Re: Upgrading Xorg 7.2 - 7.4, black screens, evdev and all that
On Mon, Apr 13, 2009 at 5:55 AM, David Jensen djensen...@windstream.net wrote: On Mon, 13 Apr 2009 08:05:57 +0100 Jeremy Henty onepo...@starurchin.org wrote: On Sun, Apr 12, 2009 at 09:07:12PM -0500, David Jensen wrote: The blank screen may be agp mode mismatch. 7.4 tried to use 4x but my original R200 just does 2x. I set it 2x in the bios and in xorg.conf (hey, doesn't hurt to be sure). Which option is that? AGPMode? in my case: Option AGPMode 2 Yep. Newer radeon stopped trying to guess the AGP mode because it was wrong too often. If you find out what the correct mode is, send a message to x...@lists.freedesktop.org with the settings you needed and they'll add it to a quirks list in the driver. It might even be fixed in newer versions of the ati driver. You can try out 6.12.1 and see if that fixes your issues: http://xorg.freedesktop.org/releases/individual/driver/xf86-video-ati-6.12.1.tar.gz I believe it should build and run with 7.4, but the server might be too old. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Upgrading Xorg 7.2 - 7.4, black screens, evdev and all that
On Mon, Apr 13, 2009 at 8:17 AM, Jeremy Henty onepo...@starurchin.org wrote: On Mon, Apr 13, 2009 at 07:49:23AM -0700, Dan Nicholson wrote: On Mon, Apr 13, 2009 at 5:55 AM, David Jensen djensen...@windstream.net wrote: in my case: Option AGPMode 2 This makes no difference. I haven't looked at the BIOS settings yet. But is this even relevant, since my Radeon 9250 is a PCI card? Assuming the AGPMode is relevant, how do I find out the right value? I've started Xorg 7.2 and looked through the server log but nothing seems relevant. There are other values you can set for AGPMode. I think the default is 1, but you can also have 2, 4 or 8. You just have to iterate till it works. See radeon(4) for more details. Can you post the log? -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: HAL - Xorg troubles
On Fri, Apr 3, 2009 at 4:12 PM, michael lang kingo...@gmail.com wrote: My apologies for double mailing, but I have to correct myself, I used the patch on http://lists.freedesktop.org/archives/hal/2008-August/012179.html and got a whole lot more info, but the error in the end is still the same, hald cannot get dbus to convince that it is privileged to own the service org.freedesktop.Hal(even though haldaemon and root are allowed to own it according to /etc/dbus-1/system.d/hal.conf) (note, the --retain-privileges option doesn't change this either) Be careful using newer dbus releases. A security issue was found in how dbus determines privileges. It's fixed in newer releases, but it means that some configuration files (i.e., /etc/dbus-1/system.d/hal.conf) that were working are now in error. You may want to give one of the permissive releases a try to make sure this is not your issue. http://www.freedesktop.org/wiki/Software/dbus#head-ed92e8f84ae0374ae3a2e3f714c2eb0037a84868 -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: libqalculate-0.9.6 for kdeedu-4.2.1
On Sun, Mar 29, 2009 at 2:35 PM, lux-integ lux-in...@btconnect.com wrote: On Sunday 29 March 2009 04:51:28 pm Ken Moffat wrote: That INT_MAX error ought to be fixable, perhaps by http://cvs.archlinuxppc.org:7647/viewvc/Extra/extra/lib/libqalculate/libqal culate-0.9.6-gcc4.3.patch?view=markupsortby=log - that just happened to be the first link I found which took me to a list of patches for libqalculate, I don't think it's ppc-specific. I found the 'useful' url to be:- http://repos.archlinux.org/viewvc.cgi/libqalculate/repos/extra-i686/ So, not only does current kde need not-properly-released tools to build it, it also needs defunct packages. Sometimes, I start to despair. If the cln-config patch you're using is anything like the one in the same arch /libqalculate/ directory, you need to invoke the autofoo magic after applying it, before you run configure. The normal shotgun approach is to just run 'autoreconf'. On the one package where I currently use that, it wasn't sufficient and I had to follow it with libtoolize -f'. I'm led to believe that should only very rarely be necessary (in that case, I found it in a gentoo ebuild). If you haven't tried autoreconf, use clean source, apply both the patches, autoreconf, configure, make. If that is not sufficient, use clean source [critical in this case!], reapply the patches, autoreconf and then 'libtoolize -f' before configure. I applied BOTH patches *fresh sources:- A) ### running autoreconf # checking for pthread_create in -lpthread... yes ./configure: line 17615: syntax error near unexpected token `1.1.0,' ./configure: line 17615: `AC_PATH_CLN(1.1.0,' make: *** No targets specified and no makefile found. Stop. B)#running libtoolize -f### checking for correct ltmain.sh version... no configure: error: *** [Gentoo] sanity check failed! *** *** libtool.m4 and ltmain.sh have a version mismatch! *** *** (libtool.m4 = 1.5.23b, ltmain.sh = 2.2.6) *** Please run: libtoolize --copy --force if appropriate, please contact the maintainer of this package (or your distribution) for help. ## for the latter I seem to have a mismatch of libtool. (I even ran the libtoolize --copy --force (on fresh sources) with the same effect. If you're going to run autoreconf, always run autoreconf -iv. The -i is important since libtoolize won't be run and stuff like the above happens. The -v is important so that you see what autoreconf is doing. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: root=UUID or root=LABEL not working
On Mon, Feb 23, 2009 at 11:43 AM, ell sam ell@e17th.com wrote: Dan Nicholson wrote: On Mon, Feb 23, 2009 at 11:32 AM, ell sam ell@e17th.com wrote: I am having trouble booting from the kernel on my hd to load Linux from my usb hd. I have installed everything and it works using the internal hd kernel booting from grub passing the root=/dev/sdb6 to the kernel. I want to pass something other than /dev/sdb6 so that if I plug in another drive it still boots the correct one. I have tries root=UUID=my uuid numbers... and root=LABEL=LFS and root=/dev/disk/by-uuid/my uuid... and root=/dev/disk/by-label/LFS all don't work. is there something in the kernel that needs to be configured so it works. Or maybe in initramfs etc.. That's handled by an initramfs. Since the /dev/disk/by-*/* symlinks are setup by udev, it needs to be running before the kernel tries to mount the root filesystem. This can only be done in an initramfs. There's no officially supported initramfs for LFS yet. -- Dan that's true for root=/dev/disk/by-uuid/123. but the root=UUID=123ABC should work. No, I'm pretty sure that's also an initramfs feature. Basically, you check if root={UUID,LABEL}, and then poll for the symlink to be created by udev. The kernel has no feature that I know of for doing that. The reading of disk labels and UUIDs is done in userspace, AFAIK. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: HAL vs autofs vs ? - need some tips
On Sat, Jan 17, 2009 at 8:20 AM, Scott harv...@montana.com wrote: Hi, I am slogging through blfs. I *think* I have finally tracked down all of the requirements for HAL, but have some questions. Looking at the config details, I see a lot of mention of gnome, and wonder if I am doing the right thing. My ultimate system will be mainly console-based; if I get so far as getting X running, it will use a lightweight wm such as fluxbox. However, it will be essential that access to samba shares and usb plug-in devices be available transparently to the user. Can anyone point me in the right direction here? I have had great success with autofs on my old (Mandrake 7.1 - can't even remember what kernel) system with samba; of course, that kernel doesn't even recognize usb. However, an updated Mandrivel 2009 seems to have difficulties: recognizes the shares, but can't access them. (Works okay if I run a gnome app such as nautilus) I don't really know what I'm doing, and want to get LFS put together to work flawlessly, so I sure will appreciate any advice! HAL runs as root and has a Mount method that allows unprivileged users to mount devices. The automatic part comes into play when there is a service that listens for HAL events that a mountable device shows up. It then tells HAL to mount it if it believes the user is privileged. For GNOME, this role is played by gnome-volume-manager/gnome-mount or nautilus/gvfs in more recent releases. Two more generic tools that would make more sense on the console are ivman (to listen to HAL events) and pmount (to handle the mounting). I've never set this up personally, but I know it can work. On the other hand, if you're comfortable with autofs and can get a working setup, it will probably be simpler to use that. HAL can provide a much richer and more dynamic experience, but the setup can be difficult. If you want a working system today, it might be easier to use autofs. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Problems installing Xorg 7.4
On Fri, Jan 9, 2009 at 9:43 AM, William Tracy afishion...@gmail.com wrote: On Fri, Jan 9, 2009 at 12:11 AM, DJ Lucas d...@linuxfromscratch.org wrote: Most of the dropped apps are still perfectly valid, and continue to be developed. XDM is no longer being supported, which I find disappointing. Somebody is still maintaining TWM, though, which is cool. I'm pretty sure xdm is being maintained, but it's just not part of the Xorg release. All the distros have it available, so things do get fixed now and again. It probably won't receive any new features, though. http://cgit.freedesktop.org/xorg/app/xdm/log/ -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Problems installing Xorg 7.4
On Thu, Jan 8, 2009 at 2:58 AM, José Carlos Carrión j...@estudiosvirtuales.es wrote: Hello colisters: I've finished the installation of LFS 6.4 without problems. I'm using BLFS-svn-20090102 in order to completing the installation of a server. Many packages installed without problems. I'm installing Xorg 7.4 with an eye on BLFS-svn-20090102 and the other on BLFS 6.3 (for example, I've installed the xorg-server-1.5.3 package with the --with-mesa-source=dir to mesalib source modifier as BLFS 6.3 reads). My big surprise rise when I've started to config X. I'm missing many X applications which has always been there. I've double checked the BLFS-6.3 app-7.2.wget list and the BLFS-svn-20090102 app-7.4.wget list. Both of them are similar to ftp.x.org dirs ftp://ftp.x.org/pub/X11R7.2/src/app and ftp://ftp.x.org/pub/X11R7.4/src/app, but a lot of X applications are missing in X11R7.4. The X11R7.2 app. list have 92 items and the X11R7.4 list only 40.This is the whole list of missing applications in X11R7.4: bdftopcf (added by BLFS team to BLFS-svn-20090102 wget list) beforelight editres fonttosfnt fslsfonts fstobdf ico lbxproxy listres mkcfm oclock proxymngr rgb rstart scripts showfont twm (added by BLFS team to BLFS-svn-20090102 wget list) viewres xbiff xcalc xclipboard xclock (added by BLFS team to BLFS-svn-20090102 wget list) xconsole xdbedizzy xditview xdm xedit xeyes xfd xfindproxy xfontsel xfs xfsinfo xfwp xgc xinit (added by BLFS team to BLFS-svn-20090102 wget list) xkbprint xload xlogo xlsfonts xmag xman xmessage xmh xmore xphelloworld Obviously the problem is in Xorg and not in BLFS team. Anyone knows why this applications (many of them have been in X from many years ago and are very useful like xdm) have been «dropped» from X11R7.,4 official distro? I'm browsing in deep in the Xorg site but I couldn't find it. Most of them are completely deprecated. The releases are still there, though, and some are even still developed. Just look in http://xorg.freedesktop.org/releases/individual/app/ and pick any extras you want. The ones added back for BLFS (xinit/twm/etc.) are just so that startx works out of the box. bdftopcf should probably be added back upstream since building the fonts fails otherwise. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: help with libexif-gtk and cdrdao
On Wed, Dec 17, 2008 at 6:22 AM, Ken Moffat k...@linuxfromscratch.org wrote: On Tue, Dec 16, 2008 at 11:27:23PM +, b-vol wrote: Greetings, I am in a spot of bother with installing two programs on an AMD64 box (gcc-4.3.2 kernel 2.6.27.7 - 64-bit (non-multilib) build. (I've never come across libexif-gtk, and have no ideas about what is wrong. A quick search suggests distros are using it, so if nobody has any other suggestions you could try looking at the distros from http://oss-security.openwall.org/wiki/distro-patches - I think Dan gave me that link originally, It can be a bit hit-and-miss trying to find a path in some of them, but fedora (fc10) is usually a good place to start. I don't know if that will help.) Oooh, that's nice. I didn't give you that link, but I'm glad you posted it. Collects all the important information in one location instead of just floating around my head. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: help with libexif-gtk and cdrdao
On Tue, Dec 16, 2008 at 3:27 PM, b-vol lux-in...@btconnect.com wrote: Greetings, I am in a spot of bother with installing two programs on an AMD64 box (gcc-4.3.2 kernel 2.6.27.7 - 64-bit (non-multilib) build. Program1:- libexif-gtk-0.3.5. This is needed for gtkam. gtkam (and libexif-gtk) are not in the blfs book but on the cblfs site. All attempts to compile libexif-gtk have failed. I tried the sed on the cblfs website as well as debian patches I found all to no avail. I tried stuff from the CVS repository but the downloaded stuff has no configure or autgen script.The first thing odd noticable is when running the configure script one gets (after makefile generation):- ./configure: line 29105: srcdir: command not found Configuration (libexif-gtk): Source code location: Version: 0.3.5 Compiler:gcc -m64 libexif: 0.6.12 (think about upgrading) I had libexif 0.6.16 and then 0.6.17 (newely relesed) installed and I still got the nonsence. Further down I get this:- libexif-gtk is looking for the exif-mem.h header file to determine if you have newer libexif. It should be in /usr/include/libexif/exif-mem.h. It looks like it's expecting `pkg-config --cflags libexif` to return -I/usr/include/libexif. Better would be to check for libexif/exif-mem.h. That's what the cblfs sed is doing. That should take care of this problem. In file included from gtk-menu-option.c:22: gtk-menu-option.h:53: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'gtk_menu_option_get_type' gtk-menu-option.c: In function 'gtk_menu_option_destroy': gtk-menu-option.c:72: warning: implicit declaration of function 'GTK_CHECK_CAST' gtk-menu-option.c:72: warning: nested extern declaration of 'GTK_CHECK_CAST' .. In file included from gtk-menu-option.c:22: gtk-menu-option.h:53: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'gtk_menu_option_get_type' gtk-menu-option.c: In function 'gtk_menu_option_destroy': gtk-menu-option.c:72: warning: implicit declaration of function 'GTK_CHECK_CAST' gtk-menu-option.c:72: warning: nested extern declaration of 'GTK_CHECK_CAST' Could you show the command being run and any other output before this? C errors can be tricky to debug. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: help with libexif-gtk and cdrdao//udf-tools-1.0.0b3
