea6/icedtea6-1.9.7/openjdk-ecj' make[2]: *** [build_product_image] Error 2 make[2]: Leaving directory `/sources/icedtea6/icedtea6-1.9.7/openjdk-ecj' make[1]: *** [stamps/icedtea-ecj.stamp] Error 2 make[1]: Leaving directory `/sources/icedtea6/icedtea6-1.9.7' - It does not look like an error with firefox development libs. And actually I do not understand what is going on. The directory /sources/icedtea6/icedtea6-1.9.7/openjdk-ecj/hotspot/make has disappeared... No time to investigate. I think I'll go back to old jdk. Pierre -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page -- This message has been scanned for viruses and dangerous content, and is believed to be clean. Hopefully I'll have an initial build of OpenJDK 7 (IcedTea-2.1) this weekend for amd64. -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xulrunner
On 05/01/2012 12:31 PM, Bruce Dubbs wrote: I think we are going to remove openjdk, but keep iced-tea. I've discussed with DJ, but he has the most experience with it. I think there is some confusion on this topic in general, so I'll try to explain. OpenJDK is the GPL'd version of Oracle's JDK with all of the closed parts removed. IcedTea is a project that replaces the closed parts with GPL software. The IcedTea download is just a build harness that provides a complete JDK with all of the functionality of the Oracle one using only GPL software, which is followed by IcedTea-Web to provide a GPL implementation of the browser plug-in and Java Web Start. I should mention that the OpenJDK license is GPLv3 with Classpath Exception. The Classpath Exception is akin to LGPL for jar files, you can release closed software using the classes in OpenJDK, but any changes must still be given back to the community. As far as future plans, unless somebody sees some value in the closed version, I do want Oracle's JDK removed from the book, and I plan to rename the IcedTea page OpenJDK as IcedTea is just the build tool. I hope that clears up any confusion. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Fwd: [ANNOUNCE] xf86-video-intel 2.19.0
On 05/01/2012 07:22 AM, Armin K. wrote: Original Message Subject: [ANNOUNCE] xf86-video-intel 2.19.0 Date: Tue, 1 May 2012 11:47:06 +0100 From: Chris Wilsonch...@chris-wilson.co.uk Reply-To: x...@lists.freedesktop.org To: xorg-annou...@lists.freedesktop.org CC: intel-...@lists.freedesktop.org, x...@lists.freedesktop.org I sent this out over the weekend and never saw it arrive, so presuming the announcement was lost in transit... Release 2.19.0 (2012-04-29) === * Remove broken acceleration for rendering glyphs directly upon the destination pixmap, exposed by cairo-1.12.0 (and coincidentally fix another Pixmap leak upon fallback handling). -- I just got this mail. This is just part of the changelog indicating that the cairo bug was in the driver itself. Perhaps I'm mistaken, but I thought this was seen with ATI and NVidia drivers too. :-/ Also, Cairo is up to 1.12.2 (current patch applies and I already symlinked it in the repo). -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] [blfs-book] r10058 - in trunk/BOOK: . introduction/welcome postlfs/editors x/lib
On 05/02/2012 11:32 AM, Bruce Dubbs wrote: I can see both sides of the discussion here. If someone works on a section of the book for quite a while, I think it's natural to develop a sense of 'ownership' of that work. I don't think anyone would object if another comes in and corrects a misspelling or md5sum or somthing like that. We should be able to touch up each other's work, but things that would appear to be style are not quite so simple. And sometimes that ownership sticks for a long time! ;-) The easy way out is to ask in advance and offer to make the change. The key is trying to avoid surprises. I agree in principal, but I'll go on record and say that if anybody wants to do something with packages that are historically mineplease speak up and then get to work! I'm thankful that Andy and Fernando teamed up to do LibreOffice. I hate that the book can suffer in any way just because I can't dedicate the time I'd like to. For instance, right now, the previously mentioned OpenJDK changes are probably a month out yet. I wouldn't be the least bit hurt if somebody else wanted to do it faster. I'd certainly appreciate consulting, but it is not necessary. Same thing for Xorg. 7.7 should be out very soon, and I think it could definitely benefit from some style other than that outdated layout I added several years ago. One suggestion, for instance, is to separate out the parts only used for testing the X server (twm, xinit, mesa demos). These aren't part of the katamari nor is Xp which was required for Java. Another, if done correctly, the packages could even be separated out instead of arbitrarily assigning a build number to it (the -2 in 7.6-2). Even though I've recently argued against it, there is some merit to separating the xorg packages. Now, I will continue to argue against adding 200+ pages to the book, but having thought about it a little more, something with the organization of the Python modules page for each section might make a nice compromise. I suggest that we still use the wget and md5 files and a loop, but also provide descriptions and dependencies on the page. I don't have the time or desire to do that, but as it is right now, there is absolutely no documentation on what might or might not be needed. I suspect that many of us just build the whole enchilada, which completely negates the point of separating the packages in the first place. We might as well just write a Makefile with a World target. The fonts are another thing. We probably need only the font-util package and an assortment of TTF fonts now that the legacy packages are gone. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Gnome 3.4 status
), missing icons (icons and xdg symlinks/paths), missing artwork (pixmap symlink), automount issues (? udisks), etc. All kinda fuzzy now as I've been building in /usr for a while now. Also, as discussed earlier in the FHS-3.0 notes, a shared libexec dir would be nice now that it is explicitly allowed (use defaults so there are no surprises). I'll be following right on your heels come Friday (hopefully). Oh, also, on GDM (which I said I'd do in a previous response to Ken and then got side tracked), we need a new bootscript or to have it started in inittab, and the PAM config file should be updated with what I posted earlier (also in response to Ken IIRC) using system values where possible. I'll post anything else I notice as I come across it. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Planning
On 04/19/2012 08:24 AM, Fernando de Oliveira wrote: On 18-04-2012 23:52, Bruce Dubbs wrote: Ken, Andy, Ragnar, Armin, Fernando, Could you please let me know what you are planning to work on for the next few weeks. I've been working tickets and we are now down to 30! No particular plans. I have tried and failed to build openjdk and web plugin (-: . I will try again. DJ, you are doing this. Please, would you think that sending me some instructions or draft, I could help you with this? Absolutely. You can build with gcj, but setting up a full JDK mockup is difficult. Early on, OpenJDK 6 wouldn't build it, but I'm told that is resolved now (haven't bothered trying it). It is much, much easier to bootstrap with another OpenJDK-7 installation. For instance, I grabbed the Fedora RPMs and used those for my initial 7 bootstrap, and from that point on I use my own. I'll upload the current patches, but that is not complete. We are going to want just about everything that has come across distro-pkg-dev since Feb 15th for the 2.1 branch. IcedTea-2.1 includes all fixes through the proprietary 7u3, and some security fixes slotted for 7u4, but there have been some since. In addition there are some other bugs. The glib 2.32's added gthread requirement, and gcc-4.7 fixes, rhino was not actually delivered in the 2.1 release, etc. All of those changes are available upstream. Also, the test suite might require some massaging or explanation. I don't know exactly how much. I had 60+ failures last run and a display lockup when I unlocked my screen, but that was the problem: I left the screen locked, forgetting about the interactive tests. I've just put it on the back burner until I've imported all of the upstream patches that are needed. Anyway, I'll submit the add-cacerts patch (Bruce's mydate() to handle 2038+ expiration dates on 32bit) and fixed-paths patches in just a few moments. As far as configure flags, I'm using '--with-jdk-home=/opt/icedtea --enable-nss --enable-pulse-java' along with the two patches mentioned above and builds are successful with what is currently in the book (also, no need to fake gcj home with ecj, and no more need for xerces and xalan). -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] Java Common page
Anybody have an objection to a Java Components page such as is done with Perl or Python as opposed to 10 new package pages? Items such as ECJ, JUnit, JAI, Jakrata, Net-Commons, Xerces, XalanJ, etc. could be placed on a single page such as is done with Python and Perl modules. Candidates include those required by Apache Ant and IcedTea (others?). This list will cover all of what is needed in the book by those two packages currently so that instructions are not reproduced on multiple pages. These of course could be separated if desired, but the majority simply copy a file and create a symlink...potentially add a .sh file for CLASSPATH (as is the case with JUnit since it is installed outside of the global directory). http://archive.apache.org/dist/jakarta/regexp/binaries/jakarta-regexp-1.5.zip http://archive.apache.org/dist/jakarta/oro/jakarta-oro-2.0.8.zip http://www.carfab.com/apachesoftware//commons/net/binaries/commons-net-1.4.1.zip http://java.sun.com/products/java-media/jai/downloads/download-1_1_3.html http://download.java.net/media/jai/builds/release/1_1_3/jai-1_1_3-lib-linux-amd64.jar.zip http://download.java.net/media/jai/builds/release/1_1_3/jai-1_1_3-lib-linux-i586.jar.zip http://ftp.osuosl.org/pub/eclipse/eclipse/downloads/drops/S-3.8M6-201203141800/ecj-3.8M6.jar ftp://ftp.mozilla.org/pub/mozilla.org/js/rhino1_7R3.zip (is it possible to be built as part of js?) http://apache.osuosl.org/xml/xalan-j/xalan-j_2_7_1-bin.zip http://apache.osuosl.org/xerces/j/Xerces-J-bin.2.11.0.zip -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] KMS, Gallium, kernel firmware, etc.