On Wed, Dec 17, 2008 at 3:39 PM, Ken Moffat k...@linuxfromscratch.org wrote: On Wed, Dec 17, 2008 at 08:43:39PM +, b-vol wrote: #first good tidings: 1) for cdrcdao-1.2.2 the patch http://www.linuxfromscratch.org/patches/downloads/cdrdao/cdrdao-1.2.2-gcc43 worked well. great. It will get into BLFS-dev real soon now (I'm short of time at the moment, and beating my head against python/sip for kdebindings-4.1.3). 2) for libexif-gtk I will have to do some more digging. The program appears not to be well maintained. I am susprised it plays is such an important part for gtkam. If you or any other know of an alternative to gtkam please let me know. I'm not quite sure how it is used, but I see it is a front-end for gphoto2 (which doesn't impress me, I'm afraid). So, my usage is probably going to be slightly different from the way you want to work. I set up a rule for my camera in udev, an entry in /etc/fstab, then I just mount it and copy the files over (needs vfat in the kernel). Then, I use the gimp (and ufraw for raw files) to do manipulation, or sometimes I just use 'display' from ImageMagick. Not necessarily the most convenient way of looking at the pics (display will open them with pixels mapped 1:1), but then I normally have to do manipulations anyway, which is why I use the gimp. Alternatively, somebody will perhaps have a good word for gwenview (kde4 - again uses gphoto2), but in all honesty I think we've some way to go in sorting out how best to build kde4 at the moment. Maybe there are other front ends, I can certainly peopel mentioning that they use gphoto2. I've been using libgphoto2 for a long time, and it works very well with my camera on the latest release. Another pretty simple GNOME gphoto frontend is gthumb. That's the default photo app on fedora and has a similar dependency set to gtkam. I think gtkam was more of an example frontend for gphoto when it was being developed. Nowadays, there are lots of them. Two that I've used are digiview and f-spot, although they'd take longer to get built. Even picasa uses gphoto via wine, I think. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: pciaccess
On Thu, Nov 27, 2008 at 11:28 AM, Ken Moffat [EMAIL PROTECTED] wrote: On Thu, Nov 27, 2008 at 12:38:42PM -0600, Ralph Porter wrote: HELP! Compiling XORG-Server is telling me that Package requirements (pciaccess 0.8.0) is not met. Looking around the web I see the libpciaccess but its all source. What do I need to do and why is this not in the BLFS docs. Of maybe it is and I missed an install. thanks in advance rp -- It's in the book for 7.4, as of yesterday or today. If you are building 7.2, I thought it was only needed for a few of the (intel?) video drivers and a quick look at my logs shows I didn't need it for the 1.2 server. It's needed for the server and all drivers now. The PCI layer in the server was removed to instead rely on libpciaccess. All drivers that needed to be converted to libpciaccess. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: gnome-desktop-2.24 compile problem
On Tue, Nov 18, 2008 at 8:03 PM, Dennis J Perkins [EMAIL PROTECTED] wrote: I'm having trouble compiling gnome-desktop. It gives me this error message: gnome-rr.c:46: error: expected specifier-qualifier-list before 'XRRScreenResources' Has anyone else encountered this and solved it? That's from libXrandr. Either your randrproto/libXrandr combo is too old (pre-1.2) or too new, depending on what gnome-desktop wants with it. I suspect too old, though, since I understand that newer gnome was trying to make use of the RandR-1.2 API. What version of randrproto/libXrandr did you install? -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: hal start fails
On Sun, Nov 2, 2008 at 7:09 AM, Dr. Edgar Alwers [EMAIL PROTECTED] wrote: Hallo, I have three BLFS systems runing, two desktops and one laptop. HAL is installed on all three. On the desktops, hal starts normal ( hal 0.5.9.1 ) but not on the laptop: starting the HAL Daemon.[ FAIL ] . d-bus was already running. Is some bug known on HAL when running on an Laptop ? How can I get a little more debug informations than FAIL ? Just run `haldaemon --verbose=yes --daemon=no'. That will log to stdout, and you should be able to see the failure. However, your current error should be in syslog. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Another problem...
On Fri, Oct 31, 2008 at 6:38 AM, Nicolas FRANCOIS [EMAIL PROTECTED] wrote: ...with my brain, I'm quite sure. I have serious problems with devices permissions : I can read a DVD with MPlayer being root, but not being me. Same thing with gphoto2 : I can download the photos from my camera being root, but not as a normal user. My normal user belongs to disk, usb, video, haldaemon... I don't know what to add ! Some of the devices created by udev in the dev directory do seem to be owned by root, so this may be the problem. You need to look at the /dev permissions and associated udev rules. The CD/DVD devices (look at the /dev/cdrom* symlinks) are created with cdrom group permissions. I'm not sure what method you followed to install libgphoto2, but the package contains a program to create udev rules with an appropriate group (I was using the camera group). However, without the rules, I think the regular usb udev rule should be used, which sets the group to usb. Need some more details here. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Vsftpd not building on LFS/BLFS 6.3 stable system.
On Wed, Oct 29, 2008 at 7:49 AM, Zach Milton [EMAIL PROTECTED] wrote: Trying some of the newer versions of libcap does not help, as it still does not compile. Hmm, that would imply that the vsftpd build is not picking up the libcap headers for some reason. I wonder why. On the other hand, Dan's patch worked like a charm, and vsftpd compiles, installs, and starts up correctly. To be fair, I don't think the code using capset is actually used unless you have chown_uploads=YES or connect_from_port_20=YES. But I'm fairly confident that the patch is correct. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Vsftpd not building on LFS/BLFS 6.3 stable system.
On Tue, Oct 28, 2008 at 1:48 AM, Zach Milton [EMAIL PROTECTED] wrote: When I issue the 'make' command that the BLFS 6.3 stable book specifies for building vsftpd, vsftpd fails to build. The system is LFS 6.3 stable w/ BLFS 6.3 stable. All optional dependencies for the package are installed with the exception of libcap (could not build libcap due to errors). Am I doing something wrong, or is this a problem with the package? Both vsftpd and libcap have a problem where they're using the old way of making syscalls that aren't supported in newer kernel headers. Try a newer libcap release: http://www.kernel.org/pub/linux/libs/security/linux-privs/kernel-2.6/ The code where vsftpd is failing to build is where it's doing the capset syscall, which libcap would do if it was installed. I'll also attach a patch for vsftpd I made a long time ago that I'm not entirely sure is correct, but I think it is. -- Dan vsftpd-2.0.5-syscall-2.patch Description: application/mbox -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Cairo dependencies
On Fri, Oct 24, 2008 at 3:56 PM, Ken Moffat [EMAIL PROTECTED] wrote: I'm still trying to document the dependencies on my new desktop build (which works, except that evince still crashes xorg (this is with gcc-4.2.4 on x86_64)). I seem to have found a circular dependency,and I'm mighty confused. At the moment, I've built cairo-1.8.0, poppler-0.8.7, and ghostscript-8.63 in that order (because that's the order I've built the previous versions in, in the days of gcc-4.1.2 it all worked). In cairo, I see it tests for poppler and ghostscript, and it won't build the pdf and ps backends without them. It did cross my mind that this might be the cause of my evince problem, but rebuilding cairo now that poppler and gs have been installed, and rebuilding evince, didn't help - I guess that was a red herring. Are you sure it's not just skipping PDF and PS tests without those guys? Certainly, poppler uses cairo for it's rendering of PDFs. I know in the past that cairo checked for gtk+, but it was only for the testsuite. The cairo testsuite is pretty large and not shy about leaning on external pieces for it, but they shouldn't be needed for cairo itself. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: xorg-server-1.2.0 build error
On Thu, Oct 16, 2008 at 2:16 PM, Nicolas FRANCOIS [EMAIL PROTECTED] wrote: Hi. Compiling Xorg-7.2 from the last SVN book (svn-20081013), I encountered this error : ... gcc -DHAVE_XORG_CONFIG_H -DXF86PM -Wall -Wpointer-arith -Wstrict-prototypes -Wmissing-prototypes -Wmissing-declarations -Wnested-externs -fno-strict-aliasing -D_BSD_SOURCE -DHAS_FCHOWN -DHAS_STICKY_DIR_BIT -I/usr/include/freetype2 -I../../include -I../../include -I../../Xext -I../../composite -I../../damageext -I../../xfixes -I../../Xi -I../../mi -I../../miext/shadow -I../../miext/damage -I../../render -I../../randr -I../../fb -g -O2 -o Xorg -rdynamic xorg.o ../../dix/.libs/libdix.a common/libinit.a loader/libloader.a ./.libs/libosandcommon.a rac/librac.a parser/libxf86config.a dixmods/.libs/libdixmods.a ../../composite/.libs/libcomposite.a ../../mi/.libs/libmi.a ../../xfixes/.libs/libxfixes.a ../../Xext/.libs/libXextbuiltin.a ../../GL/glx/.libs/libglx.a ../../GL/mesa/.libs/libGLcore.a ../../render/.libs/librender.a ../../randr/.libs/librandr.a ../../damageext/.libs/libdamageext.a ../../miext/damage/.libs/libdamage.a ../../miext/cw/.libs/libcw.a ../../miext/shadow/.libs/libshadow.a ../../Xi/.libs/libXi.a ../../xkb/.libs/libxkb.a ../../dix/.libs/libxpstubs.a ../../os/.libs/libos.a -ldl /usr/lib/libXfont.so /usr/lib/libfreetype.so /usr/lib/libXau.so /usr/lib/libfontenc.so -lz /usr/lib/libXdmcp.so -lm dixmods/.libs/libxorgxkb.a -lrt /usr/lib/libXfont.so: undefined reference to `ft_isdigit' collect2: ld returned 1 exit status make[4]: *** [Xorg] Erreur 1 make[4]: quittant le répertoire /sources/xc/xorg-server-1.2.0/hw/xfree86 I checked in a fix for this a couple months ago. http://wiki.linuxfromscratch.org/blfs/changeset/7432 Basically, ft_isdigit is a macro from freetype that got removed. So, eventually it manifests as a symbol error. We just replace it with isdigit, which is what the macro was before anyway. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: webkit with gnome-2.24
On Wed, Oct 15, 2008 at 11:52 PM, Simon Geard [EMAIL PROTECTED] wrote: On Wed, 2008-10-15 at 16:39 +0100, Ken Moffat wrote: On Wed, Oct 15, 2008 at 04:31:59PM +0100, Ken Moffat wrote: I didn't know that the firefox code can build xulrunner, and I still don't understand that - in particular, where do you get the .pc files which epiphany and yelp will look for ? Sorry, a bad 'find' on my part - I was looking for '*.pc' files, in fact they are all in the xulrunner/installer/ directory as '*.pc.in'. Yup - as far as I can tell, they're installed normally if you build the Firefox sources with --enable-application=xulrunner instead of =browser. Although the Slackware scripts also use a classic ./configure build, instead of the odd one BLFS uses - I don't know if there's a reason for that... I haven't actually tried this, but you can also just grab the xulrunner tarball: http://releases.mozilla.org/pub/mozilla.org/xulrunner/releases/ FWIW, I'm pretty sure fedora builds xulrunner from these tarballs and not from the firefox source. But there's probably not a lot of difference since they're just different branches of the same code. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: KDE MIME issues
On Tue, Oct 7, 2008 at 9:36 PM, William Tracy [EMAIL PROTECTED] wrote: At this point I would want to confirm that kdelibs did indeed build and install correctly, and to maybe run ldconfig before trying to build kdebase again. ldconfig doesn't seem to fix anything. I went through several rebuilds (make uninstall, then make clean, then follow the build instructions again) trying to see if I screwed up an environment variable or something. No dice. Anything in particular you would recommend looking at when trying see whether kdelib built correctly? I'm not a KDE person, so this is just a wild guess looking at the files that are installed by kdelibs. I have a directory full of protocols and other stuff in /usr/share/service. In particular, I see /usr/share/services/http.protocol. Maybe you're missing this file or the applications are looking in the wrong location. strace might help here. Something like: strace -f -eopen kdesktop And then look at all the places it's trying to open and see if they match your installation. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: cursor control in console
On Tue, Oct 7, 2008 at 9:34 AM, Scott Castaline [EMAIL PROTECTED] wrote: both ldd /bin/bash ldd /bin/sh get the following responses: linux-gat.so.1 = (0xe000) libdl.so.2 = /lib/libdl.so.2 (0xb7ee5000) libc.so.6 = /lib/libc.so.6 (0x7dbe000) /lib/ld-linux.so.2 (0xb7ef6000) Bash isn't using readline, although I suppose it would be using it's own internal copy in that case. $ ldd /bin/bash linux-gate.so.1 = (0xb7f84000) libreadline.so.5 = /lib/libreadline.so.5 (0xb7f3d000) libdl.so.2 = /lib/libdl.so.2 (0xb7f39000) libc.so.6 = /lib/libc.so.6 (0xb7df6000) libncursesw.so.5 = /lib/libncursesw.so.5 (0xb7da9000) /lib/ld-linux.so.2 (0xb7f85000) You may want to rebuild bash and make sure you use the --with-installed-readline option. Again, that might not be a big deal since bash should be using readline one way or another. Also, ensure that you've setup /etc/inputrc or ~/.inputrc as described here: http://www.linuxfromscratch.org/lfs/view/development/chapter07/inputrc.html -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: cursor control in console
On Mon, Oct 6, 2008 at 1:29 PM, Scott Castaline [EMAIL PROTECTED] wrote: While reading your response I had rebotted vbox vm to livecd and did a lspci. I noticed 2 lines that were different that seems like when the livecd was created they had the source code for the vbox guest additions software for linux. The 2 lines are as follows, first from my created LFS and then from the LiveCD: 00:02.0 VGA compatible controller: Unknown device 80ee:beef 00:04.0 System peripheral: Unknown device 80ee:cafe 00:02.0 VGA compatible controller: InnoTek Systemberatung GmbH VirtualBox Graphics Adapter 00:04.0 System peripheral: InnoTek Systemberatung GmbH VirtualBox Guest Service Notice VirtualBox Guest Service for the 2nd line under LiveCD boot. This doesn't really matter. It just means that the LiveCD has a newer pci.ids file than on LFS. The BLFS page describes updating this file: http://www.linuxfromscratch.org/blfs/view/svn/general/pciutils.html -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Newbie wants to connect laptop to internet with wireless router... how?