On 04/15/2012 05:33 AM, Andrew Benton wrote: I think we should add a section about kernel config to one of the xorg pages, maybe xorg-server? It's xorg-server that needs KMS enabled. Well, not exactly. The framebuffer console is very nice to work in as well, and that was the reason for the question. Placement would quite obviously be limited to the xorg section if not for the FB. If it were limited to Xorg, the xorg-server page makes more sense to me than the generic Xorg configuration pages as was done in the past. Maybe there was a typo in one of the names? No, the kernel make would barf with a file not found error. Did you use CEDAR_*.bin globing in your kernel config? No, like you said, the build would have stopped. It was exactly as you have below, that's why I'm still unsure as to why it is working now. The big thing that bothers me is that I found out about compiling in the microcode/firmware into the kernel from the Gentoo and Arch wikis, not from the ATI page at freedesktop.org's wiki. There seem to be 3 CEDAR bin files: andy@doughnut:~$ ls /lib/firmware/radeon/CE* /lib/firmware/radeon/CEDAR_me.bin /lib/firmware/radeon/CEDAR_rlc.bin /lib/firmware/radeon/CEDAR_pfp.bin CONFIG_EXTRA_FIRMWARE=radeon/CEDAR_me.bin radeon/CEDAR_rlc.bin radeon/CEDAR_pfp.bin CONFIG_EXTRA_FIRMWARE_DIR=/lib/firmware Yes, I think that's the way to go. Install the linux firmware: http://git.kernel.org/?p=linux/kernel/git/dwmw2/linux-firmware.git;a=summary If you put it in /lib/firmware so you don't have to keep copying it into the kernel source every time you recompile. Start by compiling the kernel as a module, boot and then: dmesg | grep radeon or dmesg | grep firmware Then recompile with the firmware compiled into the kernel so you get KMS, framebuffer, penguins and nice fonts right from the start. Okay, so that has been tested then. I will revisit this a little later and do the same. Intel and Nvidia users have the same issues, or is this limited to Radeon devices? -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Java Common page
On 04/15/2012 03:30 AM, Armin K. wrote: On 04/15/2012 08:39 AM, DJ Lucas wrote: Anybody have an objection to a Java Components page such as is done with Perl or Python as opposed to 10 new package pages? Items such as ECJ, JUnit, JAI, Jakrata, Net-Commons, Xerces, XalanJ, etc. could be placed on a single page such as is done with Python and Perl modules. Candidates include those required by Apache Ant and IcedTea (others?). This list will cover all of what is needed in the book by those two packages currently so that instructions are not reproduced on multiple pages. These of course could be separated if desired, but the majority simply copy a file and create a symlink...potentially add a .sh file for CLASSPATH (as is the case with JUnit since it is installed outside of the global directory). http://archive.apache.org/dist/jakarta/regexp/binaries/jakarta-regexp-1.5.zip http://archive.apache.org/dist/jakarta/oro/jakarta-oro-2.0.8.zip http://www.carfab.com/apachesoftware//commons/net/binaries/commons-net-1.4.1.zip http://java.sun.com/products/java-media/jai/downloads/download-1_1_3.html http://download.java.net/media/jai/builds/release/1_1_3/jai-1_1_3-lib-linux-amd64.jar.zip http://download.java.net/media/jai/builds/release/1_1_3/jai-1_1_3-lib-linux-i586.jar.zip http://ftp.osuosl.org/pub/eclipse/eclipse/downloads/drops/S-3.8M6-201203141800/ecj-3.8M6.jar ftp://ftp.mozilla.org/pub/mozilla.org/js/rhino1_7R3.zip (is it possible to be built as part of js?) http://apache.osuosl.org/xml/xalan-j/xalan-j_2_7_1-bin.zip http://apache.osuosl.org/xerces/j/Xerces-J-bin.2.11.0.zip -- DJ Lucas java-avalon-framework-4.2.0 java-batik-1.7 java-commons-io-1.4 java-commons-logging-1.1.1 java-xerces-2.11.0 java-xml-commons-external-1.4.01 java-xml-commons-resolver-1.2 java-xmlgraphics-common-1.4 I have installed those with Sun's JDK and FOP runs without any errors. I think xerces is not necesary at all, but it can help. And it would be a good idea to add them on one page, I tought of that too. JUnit and JAI need to be separated as they are not machine independent, which means that they do not belong in /usr/share, but I'm wondering if that shouldn't be /usr/lib/java anyway. They are not technically libraries, but they are treated similar to libraries by the Classpath exception (similar to GNU LGPL exceptions) and used in a similar manor. I'm going to put back the jar search function in icedtea.sh too and eliminate the dir search (to make it automatic as is implied). If the Oracle JDK is to remain in the book, these changes should be done there too (I'll take care of it when I do IcedTea). Any comments, objections, or suggestions for the above? -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xorg-7.6-2 upgrade
On 04/13/2012 10:30 PM, Fernando de Oliveira wrote: I have upgraded Xorg in five generation LFS machines. The sequence and instrucions are the same as in the book, only many versions change: util-macros-1.17.0 xorg-proto makedepend-1.0.4 libXau-1.0.7 libXdmcp-1.1.1 libpthread-stubs-0.3 xcb-proto-1.7 libxcb-1.8.1 xorg-lib xcb-util-0.3.8 MesaLib-8.0.2 xorg-app xcursor-themes-1.0.3 xorg-font xkeyboard-config-2.5.1 xorg-server-1.12.0.902 xorg-driver xterm-278 Many packages in the wget lists are from march/2012. All went well, but for the drivers, I had some that could not be built and I did not try to fix them to build, leaving for running X as the final test if I needed them or not. The failed build drivers: $ grep error sshfs/blfs/xc-10.08.2011/driver-7.6-famo.wget #xf86-video-apm-1.2.3.tar.bz2 # error #xf86-video-rendition-4.2.4.tar.bz2 # error #xf86-video-s3virge-1.10.4.tar.bz2 # error #xf86-video-sisusb-0.9.4.tar.bz2 # error #xf86-video-tseng-1.2.4.tar.bz2 # error #xf86-video-xgi-1.6.0.tar.bz2 # error Attached, the wget lists. Xorg is a fairly easy update. If you really want to do an internal version bump to 7.6-3, go ahead. I'm not sure about xorg-server-1.12.1 (haven't looked at it), but if x.org is preparing for a distribution version bump anyway, I'd personally just let it hang for a couple of weeks. Issues like the drivers above, and new libx11 will be fixed by the time a distribution release is made. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Xorg Build Instructions
Armin K. kre...@email.com wrote: Hello there. I've been looking at Xorg build instructions lately. Introduction page includes little script that automates building of every package in the section. Every package section have only configure (a bit different at some stages), make and make install I'd suggest to merge that script into every xorg section, but with introducing a risk that all packages must be configured and built as root, unless someone can think of some other way. I originally did this and it was not liked because of hand holding. I'm happy to have it scripted, but the book should not be recommending behavior that is generally frowned upon. Recommend sudo and they can remove it at their own risk. -- DJ -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Remove lesstif and xpdf?