On Sun, Oct 5, 2008 at 4:23 AM, Lauri Kasanen [EMAIL PROTECTED] wrote: If your wlan router uses WPA encryption instead of WEP, you'll also need wpasupplicant. For WEP your steps are fine. After connecting like that, you just need to enter your IP, and the router's: ip addr add 192.168.1.56/24 dev wlan0 ip route add default 192.168.1.56/24 via 192.168.1.1 dev wlan0 echo nameserver 192.168.1.1 /etc/resolv.conf It's very likely that your router is setup as a DHCP server, so you can just use a DHCP client to handle these details. There are two dhcp clients in BLFS: http://www.linuxfromscratch.org/blfs/view/svn/basicnet/dhcpcd.html http://www.linuxfromscratch.org/blfs/view/svn/basicnet/dhcpclient.html This has a summary of the wireless side: http://wiki.linuxfromscratch.org/blfs/wiki/WirelessTools -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Newbie wants to connect laptop to internet with wireless router... how?
On Sun, Oct 5, 2008 at 10:47 AM, Christian Gardner [EMAIL PROTECTED] wrote: It's very likely that your router is setup as a DHCP server, so you can just use a DHCP client to handle these details. There are two dhcp clients in BLFS: http://www.linuxfromscratch.org/blfs/view/svn/basicnet/dhcpcd.html http://www.linuxfromscratch.org/blfs/view/svn/basicnet/dhcpclient.html This has a summary of the wireless side: http://wiki.linuxfromscratch.org/blfs/wiki/WirelessTools Hello Dan, I remember your excellent help from the last brick wall I hit! I'd prefer to install as little as possible. If, as it would appear, I can get online just by issuing a few commands in a script, would there be any advantage in installing a dhcp client? What would that do that those few commands wouldn't? The DHCP client talks to the DHCP server (your router) to find out the IP adress to use, DNS servers, gateway, etc. While you may know these settings on your home router, if you take the laptop elsewhere, you'll most definitely need a DHCP client to do this work for you. Both DHCP clients are pretty small and probably worth it unless you're using static IP addresses everywhere. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Login Security
On Fri, Oct 3, 2008 at 11:58 AM, Scott Castaline [EMAIL PROTECTED] wrote: Started installing some security packages onto my LFS-6.3 base system. Installed the following in the order listed: libgpg-error-1.5 tetex-3.0 libgcrypt-1.2.4 gnutls-1.6.3 cracklib-2.8.12 linux-pam-0.99.10.0 (created /etc/pam.conf as shown in blfs-6.3 book under config info for this pkge) shadow-4.0.18.1 (reinstall as required.) I did not finish making the configuration of shadow when I accidently logged out. Now I can't login as either root or regular user. I'll enter the login and it'll just sit there never asking for password before finally stating Login incorrect. How do I fix this? Can I boot using the LiveCD add in the scripts and be able to boot again from my system? Also, which way should I go, using /etc/pam.conf configs or directory based security using /etc/pam.d/files? Yeah, you'll need to use a LiveCD or some other way to get to the pam configuration. I'd suggest using /etc/pam.d/login and getting rid of pam.conf (it would just get real bloated over time). There should be nothing wrong with the BLFS suggested login configuration, but in case you just can't get it working, this should at least work temporarily: cat /etc/pam.d/log EOF auth required pam_unix.so nullok account required pam_unix.so session required pam_unix.so password required pam_cracklib.so retry=3 password required pam_unix.so nullok md5 shadow use_authtok EOF Realize that that's very permissive, so you'll want to get a more secure configuration once you're up and running again. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Login Security
On Fri, Oct 3, 2008 at 1:42 PM, Bruce Dubbs [EMAIL PROTECTED] wrote: Dan Nicholson wrote: That said, pam is pretty complex. That is why I don't use it. Don't get me wrong. In a multi-user envronment, it may be necessary, but in most single user environments it really just gets in the way. The LFS servers don't use it either and I'm not aware of any security problems that have occurred in the last 9 years where PAM wold have helped. True, true. However, Scott did say later that he'd like to learn more about security. Like it or not, pam is a major piece of the puzzle on modern linux systems. So, if that is his goal, getting a first hand education on pam will serve him well. I do agree, though, that it you can certainly have a secure system without pam. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Problem building font in Xorg
On Wed, Oct 1, 2008 at 2:57 PM, Cliff McDiarmid [EMAIL PROTECTED] wrote: Hi I'm trying to install the Xorg fonts in BLFS. Both font-adobe-100dpi-1.0.0 and 75dpi are giving me the same error: checking for mkfontdir... /usr/bin/mkfontdir checking for pkg-config... /usr/bin/checking for mkfontscale... /usr/bin/mkfontscale pkg-config checking pkg-config is at least version 0.9.0... yes checking for MAPS... configure: error: Package requirements (fontutil) were not met: Am I missing an app. here that contains fontutil or is this a bug as a mail on google suggests? And what is MAPS? You need font-util, which contains maps of codes from various character sets to Unicode. If you look at the wget list, it's right at the top. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Compiling gtkmm
On Sat, Sep 13, 2008 at 5:28 PM, Nicolas FRANCOIS [EMAIL PROTECTED] wrote: Le Sat, 13 Sep 2008 19:13:24 +0100 Ken Moffat [EMAIL PROTECTED] a écrit : For the future, keeping notes of what works and what changed is always a good idea. Unfortunately, I still have trouble achieving this! It may help to put your scripts into just a few scripts which serve identifialble purposes. I have my own scripts for xorg, basics (toolkits, windowmanager, firefox and other essentials), extras (cups, gimp, etc), audio-video, gnome-stuff. When I upgrade (not very often) I try to preserve the old versions of my scripts and add the changes into the current versions. I usually take great care of what I do on my computer, with a little script. But I thought gtk+ was a piece of cake... Now I understand why you BLFS guys don't follow the progresses of the Gtk/Gnome team : these are no progress ! It's absolutely impossible to follow the successive versions and their dependencies :-( For example, even the README of the gtk+ packages don't mention a version of pango or atk. But this seems to be very important ! I had to use a dev version of pango to install the latest stable version of gtk+ ??? The whole GTK+ stack follows the GNOME release schedule. So, you can go off of what versions were part of the GNOME release you're following. http://download.gnome.org/platform/ http://download.gnome.org/desktop/ http://download.gnome.org/bindings/ http://download.gnome.org/admin/ http://download.gnome.org/devtools/ You should be able to find all the packages there sorted by GNOME release. This is how Randy sorts out the versions that go into BLFS. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: i810 - i915
On Wed, Sep 10, 2008 at 7:49 PM, William Tracy [EMAIL PROTECTED] wrote: I have all the output online here: http://users.csc.calpoly.edu/~wtracy/1420/ Thanks. The full output from running lspci under Ubuntu is here: http://users.csc.calpoly.edu/~wtracy/1420/lspci.ubuntu It's a little short on info. Can you post `lspci -v'? You can just strip out the VGA device. However, it does say Mobile Integrated Graphics. Since you said this is a 965, I'm guessing this is a 965GM. Is this a laptop? Which one? (sorry if you already answered that) and the Xorg log from after the driver update is here: http://users.csc.calpoly.edu/~wtracy/1420/Xorg.0.log.lfs.2 If you look at the end of the log, you'll see that 965GM is not listed. Most likely you'll need xf86-video-intel-2.x.x. Like I said earlier, I _think_ you can build that against the old xorg-server in Xorg-7.2. My honest best guess right now is that the Xorg driver doesn't like my kernel, at 2.6.22.5, unpatched. Not in this case. The xorg server and drivers are basically like their own kernel in userspace. In this case, the driver just doesn't know how to support your hardware. The kernel drm modules only come into play when you start using GLX/DRI, which mostly comes into play for 3D through Mesa's libGL or through AIGLX. But you're not getting that far. Work is being done in Xorg now to have more of the driver work offloaded to the kernel drm drivers, but that's not the case with your server. I highly, highly suspect that updating to xf86-video-intel-2.x.x will at least get your video up. Whether there will be other problems after that, that's another matter. BTW, the FB drivers in the kernel have no effect on X. Except that sometimes they interfere with the X drivers and cause problems. If you don't need a framebuffer driver for your video card and can deal with just versafb, don't even build the other FB drivers. They'll be autoloaded if they exist at boot. Random thought--I've been building Xorg under Ubuntu in a chroot. (I like having Ubuntu up and being able to work on other things while I wait for stuff to build.) Is there a chance that Xorg is doing some by magic by detecting the kernel version at build time, and is unhappy about running under a different kernel when I boot into LFS? No. Not until you actually run X does it interact with the system. The build is completely contained to the other components in the chroot. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: i810 - i915
On Wed, Sep 10, 2008 at 8:41 PM, William Tracy [EMAIL PROTECTED] wrote: On Wed, Sep 10, 2008 at 8:15 PM, Dan Nicholson [EMAIL PROTECTED] wrote: It's a little short on info. Can you post `lspci -v'? You can just strip out the VGA device. 00:02.0 VGA compatible controller: Intel Corporation Mobile Integrated Graphics Controller (rev 0c) (prog-if 00 [VGA]) Subsystem: Dell Unknown device 01f3 Flags: bus master, fast devsel, latency 0, IRQ 16 Memory at fea0 (64-bit, non-prefetchable) [size=1M] Memory at e000 (64-bit, prefetchable) [size=256M] I/O ports at eff8 [size=8] Capabilities: [90] Message Signalled Interrupts: Mask- 64bit- Queue=0/0 Enable- Capabilities: [d0] Power Management version 3 00:02.1 Display controller: Intel Corporation Mobile Integrated Graphics Controller (rev 0c) Subsystem: Dell Unknown device 01f3 Flags: bus master, fast devsel, latency 0 Memory at feb0 (64-bit, non-prefetchable) [size=1M] Capabilities: [d0] Power Management version 3 Ah, shoot. One more. Can you do one more? Try `lspci -nn'. I want to see the PCI ID number. I think your pci.ids file might be out of date because the online one has more information than that. It will be the number near the end of the line in the brackets. My G965 is [8086:29a2]. I suspect you have a GM965, which is [8086:2a02]. And, in fact, the online database shows that it's included in the Inspiron 1420. http://pci-ids.ucw.cz/iii/?i=80862a02 You'll get there. What's interesting is that looking at your Ubuntu Xorg.log, it show that it is the i810 driver, version 1.7.4. I suspect they just patched the driver to match the GM965, too, since it's really similar to the 965. I'm not seeing the patch in my brief look. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: i810 - i915
On Tue, Sep 9, 2008 at 8:22 PM, William Tracy [EMAIL PROTECTED] wrote: On Tue, Sep 9, 2008 at 8:32 AM, NP [EMAIL PROTECTED] wrote: Full support for Intel 965GM chipset needs a recent kernel (e.g. 2.6.24) with intel_agp module as well as a recent version of X (7.2 does not work) with i915 module. Works fine for me with DRI enabled on Dell Vostro 1510 and Dell Inspiron 1525. Ubuntu somehow has it working on my hardware with a 2.6.20 kernel and X 7.2. I wouldn't rule out some clever patching on the part of Ubuntu, though. What hardware do you have? If it's a straight G965 and not a GM965 (I think), you should be able to use xf86-video-intel-1.7.4 with Xorg-7.2. I used that for a long time on my G965. You can probably also run the newer xf86-video-2.x.x releases, which don't use the BIOS for modesetting. I _think_ they should run against xorg-server-1.2.x. If you're not sure, please post the output from `lspci' about your specific card. Also, a full Xorg.log might help. I did try building the new version of the Intel module with my current kernel and Xorg, and I can see it being loaded in the log. However, I still have no joy. Can you attach a full log from that? I'm having a hard time believing you have unsupported hardware with the new driver. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: How to get audio from flv files and gstreamer
On Mon, Sep 8, 2008 at 1:51 AM, Simon Geard [EMAIL PROTECTED] wrote: Does anyone here happen to know how to play the audio part of an FLV file under Linux (e.g downloaded clips from youtube)? Using totem (or any gstreamer-based player) with the gst-ffmpeg package installed, I can see the video just fine, but don't get any sound. Watching them in the browser via the flash plugin works fine, but I want to be able to access them offline too... swfdec uses gstreamer and decodes flv streams, but it seems like the decoder is internal to libswfdec instead of a dedicated gstreamer module. Still, you may want to try that out. There's a sample player in the tarball as well as a full player for gnome and a plugin for mozilla browsers. http://swfdec.freedesktop.org/download/ -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: i810