On 04/09/2012 11:31 PM, Bruce Dubbs wrote: I've been looking at lesstif and think it's more effort than it's worth to maintain. The maintenance is very spotty. It builds OK, but the install procedure is questionable. I wouldn't think anyone would really want to use the Motif window manager. 0.95.2 2009-05-27 19,728 downloads 0.95.0 2006-06-10 54,515 downloads 0.94.42005-04-01 30,104 downloads I had still used it until very recently for one really old app that continued to work, but I don't any longer, and it's probably a safe bet that I'm in the minority. Probably need to check as I'm not positive, but the proprietary JDK used to depend on motif (remember the really ugly java windows?). I only recall this as we had to fix linking with libXm.a very early on in the SCSL/JRL source from around the 6 release until about 6_u6 or so (best guess as to when it was fixed upstream). I'd personally be happy to see the Oracle JDK go completely. I have verified that IcedTea-1.x (OpenJDK6) is without the motif dependency and I'll be updating the existing IcedTea-1 and adding IcedTea-2 (OpenJDK7) shortly. Coincidently, does anybody have a very recent 32-bit LFS/BLFS with Gnome laying around? It's not a huge deal to build it now that my scripts are up to date, but I'm still all for time savings. Anyway, from a fairly complete system, I have: dj [ /opt/icedtea/demo/jvmti ]$ for dir in /lib /usr/lib /opt/icedtea/jre/lib/amd64; do for lib in $dir/*.so; do ldd $lib | grep libXm\.so echo $lib done done loader cannot load itself ldd: exited with unknown exit code (127) ldd: warning: you do not have execution permission for `/lib/libcap.so' ldd: warning: you do not have execution permission for `/usr/lib/libc.so' ldd: warning: you do not have execution permission for `/usr/lib/libcurses.so' ldd: warning: you do not have execution permission for `/usr/lib/libcursesw.so' libXm.so.2 = /usr/lib/libXm.so.2 (0x7fbb38acd000) /usr/lib/libDtPrint.so ldd: warning: you do not have execution permission for `/usr/lib/libform.so' ldd: warning: you do not have execution permission for `/usr/lib/libgcc_s.so' ldd: warning: you do not have execution permission for `/usr/lib/libicudata.so' ldd: warning: you do not have execution permission for `/usr/lib/libmenu.so' libXm.so.2 = /usr/lib/libXm.so.2 (0x7f81699ae000) /usr/lib/libMrm.so ldd: warning: you do not have execution permission for `/usr/lib/libncurses.so' ldd: warning: you do not have execution permission for `/usr/lib/libpanel.so' ldd: warning: you do not have execution permission for `/usr/lib/libpthread.so' ldd: warning: you do not have execution permission for `/usr/lib/libpytalloc-util.so' ldd: warning: you do not have execution permission for `/usr/lib/libtalloc.so' libXm.so.2 = /usr/lib/libXm.so.2 (0x7f28c8cb6000) /usr/lib/libUil.so ldd: warning: you do not have execution permission for `/usr/lib/preloadable_libintl.so' dj [ /opt/icedtea/demo/jvmti ]$ grep libDtPrint.so /var/log/llog/* /var/log/llog/lesstif-0.95.2.llog:/usr/lib/libDtPrint.so19 root:root 777 libDtPrint.so.1.0.0 /var/log/llog/lesstif-0.95.2.llog:/usr/lib/libDtPrint.so.1 19 root:root 777 libDtPrint.so.1.0.0 /var/log/llog/lesstif-0.95.2.llog:/usr/lib/libDtPrint.so.1.0.0 124506 root:root 755 dj [ /opt/icedtea/demo/jvmti ]$ grep libMrm.so /var/log/llog/* /var/log/llog/lesstif-0.95.2.llog:/usr/lib/libMrm.so15 root:root 777 libMrm.so.2.0.1 /var/log/llog/lesstif-0.95.2.llog:/usr/lib/libMrm.so.2 15 root:root 777 libMrm.so.2.0.1 /var/log/llog/lesstif-0.95.2.llog:/usr/lib/libMrm.so.2.0.1 217216 root:root 755 dj [ /opt/icedtea/demo/jvmti ]$ grep libUil.so /var/log/llog/* /var/log/llog/lesstif-0.95.2.llog:/usr/lib/libUil.so15 root:root 777 libUil.so.2.0.1 /var/log/llog/lesstif-0.95.2.llog:/usr/lib/libUil.so.2 15 root:root 777 libUil.so.2.0.1 /var/log/llog/lesstif-0.95.2.llog:/usr/lib/libUil.so.2.0.1 164502 root:root 755 dj [ /opt/icedtea/demo/jvmti ]$ Nothing is using it now days. As an aside, should we be setting 755 perms on these? Seems we had this discussion once before, but the warnings above, while harmless, are kind of annoying when doing an exercise like the above. xpdf has lesstif as a mandatory dependency. There are better pdf viewers. The last update was 08/16/11 and the book requires an update from 3.0.2 to 3.0.3. xpdf-3.00.tar.bz2 Oct 29 2005 xpdf-3.01.tar.gz Apr 6 2006 xpdf-3.02.tar.gz Jul 27 2007 xpdf-3.03.tar.gz Aug 16 2011 Any problems with removing these? Probably not needed any longer. Wait a bit and see if anybody expresses interest, but I'd expect it to be done. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs
Re: [blfs-dev] What happened to the x-setup.html page?
On 04/07/2012 05:13 PM, Ken Moffat wrote: O Sat, Apr 07, 2012 at 04:49:18PM -0500, Bruce Dubbs wrote: Ken Moffat wrote: Which device is that? /dev/dri/card0? I'd think the kernel/udev would create it at boot. I don't have an ATI graphics card to test, but I'm interested. I'll look at the source, but I need to know what to look for. -- Bruce Seems to be. Note that it isn't specific to ati, it's for *all* cards or vm's supported by Mesa. (Sorry for the greengrocer's apostrophe - didn't want anyone to think I was referring to VMS.) And in case anyone missed what started this - even a user NOT in the video group can use the Software Rasterizer (Mesa's swrast driver), but not hardware acceleration. ken@ac4tv ~ $ls -lr /dev/dri/ total 0 crw-rw 1 root video 226, 64 Apr 7 18:29 controlD64 crw-rw 1 root video 226, 0 Apr 7 18:29 card0 ken@ac4tv ~ $ Revised version of the proposed changes are now at http://www.linuxfromscratch.org/~ken/tmp/blfs-book-DRI/general/clutter.html (well, that's the easiest way in) with a link to the main change in the Note at the top of that page. The link renders nicely this time, so I daren't change it, too much voodoo in docbook 8) The important stuff is the changes in xorg-config, changes in clutter (apart from the link), metacity, mutter, totem are really just minor rewording now that I understand the hw/sw acceleration a little more. Looks good to me! -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Open = Libre Office
Andrew Benton a...@benton.eu.com wrote: Hello, Based on the information Fernando is spoon feeding me I think I may be close to being able to compile Libre Office. If I get it to work and run (I can't promise anything ;) I'd like to replace Open Office with Libre Office (in the book). IE, remove Open Office from the book. Would that be Ok with everyone? I had planned to do it...some day. *Please* take it! ;-) -- DJ -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] What happened to the x-setup.html page?
On 04/06/2012 07:43 PM, Ken Moffat wrote: On Fri, Apr 06, 2012 at 02:37:24PM +1000, Wayne Blaszczyk wrote: Apologies if I missed a previous discussion, but I noticed the x-setup.html page no longer exists (not linked). I was specifically looking for the following instructions: cat /etc/sysconfig/createfiles EOF /tmp/.ICE-unix dir 1777 root root EOF I gather that this is still relevant since I'm getting a warning message in my logs stating that it is not set to root. Regards, Wayne. It was commented out (from x/installing/installing.xml), svn blame shows the commented lines are from r9069 when dj updated xorg to 7.6-2. I've no idea *why* it was commented (I've already been bitten by changing the *wrong* set of instructions for something earlier this year and then wondering why nothing changed - in my sandbox ), but I guess he thought we had redundancy and overlooked that that instruction was NOT duplicated. DJ ? Ah yes, it didn't get transferred over to the single page. Will get on it in a sec. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] What happened to the x-setup.html page?