On Sun, Sep 7, 2008 at 6:18 PM, William Tracy [EMAIL PROTECTED] wrote: Okay, stupid question time again. :-) Does the i810 driver for Xorg require any special kernel support? Xorg launches and works (mostly) if I specify the vesa driver in xorg.conf, but fails if I specify i810: (II) I810: Driver for Intel Integrated Graphics Chipsets: i810, i810-dc100, i810e, i815, i830M, 845G, 852GM/855GM, 865G, 915G, E7221 (i915), 915GM, 945G, 945GM, 965G, 965G, 965Q, 946GZ (II) Primary Device is: PCI 00:02:0 (WW) I810: No matching Device section for instance (BusID PCI:0:2:1) found (EE) No devices detected. Fatal server error: no screens found I'm reasonably certain that the Xorg driver is built correctly (I see all the relevant .so files), and I have verified that i810 is in the fact the driver that Ubuntu is successfully using on the same hardware. Did you actually install the i810 driver? It should be known as intel nowadays and comes from the xf86-video-intel driver package. Nothing kernel related here. Also, what intel chipset do you have? -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: ifconfig [up|down] broken
On Thu, Sep 4, 2008 at 1:50 AM, Simon Geard [EMAIL PROTECTED] wrote: On Tue, 2 Sep 2008 22:10:08 -0700, Dan Nicholson [EMAIL PROTECTED] wrote: Wireless is not a lot of fun, especially if encryption is involved. I don't think we cover it much at all in BLFS. I personally let NetworkManager handle all the details, but getting that all built and setup is another story altogether. Although not too long a story these days... the current SVN can be run pretty easily on LFS systems, without the large patches 0.6 required. A few dependencies, but if you're running Gnome or KDE, you've already got a lot of them. True; it's very nice that most of the backend has become generic. My patch became pretty small (oh, wait I'm still using 0.6.5). I guess I'm thinking about someone who may not have a full D-Bus/HAL/X/GNOME stack going. By the time you get NM and a client (nm-applet), that's a lot of packages. Isn't there a CLI interface to NM? Have you ever tried it? I always thought that would be a neat project. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: ifconfig [up|down] broken
On Tue, Sep 2, 2008 at 9:41 PM, William Tracy [EMAIL PROTECTED] wrote: Hello, I am playing with LFS for the first time on my Dell Inspiron 1420n. (At some point, I might write a beginners' walkthrough for getting LFS working on this laptop model.) Overall, I like what I've seen so far of LFS. Except for some oddness with the SCSI drive on my machine (kernel modules don't help much for mounting the root directory when you're not using initrd ...) all of the problems I've had earlier I've traced directly to something stupid I did. Yeah, initramfs is definitely the way to go, but it can be a serious pain to setup. Bryan Kadzban created an initramfs tool, but it hasn't made it into LFS yet. Someday. Now, I successfully compiled the DHCP client without the iproute patch, as I intended to use net-tools, which I also built (successfully, I thought). Now, if I boot up with the ethernet cable already plugged in, I get an IP address and can surf the web in glorious ASCII with Lynx. I can run: ifconfig eth0 and see the information about that device and IP address. However, if I run: ifconfig down eth0 or: ifconfig up eth0 I get: eth0: Host name lookup failure ifconfig: '--help' gives usage information At a glance, does this look like a simple configuration problem, or did I screw up the build itself? That's because you need to run the DHCP client again after you bring the device up again. You can do something like /etc/rc.d/init.d/network start again to run the LFS network scripts. While I'm emailing the list, I have to comment on the BLFS book itself: Overall, the structure seems sane, but I absolutely cannot fathom why dhcpcd is covered under Basic Networking, while DHCP is covered under Servers. This seems completely the reverse of what I would expect. Am I missing something? The DHCP package contains both a client (dhclient) and a server (dhcpd) for running a DHCP server. The client is covered here, which is under Basic Networking: http://www.linuxfromscratch.org/blfs/view/svn/basicnet/dhcpclient.html Sort of unfortunate, but that's how it goes. Thanks for any replies; I look forward to inflicting a new level of pain on myself with wireless networking once I have net-tools behaving. :-D Wireless is not a lot of fun, especially if encryption is involved. I don't think we cover it much at all in BLFS. I personally let NetworkManager handle all the details, but getting that all built and setup is another story altogether. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: bash script command
On Sat, Aug 30, 2008 at 1:19 AM, arsyante [EMAIL PROTECTED] wrote: btw, i m not using kernel-vanilla for my lfs/blfs, but i m use kernel-source package from opensuse 11 (my host system, cause i think it have better hardware support) I would use the vanilla kernel unless you know exactly what patches are in opensuse's kernel and you want to keep track of them. Of course, if you know that there is some driver that opensuse has added that's not in the vanilla kernel, then I suppose you can try to use the .src.rpm. however the im not using .config from suse because it not bootable (kernel panic) so i using .config from vectorlinux and it works (bootable) are this is legal? Any .config you want to use is perfectly legal (the kernel source will add/remove any config settings it needs with the defaults). The reason your suse config doesn't work is because it expects to use an initramfs to load your the kernel modules for your hard drive. You'll need to build those into the kernel unless you want to make an initramfs, too. If you have no experience with an initramfs, I'd suggest just building stuff into the kernel. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: bash script command
On Thu, Aug 28, 2008 at 10:39 AM, arsyante [EMAIL PROTECTED] wrote: i m sorry maybe this is irrelevant with this milis but i new in linux in past i ussualy use batch file now i try to use shell script i got problem in batch file i ussualy use @echo off some command some command goto end :end some command --- what is replacement goto end and :end in bash script i've tried function(), but that is not excatly what i want i need command that jump to another part of that script There are no labels and jumps in shell. The best you can do is use conditionals. foo() { commands } bar() { commands } if [ $somecondition = 1 ]; then foo else bar fi Here is some good documentation on bash (which is probably the shell you're using): http://tldp.org/HOWTO/Bash-Prog-Intro-HOWTO.html http://tldp.org/LDP/abs/html/ -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Thunderbird 2.0.0.12 Segmentation Fault
On Sun, Aug 3, 2008 at 11:17 AM, Dan McGhee [EMAIL PROTECTED] wrote: Dan Nicholson wrote: I think moz_pis_startstop_scripts is probably a function in run-mozilla.sh. So I don't think the script part is the problem. This would need to be run under a debugger to see what's crashing thunderbird. strace won't help much since the crash is happening within thunderbird-bin. Install gdb. It's pretty straightforward except for the install command: ./configure --prefix=/usr make make -C gdb install. Then run thunderbird under a debugger. The scripts actually accommodate this already: thunderbird -g That should find gdb and run thunderbird-bin through it. When the crash happens, run bt in the debugger. That will give a backtrace from where the program crashed and we can take a gander at the thunderbird source and see why it might be crashing. I was that far when I scaled back this weekend to do some honey do's. gdb installed OK but I don't remember using `make -C gdb install.` If that is important, I will go back and re-do it. But I sure will send a bt soon, if the font thing doesn't clear it up. Since gdb shares the same top-level tree as binutils, you may overwrite your libbfd.a and libiberty.a. But maybe it doesn't do that anymore. Not that it's a big deal anyway since they're already statically linked into the binutils binaries. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Thunderbird 2.0.0.12 Segmentation Fault
On Tue, Jul 29, 2008 at 6:41 PM, Dan McGhee [EMAIL PROTECTED] wrote: I compiled and built Thunderbird, on my laptop, in accordance with the instructions in BLFS-svn-20080712. There were no errors. When I try to run it as root, I get: /usr/lib/thunderbird-2.0.0.12/run-mozilla.sh line 131: 2134 Segmentation fault prog ${1+ $@} `run-mozilla.sh` gives: Cannot execute; and `/usr/lib/thunderbird-2.0.0.12/thunderbird-bin gives: Segmentation fault Line 177 of run-mozilla.sh ( and I don't know if this is relevant) moz_pis_startstop_scripts start Using find tells me that that file does not exist on my system--even in the source directory. I think moz_pis_startstop_scripts is probably a function in run-mozilla.sh. So I don't think the script part is the problem. This would need to be run under a debugger to see what's crashing thunderbird. strace won't help much since the crash is happening within thunderbird-bin. Install gdb. It's pretty straightforward except for the install command: ./configure --prefix=/usr make make -C gdb install. Then run thunderbird under a debugger. The scripts actually accommodate this already: thunderbird -g That should find gdb and run thunderbird-bin through it. When the crash happens, run bt in the debugger. That will give a backtrace from where the program crashed and we can take a gander at the thunderbird source and see why it might be crashing. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: libXft-2.1.12 fails to configure
On Wed, Jul 23, 2008 at 7:34 AM, Dan McGhee [EMAIL PROTECTED] wrote: Installing Xorg-7.2 using latest BLFS svn book on top of LFS-6.3. All other xorg-libs have installed with no problems. The configure error is: Package requirements (fontconfig = 2.2) were not met: No package 'fontconfig' found Consider adjusting the PKG_CONFIG_PATH environment variableif you installed software in a non-standard prefix. [of course the BLFS standard is prefix=/usr] Alternatively, you may set the envrionment variables FONTCONFIG_CFLAGS and FONTCONFIG_LIBS to avoid the need to call pkg-config. $PKG_CONFIG_PATH=/usr/bin/pkg-config:/usr/lib/pkgconfig:/usr/share:/usr/include I'm not sure if this setting is causing the problems, but it's not correct. PKG_CONFIG_PATH should only contain paths where there is .pc metadata files. So, a valid setting would be PKG_CONFIG_PATH=/usr/lib/pkgconfig:/usr/share/pkgconfig. However, those two paths are the default search path, so if you're only installing in /usr, you can omit PKG_CONFIG_PATH entirely. You can test this easily enough: $ pkg-config --modversion fontconfig $ pkg-config --cflags fontconfig $ pkg-config --libs fontconfig $ pkg-config --cflags fontconfig = 2.2 $ pkg-config --libs fontconfig = 2.2 The last two commands being what configure is doing, but storing the results in FONTCONFIG_{CFLAGS,LIBS}. Maybe your fontconfig is too old? export FONTCONFIG_CFLAGS=/usr/include/fontconfig export FONTCONFIG_LIBS=/usr/lib ./configure $XORG_CONFIG This is not exactly what you're looking for, and I'm surprised it passed. The _CFLAGS variable is stuff that will be used when compiling, so a valid setting would be -I/usr/include/fontconfig. And the _LIBS variable is stuff that will be used when linking, so -L/usr/local/lib -lfontconfig would be valid. and it configured, made and installed successfully. I'm glad, but also confused. I didn't even get any me too's when I googled and found nothing similar in the archives. Therefore, I'm assuming that it was something I did or overlooked. Yeah, this seems like a configuration error rather than in the source of Xft or fontconfig. Try the commands above; if they're failing, then there's no reason to continue until they're resolved. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: BLFS 6.3-rc2, NSS-3.11.7
On Wed, Jun 18, 2008 at 11:03 AM, Randy McMurchy [EMAIL PROTECTED] wrote: Michael Franzl wrote: My machine's hostname is 'quantum', which is connected to a DHCP Router. On this router the domain name is set to 'lan'. Therefore, my FQDN should be 'quantum.lan'. I can ping quantum.lan, quantum and localhost. But no matter which value I set for the DOMSUF variable (lan, , (none), etc.), the Tests fail with an Error similar to the following: My /etc/hosts : 127.0.0.1 localhost Any hints appreciated! This is just a guess, but in the LFS network setup you are asked to create a valid 'hosts' file, which includes an entry for the computer (with the FQDN and any aliases). I don't see this in your 'hosts' file above. Not sure if it will help in your particular setup, but to me it is sort of unusual to not identify the computer in /etc/hosts, unless you have some sort of bind/dns service on your lan. To test, try getent hosts quantum.lan. If glibc can't resolve the address, neither can NSS. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: xorg-7.1 and xorg-server-1.2.0/Mesa-6.5.2
On Wed, Jun 4, 2008 at 3:56 AM, Richard Melville [EMAIL PROTECTED] wrote: Thanks for the advice. I pressed on and ran a diff on the two files and they didn't look much different, so I took a chance and symlinked *slang_version_syn.h* to *slang_pp_version_syn.h* and it built OK. I haven't tested it yet though. I wouldn't worry about them being different at all. They're internal mesa headers, so they're free to be whatever they want. You built mesa-6.5.2 using them, so they're working. Building GLcore is a special case, though. GLcore is written as a sort-of mesa 3d driver. In order to build GLcore, it needs to link against the mesa driver interfaces. These are part of an internal library, libmesa.a, which is statically linked into all the DRI libraries you installed earlier. So, to build GLcore, you need a copy of all the mesa source in order to build this static library again. Unfortunately, xorg-server uses automake, and mesa uses it's own thing entirely. To rebuild the static library with libtool, all the names of the mesa source files are hardcoded into xorg-server Makefiles and this other hacky script, symlink-mesa.sh. I imagine the problem is simply that the list of sources is out of sync from when xorg-server-1.1 was released (with whatever mesa release was available at the time) and mesa-6.5.2. So, it just needs to be resynced. Like I said, this stuff has finally gone away in mesa and xserver master. Now when you build mesa, you get a software DRI driver in addition to the hardware ones. The xserver falls back to using this when it can't find an appropriate DRI driver. I did run into another problem later on with *lnx_agp.c* when building xorg-server with *glibc 2.3.4*. I found your remedy on the freedesktop lists, and it worked a treat. It really had me beaten. I think that was DJ that put together that fix, and I thought it was in one of the BLFS books. You just picked up me proliferating it onto the xorg lists. I can't thank you enough - you really do know your stuff. You're welcome. This happens to be an area that I find very interesting. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Developement under LFS/BLFS question