On 04/06/2012 11:14 PM, DJ Lucas wrote: On 04/06/2012 07:43 PM, Ken Moffat wrote: On Fri, Apr 06, 2012 at 02:37:24PM +1000, Wayne Blaszczyk wrote: Apologies if I missed a previous discussion, but I noticed the x-setup.html page no longer exists (not linked). I was specifically looking for the following instructions: cat /etc/sysconfig/createfiles EOF /tmp/.ICE-unix dir 1777 root root EOF I gather that this is still relevant since I'm getting a warning message in my logs stating that it is not set to root. Regards, Wayne. It was commented out (from x/installing/installing.xml), svn blame shows the commented lines are from r9069 when dj updated xorg to 7.6-2. I've no idea *why* it was commented (I've already been bitten by changing the *wrong* set of instructions for something earlier this year and then wondering why nothing changed - in my sandbox ), but I guess he thought we had redundancy and overlooked that that instruction was NOT duplicated. DJ ? Ah yes, it didn't get transferred over to the single page. Will get on it in a sec. -- DJ Lucas Actually, I wound up putting it on the server page. It really didn't read well on the configuration page as it is required, similar to the /etc/X11/xorg.conf.d directory which is also on that page (which I had previously forgotten to provide a command explanation), whereas items on the configuration page are optional. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] [resend] gdm
On 04/03/2012 10:00 PM, Bruce Dubbs wrote: Ken Moffat wrote: I've never heard of AD I read that as Active Directory. Don't knwo for sure if that's what DJ meant. -- Bruce Yes. It's been a long time, but I have managed to make linux workstations play nice in AD. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] xulrunner breakage
On 04/03/2012 09:06 PM, Ken Moffat wrote: DJ is known as an iced-tea person - I suppose you need xulrunner for that ? I did note a suggestion recently (perhaps on lwn.net) that libreoffice can now be built without java, but so far I haven't attempted that. ĸen Before the Libre organization came about, Go-oo had a method to build without Java completely, and OOo from sun had it down to just gcj/gnu classpath so that you could have a completely open office suite, but in both cases, the wizards were broken or non-existent. Now, I'm not sure about the latest versions of IcedTea and its dependencies now days. I _think_ just NSS for the crypto parts and SpiderMonkey for JS now days...even the Xerces dep is gone IIUC, in favor of libxml2. I'll be getting back to that next week for x86_64, not sure when I'll get to x86. As far as rebuilding everything WRT ff/tb/xul, I wonder if that is only libcrfm or do they just regularly break API/ABI in the same minor? -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] xulrunner breakage
Andrew Benton a...@benton.eu.com wrote: On Wed, 04 Apr 2012 01:38:07 +0100 DJ Lucas d...@linuxfromscratch.org wrote: IOW, add the following at the end of the instructions (fixes my libproxy issue, and allows libmusicbrainz to install): Finally, while still the root user, add symlinks to the libraries so that other packages can link against them: ln -svf xulrunner-devel-11.0/lib/libmozalloc.so /usr/lib/libmozalloc.so ln -svf xulrunner-devel-11.0/lib/libxpcom.so /usr/lib/libxpcom.so ln -svf xulrunner-devel-11.0/lib/libxul.so /usr/lib/libxul.so Looking at the *.pc files, libxul-embedding.pc lists -lxpcomglue and libxul.pc lists -lxpcomglue_s. I think it would be simpler to do symlink them all with something like: for thing in /usr/lib/xulrunner-devel-xulrunner-version;/sdk/lib/*.{a,so} do ln -sfv ${thing#/usr/lib/} /usr${thing#*sdk} done Or leave the -L in there that I suggested you remove. :-) Sorry about that...shooting from the hip. Not necessary to have the static libs in /usr/lib other than reducing the length of the linker flags. And that means -L/usr/lib _should_ be added too (even though it is completely unnecessary). What is there now works fine, we can revisit at next update. --DJ -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] xulrunner breakage