On Wed, Jun 4, 2008 at 5:35 AM, john q public [EMAIL PROTECTED] wrote: Hello all, I've asked about IDEs in the past and there are a lot to choose from so for now I've put that off and just edit source without anything else. BUT now I'm running into segfaults and such (worse theyre sporadic so sometimes things mostly work other times everything grinds to a halt). I saw no mention of gdb in either the LFS or BLFS book. Is it not a good thing to use? gdb is a fine thing to use. I think it's mentioned briefly in the programming section. I have a half completed patch that adds strace and gdb to the book, but I never got around to finishing it. Building gdb is easy, but watch out on the install since it steps on parts of binutils if you don't do it right: ./configure --prefix=/usr make make -C gdb install My REAL question is does anyone have advice on when its better to comb the source versus chasing problems with the debugger? I believe that Mr. Torvalds thinks debuggers are almost always the wrong thing and a waste of time vs. just thinking about the source. But everyone has their preference. I personally don't know gdb that well, so it's usually is faster for me to debug problems by reading the code. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: halt: must be a superuser to use halt
On Thu, May 22, 2008 at 2:38 AM, Bharath Singh [EMAIL PROTECTED] wrote: Hi, I have successfully got the 64-bit CLFS installed with BLFS 6.2 installed on top it. Shutting down using halt from normal users is not happening. I get this halt: must be a superuser to use halt and the same goes for mount: must be a super user to use mount. Any directions to the configuration files to uses these commands as normal user is welcome. Thank you for all your support for keeping the LFS project going. Cheers to you !!! This is really a CLFS question, but... halt must be run as the superuser. Read halt(8). Same with mount, except that mount and umount are usually made suid to root. If they're not suid, then something went wrong during the installation. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: missing xproxymngproto (typo)
On Thu, May 22, 2008 at 5:45 AM, nettxzl [EMAIL PROTECTED] wrote: Sorry, there was a typo in my last email I wrote two other packages in Xorg Applications, xfwp and xrw, still require xproxymngproto Should these two packages also be removed from the install? It should read xfwp and xrx (not xrw) Good catch. Just comment them out from the wget file for now, and we'll get them updated in the book. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: missing xproxymngproto (typo)
On Thu, May 22, 2008 at 11:10 AM, Randy McMurchy [EMAIL PROTECTED] wrote: Dan Nicholson wrote these words on 05/22/08 12:56 CST: Good catch. Just comment them out from the wget file for now, and we'll get them updated in the book. Hmmm. I remember when DJ commented those out recently. He also said he did *two* full builds after that and this wasn't discovered. Odd. He commented out a couple other apps that depend on xproxymngproto, but xfwp and xrx definitely do, too. http://cgit.freedesktop.org/xorg/app/xfwp/tree/configure.ac http://cgit.freedesktop.org/xorg/app/xrx/tree/configure.ac I have no idea how his builds were completing without xproxymngproto (xproxymanagementprotocol) unless he was just testing by removing liblbxutil (which, in turn, uses the xproxymngproto headers). xproxymngproto is definitely dead, though. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Questions about GTK/Glade development
On Sun, May 18, 2008 at 6:50 PM, David Jensen [EMAIL PROTECTED] wrote: On Sun, 18 May 2008 23:00:42 +0200 Thomas Trepl [EMAIL PROTECTED] wrote: Am Sonntag, 18. Mai 2008 17:10:03 schrieb David Jensen: ... You will probably want to learn autoconf and friends, sigh... hmm, any suggestions one that one? I started to look around a bit but i found only very basic samples or too sophisticated things or outdated ones... Yes, it's a pain. As you said, a lot of examples are outdated. Actually, I started with the original BLFS hint. The autotools are really not that hard. The best way to learn it is just to make a silly project. Here is a barebones project making use of autoconf, automake and libtool: cat configure.ac EOF AC_INIT([foo],[0.1],[EMAIL PROTECTED]) dnl foreign is just so automake doesn't complain about missing COPYING, etc. AM_INIT_AUTOMAKE([foreign]) AC_PROG_LIBTOOL AC_CONFIG_FILES([Makefile]) AC_OUTPUT EOF cat Makefile.am EOF # create a shared libary libbaz and a program foo linking to it lib_LTLIBRARIES = libbaz.la libbaz_la_SOURCES = baz.c baz.h include_HEADERS = baz.h bin_PROGRAMS = foo foo_SOURCES = foo.c foo_LDADD = libbaz.la EOF cat baz.h EOF void jimmy(void); EOF cat baz.c EOF void jimmy(void) { return; } EOF cat foo.c EOF #include baz.h int main(void) { jimmy(); return 0; } EOF Rebuild the autotools, configure, make autoreconf -iv ./configure make And there is a ton of documentation on the autotools, with the autobook being a great high level tutorial. http://sourceware.org/autobook/autobook/autobook_toc.html http://www.gnu.org/software/autoconf/manual/ http://www.gnu.org/software/automake/manual/ http://www.gnu.org/software/libtool/manual/ What I've also found useful is just to read some of the autoconf macros. They're in m4, but not that difficult to understand without knowing m4. Look in /usr/share/autoconf/autoconf and /usr/share/aclocal. You'll probably want to be somewhat familiar with the pkg-config macros in aclocal/pkg.m4. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Linux-PAM-0.99.10.0.tar.bz2 md5sum
On Thu, May 15, 2008 at 9:07 PM, nettxzl [EMAIL PROTECTED] wrote: Hi, I downloaded Linux-PAM-0.99.10.0.tar.bz2 from http://www.kernel.org/pub/linux/libs/pam/pre/library/Linux-PAM-0.99.10.0.tar.bz2 and the MD5 sum I got was be4dd1d34ac5933408e13e48f3eb710a I repeated the download several times and also once from the ftp site and the MD5 sum always came out as this value. I checked the GPG signature in the .sign file, and your md5sum is correct. It looks like the f1df... hashsum is for the .tar.gz. I'll fix it. Thanks. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Little problem with ntp.conf ?
On Fri, May 2, 2008 at 1:50 PM, Nicolas FRANCOIS [EMAIL PROTECTED] wrote: Maybe it's me, but I experienced a few problems with the BLFS ntp.conf file : This is a transcript of my dameon.log : May 2 22:34:09 agecanonix ntpd[4854]: ntpd [EMAIL PROTECTED] Mon Apr 14 02:55:25 UTC 2008 (1) May 2 22:34:09 agecanonix ntpd[4854]: precision = 1.000 usec May 2 22:34:09 agecanonix ntpd[4854]: ntp_io: estimated max descriptors: 1024, initial socket boundary: 16 May 2 22:34:09 agecanonix ntpd[4854]: Listening on interface #0 wildcard, 0.0.0.0#123 Disabled May 2 22:34:09 agecanonix ntpd[4854]: Listening on interface #1 wildcard, ::#123 Disabled May 2 22:34:09 agecanonix ntpd[4854]: Listening on interface #2 lo, ::1#123 Enabled May 2 22:34:09 agecanonix ntpd[4854]: Listening on interface #3 eth0, fe80::216:17ff:fef1:f0e0#123 Enabled May 2 22:34:09 agecanonix ntpd[4854]: Listening on interface #4 lo, 127.0.0.1#123 Enabled May 2 22:34:09 agecanonix ntpd[4854]: Listening on interface #5 eth0, 192.168.10.4#123 Enabled May 2 22:34:09 agecanonix ntpd[4854]: kernel time sync status 0040 May 2 22:34:10 agecanonix named[2505]: unexpected RCODE (SERVFAIL) resolving 'tock.nml.csir.co.za//IN': 196.26.5.8#53 Oh, looks like the African NTP server we have in the default ntp.conf is having issues. Looking at pool.ntp.org for Africa: http://www.pool.ntp.org/zone/africa it looks like we should probably change that to za.pool.ntp.org and/or africa.pool.ntp.org. On the other hand, probably all the servers in that file should be commented by default, and the only enabled servers should be the global zone servers: 0.pool.ntp.org. The text should probably just instruct you to uncomment the servers in your region. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: where to find the Trebuchet MS font ?
On Mon, Apr 28, 2008 at 9:42 PM, anonymous anonymous [EMAIL PROTECTED] wrote: I've successfully built WindowMaker from source on my LFS system (with XFree86 4.7). Following the documentation, wmaker.inst is invoked to create ~/GNUstep and modify ~/.xinitrc. Then I run xinit as usual but it fails to start. The console error messages read : wmaker warning: could not load font: Trebuchet MS,Luxi Sans:pixelsize=11. wmaker warning: could not load font: Trebuchet MS,Luxi Sans:weight=200:pixelsize=11. wmaker warning: could not load any fonts. Make sure your font installation and locale settings are correct. wmaker fatal error: could not initialize WINGs widget set wmaker fatal error: it seems that there is already a window manager running It seems that my system does not have the required fonts. Luxi Sans comes with XFree86, but I can't find Trebuchet MS. I'd like to ask where can I get this font ? They are in the Microsoft Core Fonts. See this section for details: http://www.linuxfromscratch.org/blfs/view/svn/x/x-setup.html#xft-font-protocol -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Building Thunderbird-2.0.0.12 64 bit from CBLFS
On Mon, Apr 21, 2008 at 7:07 AM, Arnie Stender [EMAIL PROTECTED] wrote: Hi All, Interesting, this is the 2nd or 3rd time I went to post and couldn't till I unsubscribed then re-subscribed. I keep getting mail from the list but every so often... Does anyone else ever have that happen?? At any rate the reason for the post. I had everything running fine on my new BLFS workstation with X Thunderbird, Firefox and all the tools I needed to continue building and all of a sudden my mail and GDM stopped working. I have been trying to find out what happened but finally decided to just re-build and re-install the failed components but for some reason I couldn't get the same package that worked before to compile so I downloaded Thunderbird-2.0.0.12 and can't get that to compile either. I get the following errors. I need the fix but what I would really like is to know what is happening here. Is this just a bit of bad coding or a bug or a switch that isn't being passed to GCC or what? It has been a lot of years since I have been into serious coding and compilers require a lot of things they didn't used to care about. If possible I would like to get back up to speed so I can fix some of these problems myself. Can some kind soul throw me a bone? Thanks in advance for any help or instruction. Arnie g++ -m64 -o nsAppShell.o -c -DMOZILLA_INTERNAL_API -DOSTYPE=\Linux2.6.18.8-0\ -DOSARCH=\Linux\ -DBUILD_ID=00 -DUSE_XIM -I../../../dist/include/xpcom -I../../../dist/include/string -I../../../dist/include/gfx -I../../../dist/include/pref -I../../../dist/include/dom -I../../../dist/include/necko -I../../../dist/include/uconv -I../../../dist/include/intl -I../../../dist/include/gtkxtbin -I../../../dist/include/imglib2 -I../../../dist/include/widget -I../../../dist/include -I/usr/include/nspr-I../../../dist/sdk/include -I/usr/src/blfspackages/xfce4/mozilla/widget/src/gtk2/../xpwidgets -fPIC -fno-rtti -fno-exceptions -Wall -Wconversion -Wpointer-arith -Wcast-align -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wno-long-long -pedantic -fshort-wchar -pthread -pipe -DNDEBUG -DTRIMMED -O -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12 -DMOZILLA_CLIENT -include ../../../mozilla-config.h -Wp,-MD,.deps/nsAppShell.pp /usr/src/blfspackages/xfce4/mozilla/widget/src/gtk2/nsAppShell.cpp /usr/src/blfspackages/xfce4/mozilla/widget/src/gtk2/nsAppShell.cpp: In member function 'virtual nsresult nsAppShell::ListenToEventQueue(nsIEventQueue*, PRBool)': /usr/src/blfspackages/xfce4/mozilla/widget/src/gtk2/nsAppShell.cpp:230: error: cast from 'void*' to 'gint' loses precision That certainly looks like the type of error -Werror would throw, but I don't see it anywhere. What version of gcc is this? On the other hand, casting from a pointer to a non-pointer seems totally wrong even if it's probably harmless in this instance, so I'm guessing that the real fix is a patch. There should be some way to work around it, though. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Building Thunderbird-2.0.0.12 64 bit from CBLFS
On Mon, Apr 21, 2008 at 7:34 AM, Arnie Stender [EMAIL PROTECTED] wrote: Dan Nicholson wrote: That certainly looks like the type of error -Werror would throw, but I don't see it anywhere. What version of gcc is this? On the other hand, casting from a pointer to a non-pointer seems totally wrong even if it's probably harmless in this instance, so I'm guessing that the real fix is a patch. There should be some way to work around it, though. -- Dan Hi Dan, Thanks for the quick response and the instruction. This is gcc (GCC) 4.2.1. I'll try looking on the Mozilla site. A little googling leads me to believe that this is a 64 bit problem, and I see that you had the same problem with firefox (as you should have since this is in the shared gtk2 widget source for gecko). It's trying to use the macro GPOINTER_TO_INT, which is defined in $libdir/glib-2.0/include/glibconfig.h. In your case, it looks like it's including /usr/lib/glib-2.0/include. So, the first question is: is this multilib 64 bit? Do you have /usr/lib64/glib-2.0/include? Could you grep for GPOINTER_TO_INT in /usr/lib/glib-2.0/include/glibconfig.h? I could be wrong, but I think it also needs to be casting it to (glong) instead of just (gint) by looking at how glibconfig.h is generated. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: transcode fails
On Wed, Apr 16, 2008 at 9:10 PM, Alexander E. Patrakov [EMAIL PROTECTED] wrote: Nicolas FRANCOIS wrote: I encountered this error while compiling transcode : Could you please explain why you use transcode instead of calling ffmpeg directly? Shouldn't transcode be removed from the book, because it stayed broken for so long and nobody noticed it before you? Although it's been a while since I built it (last time was against 2.6.19 headers, or maybe 2.6.22), I use transcode from time to time. A few things I like vs. ffmpeg: * In certain places the interface is easier to understand for handling the different formats/backends/filters/etc. This is pretty subjective, though, since both tools encompass an explosion of options. * There are some nice tutorials on the transcode website for various activities, meaning that the learning curve isn't as steep. * Transcode is a frontend for more than just ffmpeg. So, I can use the mjpegtools for mpeg2 encoding if I feel like it. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: dvd+rw-tools compilation problem (and a workaround)
On Mon, Apr 14, 2008 at 4:59 AM, Nicolas FRANCOIS [EMAIL PROTECTED] wrote: Hi. I encoutered a problem with the dvd+rw-tools package : root [ /sources/dvd+rw-tools-7.0 ]# make all rpl8 btcflash make[1]: entrant dans le répertoire « /sources/dvd+rw-tools-7.0 » gcc -O2 -D_REENTRANT -c -o growisofs.o growisofs.c growisofs.c: In function 'setup_fds': growisofs.c:724: erreur: 'INT_MAX' undeclared (first use in this function) growisofs.c:724: erreur: (Each undeclared identifier is reported only once growisofs.c:724: erreur: for each function it appears in.) make[1]: *** [growisofs.o] Erreur 1 make[1]: quittant le répertoire « /sources/dvd+rw-tools-7.0 » make: *** [all] Erreur 2 A workaround for this consists in adding #include limits.h to transport.hxx and growisofs.c, as mentionned in http://mail-index.netbsd.org/pkgsrc-bugs/2008/01/09/msg000208.html. I would gladly make a patch, but don't know how to do it :-( More than a workaround, it's the right thing. INT_MAX is defined in the C standard to be in limits.h, so that header should be explicitly included instead of relying on another header implicitly pulling it in. I haven't looked at the source, but it should be pretty straightforward. Just lump limits.h in with the other system headers. This may be due to the new kernel API, or to Gcc ? I would suspect kernel headers changes. Probably before one of the kernel headers was pulling in limits.h, but now it's not. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: dvd+rw-tools compilation problem (and a workaround)
On Mon, Apr 14, 2008 at 12:14 PM, Nicolas FRANCOIS [EMAIL PROTECTED] wrote: Le Mon, 14 Apr 2008 20:00:00 +0200 Thomas Trepl [EMAIL PROTECTED] a écrit : Am Montag, 14. April 2008 15:31:08 schrieb Dan Nicholson: On Mon, Apr 14, 2008 at 4:59 AM, Nicolas FRANCOIS ... I haven't looked at the source, but it should be pretty straightforward. Just lump limits.h in with the other system headers. Yes it is. I have prepared a patch for that but forgot to post it. Here it is... Could you explain how you make this patch ? And how comes that some patches foound on Internet work the basic LFS way (patch -Np -i ../toto.patch) and others don't ? For starters, read the patch(1) and diff(1) manpages. But basically, you create a patch with diff and apply it with patch. diff just takes two arguments and finds the differences between them. The output is then just redirected to a file. There are a few other options that change the behavior, but most of the time you can use same set for each utility, respectively. Here's my typical patch this tarball routine: Unpack a fresh copy; we don't want unrelated junk creeping into the patch. $ tar -xf somepkg.tar.gz Make a recursive copy of the pristine sources to diff against later. $ cp -a somepkg somepkg.orig Make changes to the source. $ cd somepkg; hack; hack; hack Go back to the parent directory. When diff is called from the parent, the somepkg directory will be prefixed in the diff output. This becomes important later when we tell patch that we want to strip the first path component. $ cd .. Recursively diff the pristine sources to the altered sources. The -p and -u options just control the format of the generated diff, but are unnecessary. -N means that any new or removed files will be considered instead of being ignored. -r means to act recursively. The ordering of the arguments is important, too: somepkg.orig goes first because I want to find the differences from the unaltered source to my changes. $ diff -pNur somepkg.orig somepkg somepkg.patch Apply the changes to an unaltered source tree. The -p1 option means that we will strip 1 leading component. This is necessary since we prefixed the patch with the name of the source directory when creating the diff _and_ we've entered the source tree. The paths wouldn't match without it. The -N tries to detect if a patch hunk has already been applied and skips it. Usually I don't use this option since I like to know when my patches have been obsoleted by something else. $ cd somepkg.orig $ patch -Np1 -i ../somepkg.patch That's it. There are a ton of other options for diff and patch, but generally that's what I always use (especially if I'm the one who's actually generated the patch). The case of why doesn't this random patch from the internet work is usually because of a difference in leading components when the patch was generated. Many patches don't include the leading directory name, so using the -p0 option to patch is needed instead of -p1. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Question on users and groups
On Sat, Apr 12, 2008 at 9:59 AM, Andrew Barnes [EMAIL PROTECTED] wrote: OK, thanks for your advice. As I understand it then: (i) It is important to distinguish between what the /etc/passwd and /etc/group file formats allow, and what the useradd utility can do, the first being more general than the last useradd and usermod are for controlling /etc/passwd, but will also affect /etc/group due to any group settings you give users. groupadd and for controlling /etc/group. For information on the passwd and group formats, try `man 5 passwd' and `man 5 group'. (ii) It is possible to create groupless users, by editing the /etc/passwd file directly, and possibly in other ways as well. However, there doesn't seem to be any reason to do so. The group field is not optional. That's why utilities like useradd exist: they don't allow you to enter invalid settings. (iii) If I install an application that uses one of the users that appears in Chapter 3 of the 6.2.0 BLFS books, but that has no indicated GID, I should just create a group with the same name, using any number I like (provided it is not already an assigned GID), and associate the user with that group. That sounds like a sane plan. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Problem with autoFS
On Fri, Apr 11, 2008 at 12:35 AM, Abraão Ferreira [EMAIL PROTECTED] wrote: I'm trying to install the package autoFS version 4.1.4 and after the package version 5.0.3. I'm using kernel version 2.6.22.5, but receive the follow message: /usr/bin/rpcgen -h -o mount.h mount.x cannot find any C preprocessor(cpp) /usr/bin/rpcgen: C preprocessor failed with exit code 1 make[1]: ***[mount.h] error 1 I just tested, and I think you must be missing the symlink to cpp at /lib. I.e., $ ls -l /lib/cpp lrwxrwxrwx 1 root root 14 Feb 29 22:30 /lib/cpp - ../usr/bin/cpp After that, run $ echo | rpcgen -h $ echo $? If you still have errors, check that /usr/bin/cpp actually exists. I tried the suggestion of the websites: sed -i 's:^\(RPCGEN = .*\)$/\1 -Y /usr/bin:' lib/Makefile But it give me another error: sed: -e expression #1, char 35: unterminated 's' command You shouldn't need this fix, but the problem is that the delimiter is the : in this case not a / like you normally do with sed 's/old/new/'. So, the / after the $ needs to be a :. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: building seamonkey
On Sun, Mar 9, 2008 at 11:50 AM, john q public [EMAIL PROTECTED] wrote: font-adobe-100dpi-1.0.0.tar.bz2 font-misc-misc-1.0.0.tar.bz2 font-adobe-75dpi-1.0.0.tar.bz2 font-util-1.0.1.tar.bz2 font-dec-misc-1.0.0.tar.bz2 font-xfree86-type1-1.0.0.tar.bz2 Installed these then reinstalled fontconfig (2.5.0) with pointers to all of my font directories (--default --with font options in configure) in pretty much the order they occur in xorg.conf and things are looking much better. Those are bitmapped fonts, though. You should look into the TrueType fonts like DejaVu and FreeFont. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: building seamonkey
On Thu, Mar 6, 2008 at 9:09 PM, john q public [EMAIL PROTECTED] wrote: I have built seamonkey 1-1-8 according to the svn book instructions for the other version but it has problems with fonts. In particular #25bc shows up as a little box with the hex code in it at google so I know its not good. Most likely, you just need more fonts installed to cover the characters you want. See the section about Xft fonts for X: http://www.linuxfromscratch.org/blfs/view/svn/x/x-setup.html#xft-font-protocol -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Firefox fails only with external NSS
On Sun, Feb 24, 2008 at 11:25 PM, Karel Kulhavy [EMAIL PROTECTED] wrote: I tried to compile Firefox 1.5.0.9 on my freshly installed box with LFS 6.3, BLFS 6.2, Core 2 Duo and GCC 4.1.2. Try using the development version for BLFS. 6.2 is ancient and might not work at all anymore. If I follow the instructions with external NSS and without Pango, I get In file included from /home/clock/mozilla/security/manager/ssl/src/nsCipherInfo.h:40, from /home/clock/mozilla/security/manager/ssl/src/nsCipherInfo.cpp:38: ../../../../dist/include/system_wrappers/sslt.h:3:23: error: sslt.h: No such file or directory Could you show the command that was run? What version of NSS is this? -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Firefox sensitive to -j4
On Sun, Feb 24, 2008 at 2:28 PM, Karel Kulhavy [EMAIL PROTECTED] wrote: I would suggest to add a note into the Firefox instructions to avoid -jn. Such a big program is especially tempting to be compiled this way on a SMP system. Once it made only the configuration phase and didn't do anything. Other time it said it cannot compile even a simple X11 program and when I looked at the gcc call, there was something prepended in front of the gcc command which triggered no such file or directory. Here's the trick with firefox (and all the moz programs). The configure scripts need to run synchronously, but the build can all be done with parallel jobs. So, instead of just running `make -j4 -f client.mk build', run: $ make -j1 -f client.mk configure $ make -j4 -f client.mk build Then you only have to slow down a little to wait for the configure steps to complete (which is the same as any autotooled package, anyway). -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: libxml-1.8.17 fails
On Sun, Feb 24, 2008 at 8:54 AM, [EMAIL PROTECTED] wrote: Joe [EMAIL PROTECTED] schrieb: [EMAIL PROTECTED] wrote: LFS 6.3, GCC 4.1.2. The ./configure complains about nonexisting example/Makefile.in but it finishes without error. Then if I type make, it fails. I do see creating example/Makefile sed: can't read ./example/Makefile.in: No such file or directory but the make step completes successfully. The example directory has a I just tried again. My make doesn't complete successfully. It prints: No rule to make target libxml/tree.h, needed by SAX.lo I just got an idea it could be caused by my alias make=make -j4 (I have a multiprocessor system and without it, compilation is twice as slow). When I do unalias make, it works. Does it mean the libxml makefiles are buggy? Almost certainly, yes. A lot of handwritten Makefiles (i.e., not generated by automake) do not handle parallel jobs well. You'll just have to work around this one by dropping the -j4. Also, instead of aliasing make, you can use the variable MAKEFLAGS to store your options for make. export MAKEFLAGS=-j4 make # buggy makefile that needs workaround make -j1 -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: looking for an integrated motherboard graphics processor that works well with X.org
On Feb 17, 2008 3:35 PM, Troy Will [EMAIL PROTECTED] wrote: I built X.org for a machine with a VIA graphics chip. It was a pain in the ass to get 3d graphics working, and then 3d was buggy. I'm looking for a graphics processor that works nicely with X.org. It seems that Intel does a good job of providing open source drivers, and I've been looking at the ASUS P5E-VM HDMI motherboard because it has the Intel G35 Express Chipset Graphics Controller with GMA X3500. Would anyone suggest a different motherboard for building X.org? I'm not entirely sure about the G35 series being supported well yet. But, as it's derived from the 965 chipset (GMA X3000), I believe it should be pretty similar and have seen commits from some of the Intel developers specific to the G35 chipset. I have an ASUS P5B-VM with the 965 chipset and it's been working fine for me. http://usa.asus.com/products.aspx?l1=3l2=11l3=332l4=0model=1312modelmenu=2 I haven't tried using it on a TV or anything like that, though. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: OpenSSH installation problem.