The pc files for libxul and friends do not tell the build machinery that it is out of the library search path. The technically correct fix is to add '-rpath=${sdkdir}/lib' to the Libs: line, but this would break any chance of upgrading libxul on a live system. Better is to symlink the libs into /usr/lib as was done in the past. Andy, you've been doing all the work there, what are your thoughts? I'm more partial to the symlink method in that nothing would break on update (even though the -L flag is not needed in the pc file at that point). Same thing would need to be done if firefox, seamonkey, or thunderbird provide the only libxul. dj [ /sources ]$ ldd /usr/lib64/libproxy.so linux-vdso.so.1 = (0x7fff2e5c7000) libpthread.so.0 = /lib/libpthread.so.0 (0x7f26e65e7000) libdl.so.2 = /lib/libdl.so.2 (0x7f26e63e3000) libxul.so = not found libxpcom.so = not found libmozalloc.so = not found libplds4.so = /usr/lib/libplds4.so (0x7f26e61dd000) libplc4.so = /usr/lib/libplc4.so (0x7f26e5fd8000) libnspr4.so = /usr/lib/libnspr4.so (0x7f26e5d88000) libstdc++.so.6 = /usr/lib/libstdc++.so.6 (0x7f26e5a89000) libm.so.6 = /lib/libm.so.6 (0x7f26e5807000) libgcc_s.so.1 = /usr/lib/libgcc_s.so.1 (0x7f26e55f2000) libc.so.6 = /lib/libc.so.6 (0x7f26e5269000) /lib64/ld-linux-x86-64.so.2 (0x7f26e6a43000) -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] FHS 3.0 adaptation
On 04/03/2012 05:36 AM, Armin K. wrote: if there are only executables installed, I recommend moving them in /usr/bin or /usr/sbin (depends on how it would be run). No, that is absolutely the _wrong_ thing to do. In fact, my earlier example with gvfs was with gio (under the same umbrella). It was GDM that had issues with Console-Kit and the ck-* files (not policy-kit) as brought up elsewhere in this thread. The ck-* programs are installed into a libexec directory and should not be used by any package other than Console-Kit, but Gnome packages have broken this rule until recently (at least I hope they are using the new library in the newer version of GDM for the Console-Kit stuff). At any rate, upstream *is* making assumptions based on the default behavior of Console-Kit, GSD, and gio, and even though upstream remains blatantly ignorant in this case, the proposed change would actually avoid a slightly intrusive patch for gvfs, assuming GNOME_PREFIX=/usr (sorry /opt/gnome users). I have not verified that the problem still exists, just using context from elsewhere in this thread. Here are some other related bugs to display the problem (and also to support Armin's distaste for non /usr gnome): https://bugzilla.gnome.org/show_bug.cgi?id=543064 -- Still broken I think, tell you in a few hours https://bugzilla.gnome.org/show_bug.cgi?id=596388 -- Fixed https://bugzilla.gnome.org/show_bug.cgi?id=554140 -- Still open..no idea if still broken https://bugs.freedesktop.org/show_bug.cgi?id=18427 -- Fixed Notice my examples all revolve around Gnome? And I thought KDE was a PITA. :-) Still a Gnome user. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] [resend] gdm
On 04/03/2012 06:36 PM, Andrew Benton wrote: On Tue, 03 Apr 2012 17:07:23 +0100 Armin K.kre...@email.com wrote: Great. And then everyone told me how /etc/gnome and prefix other than /usr can work. Well, yes they can. But there is a lot of additional configuration that needs to be done, and also I still haven't found a way to make policykit rules installed somewhere else than in /usr available to the daemon. So, I am asking again. If everyone else agrees, I'd like to drop GNOME_SYSCONFDIR - lot easier than making symlinks and if possible GNOME_PREFIX or keep it with fat warning that it might break things (that is if anyone wants to test if that works, I'm not going to do that). As I've said before, I agree with you. I think GNOME_SYSCONFDIR should be /etc, GNOME_PREFIX should be /usr. Andy Add a third. See the bugs I just posted in the libexec thread. All related and avoidable by using /usr as the prefix and standard libexec dirs. -- DJ Lucas -- This message has been scanned for viruses and dangerous content, and is believed to be clean. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page