On Jan 30, 2008 12:58 PM, amarsoft amarsoft [EMAIL PROTECTED] wrote: My host is lfslivecd which has /dev/urandom. But my newly built lfs system doesn't have it. If you're building in a chroot and the host is the livecd, then you should have $LFS/dev/urandom unless you did not bind mount the /dev partition correctly. In that case, I would guess that the ssh-rand-helper program should succeed. If you've booted your new LFS system and you're building from there, then you should have /dev/urandom as long as udev has run successfully. Which environment are you building ssh from? Are you missing any other key devices from /dev (/dev/mem)? -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Cannot browse the internet
On Jan 29, 2008 1:30 PM, Bharath Singh [EMAIL PROTECTED] wrote: Hi The blfs 6.2 (kde) system is up and running successfully expect for browsing the internet. I use a data-card and the internet connection has been established using the comgt package. The ppp0 is up and dns servers appear in resolv.conf file. However when i type www.google.com either in mozilla or konquerer, the page cannot be displayed. However when i type the IP address(64.233.167.104) directly in the address bar (both mozilla konquerer), the google home page appears. Not sure which configuration i missed, that its unable to translate the www.google.com address. Thanks in advance for all the help.- Bharath It would sound like your DNS settings in /etc/resolv.conf aren't being used. I'd guess that you missed the dns setting for hosts in /etc/nsswitch.conf. hosts: files dns You can test address resolving at a low level using the glibc utility getent. $ getent hosts www.google.com 72.14.253.147 www.l.google.com www.google.com 72.14.253.104 www.l.google.com www.google.com 72.14.253.99www.l.google.com www.google.com 72.14.253.103 www.l.google.com www.google.com -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: DHCP Client
On Jan 26, 2008 10:05 AM, Matthew Plumb [EMAIL PROTECTED] wrote: I've followed the instructions for the setup of DHCP Client and am getting the following error on boot: /etc/sysconfig/network-devices/services/dhclient: line 24: /sbin/dhclient: No such file or directory I checked and, sure enough, this file does not exist. There are no instructions to add this file in the blfs book, so i'm not sure where i went wrong here... Did you actually install the DHCP package? The first link on the DHCP Client page points you to the DHCP page for installation instructions. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: rotate syslogs
On Jan 23, 2008 9:08 PM, Rick Shelton [EMAIL PROTECTED] wrote: On 1/23/08, Jon Fullmer [EMAIL PROTECTED] wrote: I'm interested to hear what other BLFS users use to rotate their syslogs. I thought it was odd that nothing was listed in the book, as it seems like a basic systems need. What do you use? I use logrotate, as outlined in the LFS hint http://www.linuxfromscratch.org/hints/downloads/files/logrorate.txt Same here, pretty much. Except I'm using the 3.7.1 version from Debian with their patches. http://ftp.debian.org/debian/pool/main/l/logrotate/ The logrotate source actually comes from Fedora, which is up to 3.7.6. http://download.fedora.redhat.com/pub/fedora/linux/development/source/SRPMS/logrotate-3.7.6-2.2.fc9.src.rpm -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: JDK and ld.so.conf
On Jan 24, 2008 6:08 AM, Richard Melville [EMAIL PROTECTED] wrote: On checking /etc/ld.so.conf I noticed that /opt/jdk/lib was not present. I've now added it, but I'm not sure whether I needed to or not. There are no actual DSOs there (at least on my system), so you don't gain anything from having the dynamic linker search there. My /opt/jdk/lib (1.5.0.11) just has a few .jar and .idl files. I don't know if this is different for 1.6, but, in general, I think java handles all the DSOs itself. I.e., the ones in /opt/jdk/jre/lib/i386. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: rotate syslogs
On Jan 24, 2008 6:25 PM, Jon Fullmer [EMAIL PROTECTED] wrote: Wow. *blush* Don't know how I missed that one, but thanks for pointing it out, Rick. And thank you, Dan, for going over where to find it. It's simple, I know, but perhaps this reference should be included in the BLFS book. I just made it sound like I got that answer right away :) I did the same thing you did about a year ago, and it took me a long time to figure out where the magical logrotate utility that all the distros were using came from. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Discussion: is building X.org from source now a joke?
On Jan 23, 2008 2:16 AM, Jeremy Henty [EMAIL PROTECTED] wrote: [I'm not sure if blfs-support is the best list for this - I'll happily take it to another list if people think it's appropriate.] There's a big discussion on the x.org list about the state of the Xorg tree. Apparently the complete tree hasn't built in over a year. Some people are arguing for merging the drivers back into the server tree. One contributor flamed that the whole idea of building Xorg from source is now a joke. (Others disagreed and posted various build scripts.) Also, in an earlier discussion more than one person claimed that the 7.3 release was inferior to previous 7.x releases in important ways. The 7.3 release is inferior in some ways, but I don't believe it has anything to do with building it from source. The input system received a major overhaul, and it's just pretty raw in spots. I've been using 7.3 for quite some time now, and I've haven't had any problems for my use. So just out of interest I'm wondering what do BLFS developers think of the state of Xorg, particularly of building from source? Is it getting flaky? Should the Xorg devs be sorting their act out or are things OK as they are? I regularly follow xorg and don't think there are any major problems. When you say the complete tree, it needs to be said that the part of the tree that didn't build are ancient, unmaintained drivers. Do people need the xf86-input-aiptek? Do most people even know what it is? So, to me, the only major problem is that it's not being communicated which drivers are unmaintained and need somebody to step up if they want that driver to continue to work. All the drivers you care about have always built. And in fact, someone pointed out that there is a document that shows which parts are maintained: http://gitweb.freedesktop.org/?p=xorg/doc/xorg-docs.git;a=blob;hb=HEAD;f=MAINTAINERS As for who's build script to use, it would certainly be nicer if there was a single script that received all the maintenance and could be the entry point for anyone to build xorg by source. In fact, there is one (quite a few, actually), but it doesn't get enough love. But the fact so many people showed their build scripts highlights how trivial it really is. The difficult part is figuring out what order to build in. It would be nice, though, if someone stepped up and maintained one good script. It could be you or me, it's not rocket science. http://gitweb.freedesktop.org/?p=xorg/util/modular.git;a=blob;hb=HEAD;f=build-from-tarballs.sh It should also be mentioned that people discussing the modular/monolithic merits aren't discussing putting the whole thing back together. Nobody has argued against having the libraries and apps away from the server. They're talking about moving the drivers back in with the server build since the API between them is what breaks and causes ancient drivers to not be able to build anymore. Having them together would highlight the breakage immediately instead of relying on an extra step to build all the ancient drivers. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: totem and startup-notification
On Jan 21, 2008 4:40 PM, Ken Moffat [EMAIL PROTECTED] wrote: On Sun, Jan 20, 2008 at 11:01:29PM +, Ken Moffat wrote: Thanks for this, and your other response - I'll play with the patches once I'm back on LFS-6.3 x86. After I looked at the patch, I started to think it might be trying to fix a different problem. The MOZILLA_NOT_LINKED_CFLAGS doesn't sound like what I think I'm building. To clarify, at the end of configure I get the following report: configure: Totem was configured with the following options: configure: ** Using the GStreamer-0.10 backend configure: ** Easy codec installation support enabled configure:nvtv support disabled configure:vanity compilation disabled configure: ** GNOME version enabled configure: ** Browser plugin enabled (using firefox) configure: ** Basic browser plugin enabled configure: ** GMP (Windows Media) plugin enabled configure: ** Complex (Real) plugin enabled configure: ** NarrowSpace (QuickTime) plugin enabled configure: ** MullY (DivX) plugin enabled configure: ** Nautilus properties page enabled configure:Media player keys support disabled configure:LIRC support disabled configure:HAL support disabled configure: ** XTest (legacy screensaver) support enabled configure: ** D-Bus (gnome-screensaver) support enabled configure: ** XVidmode support enabled configure: ** XFree86 multimedia keys support enabled configure: End options I'm surprised by 'Nautilus properties page enabled', and also by 'legacy screensaver', but that's what it finds. I don't know what Nautilus properties page is, but I think legacy screensaver just means it will fall back to using libXScrnSaver if it doesn't find the DBus interface to gnome-screensaver at runtime. FWIW, I couldn't persuade 'patch' to apply your patch to 2.18.2, which is what is in blfs-svn, even after I sorted out the line-wrap at the end - the hunk is at line 379 instead of 466, but even after editing that it still wouldn't apply for reasons that escape me. Yeah, that was against 2.20.3, which I thought you mentioned you were using. Maybe the variables are named something different in 2.18.x. The quick and dirty fix needs cat browser-plugin/Makefile.in instead of cat ..., I think. Anyway, it still doesn't help, the error messages with fresh source still begin totem-plugin-viewer.c:38:22: error: libsn/sn.h: No such file or directory totem-plugin-viewer.c:587: error: expected ')' before '*' token totem-plugin-viewer.c:594: error: expected ')' before '*' token totem-plugin-viewer.c:601: error: expected ')' before '*' token totem-plugin-viewer.c: In function 'free_startup_timeout': totem-plugin-viewer.c:662: error: 'sn_launcher_context_unref' undeclared (first use in this function) totem-plugin-viewer.c:662: error: (Each undeclared identifier is reported only once totem-plugin-viewer.c:662: error: for each function it appears in.) totem-plugin-viewer.c: In function 'startup_timeout': totem-plugin-viewer.c:690: error: 'SnLauncherContext' undeclared (first use in this function) totem-plugin-viewer.c:690: error: 'sn_context' undeclared (first use in this function) So, clearly it's not picking up the startup-notification CFLAGS. So, take a look at browser-plugin/Makefile.am. Look for totem_plugin_viewer_CFLAGS. In 2.20.3, it was using $(BROWSER_PLUGIN_CFLAGS), which gets generated using the pkg-config check in configure.ac. So, the hack is to find some variable it's using for totem_plugin_viewer_CFLAGS and tack on your extra information onto the end of browser-plugin/Makefile.in (not Makefile.am unless you want to regenerate the autotools). You could probably even just use `echo totem_plugin_viewer_CFLAGS += $(pkg-config --cflags libstartup-notification-1.0) browser-plugin/Makefile.in'. And the libraries into totem_plugin_viewer_LDADD. Or, you can just kludge the whole thing in through configure. export CPPFLAGS=$(pkg-config --cflags libstartup-notification-1.0) export LIBS=$(pkg-config --libs libstartup-notification-1.0) ./configure ... -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: totem and startup-notification
On Jan 21, 2008 5:56 PM, Ken Moffat [EMAIL PROTECTED] wrote: On Mon, Jan 21, 2008 at 04:59:49PM -0800, Dan Nicholson wrote: So, clearly it's not picking up the startup-notification CFLAGS. So, take a look at browser-plugin/Makefile.am. Look for totem_plugin_viewer_CFLAGS. In 2.20.3, it was using $(BROWSER_PLUGIN_CFLAGS), which gets generated using the pkg-config check in configure.ac. FWIW, I don't see that (this is 2.20.3, 2.18.2 looks the same) - totem_plugin_viewer_CFLAGS = \ $(EXTRA_GNOME_CFLAGS) \ $(WARN_CFLAGS) \ $(DBUS_CFLAGS) \ $(NVTV_CFLAGS) \ $(AM_CFLAGS) Maybe EXTRA_GNOME_CFLAGS could be used, but I went with one of your later alternatives... Looks like it got fixed upstream now. http://svn.gnome.org/viewvc/totem?view=revisionrevision=4977 I'm still guessing, but the reason it worked for me was because EXTRA_GNOME uses gnome-desktop, which was pulling in startup-notification. $ grep Requires: /usr/lib/pkgconfig/gnome-desktop-2.0.pc Requires: gtk+-2.0 libgnomeui-2.0 libstartup-notification-1.0 Maybe you're not getting that? I think startup-notification is an optional dep for gnome-desktop. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: totem and startup-notification
On Jan 20, 2008 7:30 AM, Ken Moffat [EMAIL PROTECTED] wrote: I've got an odd one here, and I suspect it's down to something I'm doing or omitting, but for the life of me I can't see what. My current builds include totem, mainly because I'm building gstreamer as a dependency for gnash (so, this is totem with the gst backend). If I don't patch totem, both the 2.18.2 and 2.20.1 versions bomb out in the build when they try to reference libsn/sn.h totem-plugin-viewer.c:37:22: error: libsn/sn.h: No such file or directory and then the usual mass of errors. From my quick check, it seems that totem is only pulling in startup-notification indirectly through gnome-desktop. And that only happens on a non-GTK-only build. totem-plugin-viewer.c seems to require startup-notification unconditionally, though. So, I would say that if you're using --enable-gtk, there's definitely a path to build breakage. The right solution is that the enabling the browser plugin needs to check for startup-notification instead of relying on something else pulling it in. This could probably go upstream (discounting gmail breaking formatting): --- configure.in.orig 2008-01-20 12:52:35.0 -0800 +++ configure.in2008-01-20 12:53:30.0 -0800 @@ -466,7 +466,8 @@ if test $enable_browser_plugins = yes [glib-2.0 gnome-vfs-2.0 = $GNOMEVFS_REQS gnome-vfs-module-2.0 = $GNOMEVFS_REQS -gthread-2.0], +gthread-2.0 +libstartup-notification-1.0], [],[enable_browser_plugins=no]) BROWSER_PLUGIN_CFLAGS=$MOZILLA_NOT_LINKED_CFLAGS $BROWSER_PLUGIN_CFLAGS which removes the references to libsn and quite a large chunk of code. With that, both 2.18.2 and 2.20.1 build for me, and seem to work. The oddity is that fedora dropped this for 2.18.2 because of upstream fixes. I'm using startup-notification-0.9, but so are they. I can't see anybody else needing to work around this (e.g. ubuntu, gentoo) so I guess it's something wrong with my builds, but I'm out of ideas. Are you using --enable-gtk? I just built it all fine a couple days ago, but I pulled in all the optional dependencies. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: totem and startup-notification
On Jan 20, 2008 12:56 PM, Dan Nicholson [EMAIL PROTECTED] wrote: --- configure.in.orig 2008-01-20 12:52:35.0 -0800 +++ configure.in2008-01-20 12:53:30.0 -0800 @@ -466,7 +466,8 @@ if test $enable_browser_plugins = yes [glib-2.0 gnome-vfs-2.0 = $GNOMEVFS_REQS gnome-vfs-module-2.0 = $GNOMEVFS_REQS -gthread-2.0], +gthread-2.0 +libstartup-notification-1.0], [],[enable_browser_plugins=no]) BROWSER_PLUGIN_CFLAGS=$MOZILLA_NOT_LINKED_CFLAGS $BROWSER_PLUGIN_CFLAGS The quick and dirty fix being: cat browser-plugin/Makefile.in EOF BROWSER_PLUGIN_CFLAGS += $(pkg-config --cflags libstartup-notification-1.0) BROWSER_PLUGIN_LIBS += $(pkg-config --libs libstartup-notification-1.0) EOF -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: LFS/BLFS on Dell Inspiron
On Jan 8, 2008 6:09 AM, lists [EMAIL PROTECTED] wrote: randhir phagura wrote: Hi, Thanks a lot for encouraging comments received. The detailed configuration is as below: Dell Inspiron - 1520: Intel(R) PRO/Wireless 3945 Dual Band 802.11a/g 54Mbps Wireless Mini Card This device is now supposed to be supported by: http://intellinuxwireless.org/ open source iwlwifi project Yep. They've been merged into 2.6.24, too. I'm using a 2.6.22 kernel with mac80211-9.0.4 and iwlwifi-0.1.12 (a little old now) patched in on my laptop with the 3945 card and it works fine. I sometimes have trouble with WEP when using NetworkManager, but WPA and WPA2 works great. Integrated Intel(R) Graphics Media Accelerator X3100 This may be a problem device. the closest model I could find to it: Laptop Inspiron 1300 915GM Gentoo ~x86 Works with the latest xf86-video-i810 (4/25/2007) and xserver packages, no 915resolution needed with latest drivers. I have this card. It works fine using xf86-video-i810-1.7.4 or the newer xf86-video-intel-2.x releases. What's in LFS stable and BLFS SVN should work fine for you except for the wireless card. You'll have to either upgrade the kernel to 2.6.24 or patch in the mac80211 and iwlwifi releases to older kernels (you also need the firmware). I would recommend using a current distro and checking the hardware compatibility before starting the build, just so you know exactly what chipset drivers you need for the functionality you want before starting the build. That's a very helpful suggestion. Finding out what kernel modules are in use can help a lot when trying to create a new kernel .config. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: iptables compile error
On Jan 7, 2008 4:09 AM, S. Martinez [EMAIL PROTECTED] wrote: I am having problems compiling iptables. First of all, I am using a new install of LFS 6.3 via the Live CD. I am working on two different computers and having the same issue on both. I did do the automated jhalfs build on both (for LFS), and have tried to retrace to see if there might be an issue with that (the install(s) went fine with the exception of the kernel which I did manually). The version of BLFS that I am using is 6.2.0. The problem that I am getting is as follows. I entered the commands as stated in the book, and get the following two errors when compiling. Unable to resolve dependency on linux/netfilter_ipv4/ip_conntrack.h Try 'make clean' Unable to resolve dependency on linux/netfilter_ipv4/ip_nat_rule.h Try 'make clean' Yeah, those headers got removed in the 2.6.22 series. My first thought was maybe there was a conflict because the KERN_DIR=/usr part of the command points to the sanitized headers in /usr rather than the kernel source headers from my running kernel, so I changed that parameter to KERNEL_DIR=/usr/src/linux-2.6.23.12 and still got the same error. Try `make KERNEL_DIR=.' to use the headers shipped in the tarball. I used that last time I built iptables and it seemed to come out correctly. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Mesa error
On Jan 7, 2008 11:01 AM, zux [EMAIL PROTECTED] wrote: Hi I think something happened to my previous mail :) i get this error after make (make OPT_FLAGS=-O2 -fno-strict-aliasing linux-dri-x86): mach64_ioctl.c: In function 'mach64FireBlitLocked': mach64_ioctl.c:190: error: 'drm_mach64_blit_t' has no member named 'idx' make[6]: *** [mach64_ioctl.o] Error 1 make[6]: Leaving directory `/sources/xc/Mesa-6.5/src/mesa/drivers/dri/mach64' make[5]: *** [subdirs] Error 1 make[5]: Leaving directory `/sources/xc/Mesa-6.5/src/mesa/drivers/dri' make[4]: *** [linux-solo] Error 2 make[4]: Leaving directory `/sources/xc/Mesa-6.5/src/mesa' make[3]: *** [default] Error 2 make[3]: Leaving directory `/sources/xc/Mesa-6.5/src/mesa' make[2]: *** [subdirs] Error 1 make[2]: Leaving directory `/sources/xc/Mesa-6.5/src' make[1]: *** [default] Error 1 make[1]: Leaving directory `/sources/xc/Mesa-6.5' make: *** [linux-dri-x86] Error 2 You need to use Mesa-6.5.2 (or 6.5.3) where this is fixed. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Mesa error
On Jan 7, 2008 12:34 PM, zux [EMAIL PROTECTED] wrote: You need to use Mesa-6.5.2 (or 6.5.3) where this is fixed. -- Dan hmm, a litle different with 6.5.3: In file included from nouveau_bufferobj.c:6: nouveau_context.h:34:25: error: nouveau_drm.h: No such file or directory Ohh, that's a mistake in 6.5.3 that's fixed upstream. You shouldn't bother trying to build nouveau unless you really know what you're doing. Try: sed -i 's/nouveau//' configs/linux-dri Or just edit configs/linux-dri by hand so nouveau is not in DRI_DIRS. For a surefire build with Xorg-7.2, you can just follow the BLFS SVN book and use 6.5.2. http://www.linuxfromscratch.org/blfs/view/svn/x/mesalib.html -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Mesa error
On Jan 7, 2008 1:45 PM, zux [EMAIL PROTECTED] wrote: ok MesaLib now compiled with no problems, but xorg-server still fails on the configure script: Creating destination directories for mesa module ... error: Source directory /sources/Mesa-6.5.3/src/mesa/array_cache does not exist configure: error: Failed to link Mesa source tree. Please specify a proper path to Mesa sources, or disable GLX. the source is in /sources/Mesa-6.5.3/ but there realy is no such file as array_cache in the location mentioned Yeah, that's part of the problem with changing Mesa versions. The xorg-server source is hacked up so that it can build some internal parts of Mesa to build the GLX module. However, the filenames are hardcoded into the Makefiles, so if the Mesa developers change any names (delete/rename a file), then the build breaks. That's why I suggested using Mesa-6.5.2 as the surefire build. On the other hand, you could probably just use 6.5.2 for the xorg part and leave 6.5.3 installed. I doubt the GLX interfaces changed significantly between those releases. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: twinkle and generic compile question
On Dec 24, 2007 9:44 AM, alberto hernando [EMAIL PROTECTED] wrote: Hi. I'm trying to compile twinkle-1.1. After fighting with the boost libs, I have all the dependencies and am ready to compile twinkle. configure goes fine and make seems to work too, until... En el fichero incluído de /usr/include/kdeversion.h:23, de /usr/include/kapplication.h:25, de main.cpp:22: /usr/include/kdelibs_export.h:27:21: error: qglobal.h: No existe el fichero o el directorio It can't find the qt headers, which should be in $QTDIR/include. snip QTDIR=/usr ./configure --prefix=/usr --with-qt-dir=/usr --with-qt-libraries=/usr/lib/qt/ --with-qt-includes=/usr/include/qt/ QTDIR=/usr is not valid. The build expects to see everything under one flat directory. I.e., $QTDIR/include should be a symlink to /usr/include/qt. The BLFS dev book suggests making these symlinks in /usr/share/qt. See http://www.linuxfromscratch.org/blfs/view/svn/x/qt.html#qt-config -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: cups ideas
On Dec 23, 2007 5:39 PM, Olaf Grüttner [EMAIL PROTECTED] wrote: installing cups on my system was easily done with the blfs instructions, many thanks. But I had problems configuring. I found out that my kernel didn't include the parallel port support and printing support from the beginning. Maybe mentioning this in the cups instructions will be helpful for others, too. That's probably a good idea. Could you open a ticket? Maybe named printing kernel options or something. Second thing, an icon is installed in the gnome menus: Under Systemwerkzeuge you can find a program called Druckerverwaltung. By clicking on it, it tells that you have not installed htmlview and it stops working further. You can change this for example in the file /usr/share/applications/cups.desktop Just replace htmlview localhost:631 to epiphany localhost:631. Maybe this is worth mentioning? Yes, that's a good one, since htmlview is just a wrapper script, anyway. Another option is to use xdg-open from xdg-utils. These wrapper scripts account for GNOME/KDE/XFCE behavior and will query the default browser from them for an http:// url. http://portland.freedesktop.org/wiki/ -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: gdm won't compile - gdm_xdmcp_host_allow
On Dec 23, 2007 5:28 PM, IVAN ANGELOV [EMAIL PROTECTED] wrote: cc1: warnings being treated as errors gdm-xdmcp-display-factory.c: In function 'gdm_xdmcp_host_allow': gdm-xdmcp-display-factory.c:606: warning: nested extern declaration of 'hosts_ctl' make[3]: *** [gdm-xdmcp-display-factory.o ] Error 1 The Internet doesn't offer many info about that issue. Any ideas how I might cope with it will be more than welcome. You didn't show the command, but the first message suggests that -Werror is being used, causing the compiler to bomb on warnings. I'd see if there's a way to turn that off. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Glib2 assembly errors
On Dec 22, 2007 9:19 AM, Arnie Stender [EMAIL PROTECTED] wrote: Hello All, I have been following the CBLFS instructions for compiling glib-2.14.4 and got the error below while doing the 32bit compile on a multi-lib CLFS system. It was compiling gatomic.o. Has anyone seen this before? I haven't tried the 64 bit build yet. BTW, if I am supposed to be posting questions about this somewhere else please let me know where. I didn't see a reference to a separate list. As always, thanks in advance for any pointers. Arnie gcc -m32 -DHAVE_CONFIG_H -I. -I. -I.. -I.. -DG_LOG_DOMAIN=\GLib\ -DG_DISABLE_CAST_CHECKS -DG_DISABLE_DEPRECATED -DGLIB_COMPILATION -DPCRE_STATIC -pthread -g -O2 -Wall -MT gatomic.lo -MD -MP -MF .deps/gatomic.Tpo -c gatomic.c -fPIC -DPIC -o .libs/gatomic.o /tmp/ccg5TQHp.s: Assembler messages: /tmp/ccg5TQHp.s:119: Error: suffix or operands invalid for `cmpxchg' make[4]: *** [gatomic.lo] Error 1 make[4]: Leaving directory `/usr/src/blfspackages/xorg-7.3/util/glib-2.14.4/glib' When using the autotools on a multiarch host, you want to pass --build=$target_arch (and optionally --host=$target_arch) so that the configure script can make the correct decisions based on $host_cpu, etc. So, in this case, try passing --build=i686-pc-linux-gnu to configure. If you also pass --host=i686-pc-linux-gnu and $CC is not set, configure will also use i686-pc-linux-gnu-gcc, and you won't have to pass -m32. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Glib2 assembly errors
On Dec 22, 2007 10:40 AM, Arnie Stender [EMAIL PROTECTED] wrote: Hi Dan, That did it. In the command they gave it included a variable for the --host that was not set. I replaced it with what you suggested and it ran as it should. It's odd because I have already compiled two thirds of Xorg and this is the first time they used that variable. Oh well. Merry Christmas to all in case I don't have any more problems before then. Thanks Dan. --host/--build/--target is a scary corner of autoconf that you're best to avoid if possible. However, it makes multiarch do the right thing in some cases. In most cases, you don't have to specify what target you're building for to configure. gcc will run the compiler and linker the right way so long as you've specified -m32/-m64. However, as far as configure is concerned, it thinks you're building for x86-64 because that's the output of uname. If a package needs to make decisions based on your architecture (such as what kind of assembly to use), then it might fail unless you specifically tell it what you're building on/for with --build. Fortunately, there are few spots in Xorg that need to make cpu specific decisions (one exception I know of is pixman). For what it's worth, the standard rpm macro for running ./configure always specifies --host, --build and --target. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: New KDE-3.5.8
On Nov 29, 2007 11:38 AM, Andreas Leuner [EMAIL PROTECTED] wrote: Am Mittwoch 28 November 2007 14:58:18 schrieb Dan Nicholson: On Nov 28, 2007 3:36 AM, Alberto Hernando [EMAIL PROTECTED] wrote: El Miércoles, 28 de Noviembre de 2007 12:05, Andrey escribió: /usr/bin/ld: cannot find -lqt-mt snip If you install qt in /opt, make sure that /opt/qt3/lib (or similar) is in /etc/ld.so.conf and run ldconfig later. No, this is a build time failure, nothing to do with ld.so.conf or ldconfig. Can you show the specific error? I don't recall exactly how KDE and Qt play together, but it may be that qmake isn't supplying the correct -L option so the linker finds libqt-mt. Normally setting QTDIR during the build is enough, no entry in /etc/ld.so.conf or setting LD_LIBRARY_PATH or setting LDFLAGS manually is necessary. Shouldn't the KDE build know how to get QTDIR? I should just look... But I have had such failures since building kde (even before version 3.5.6) from blfs. Could this be because of ./configure --enable-new-ldflags ? It adds -Wl,--as-needed -Wl,--enable-new-dtags to the LDFLAGS. This isn't in the book, but maybe Andrey enabled it? I did :-) I use --as-needed a lot and it often causes problems if the command line for the linker is not constructed carefully. Andrey didn't show the failing command, only the error, though. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: New KDE-3.5.8
On Nov 28, 2007 3:36 AM, Alberto Hernando [EMAIL PROTECTED] wrote: El Miércoles, 28 de Noviembre de 2007 12:05, Andrey escribió: /usr/bin/ld: cannot find -lqt-mt snip If you install qt in /opt, make sure that /opt/qt3/lib (or similar) is in /etc/ld.so.conf and run ldconfig later. No, this is a build time failure, nothing to do with ld.so.conf or ldconfig. Can you show the specific error? I don't recall exactly how KDE and Qt play together, but it may be that qmake isn't supplying the correct -L option so the linker finds libqt-mt. You may need to add -L/opt/qt3/lib to the LDFLAGS variable, but that seems wrong since it's never been necessary before. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: New KDE-3.5.8
On Nov 28, 2007 6:50 AM, Andrey [EMAIL PROTECTED] wrote: -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Dan Nicholson Sent: Wednesday, November 28, 2007 3:58 PM To: BLFS Support List Subject: Re: New KDE-3.5.8 No, this is a build time failure, nothing to do with ld.so.conf or ldconfig. Can you show the specific error? I don't recall exactly how KDE and Qt play together, but it may be that qmake isn't supplying the correct -L option so the linker finds libqt-mt. You may need to add -L/opt/qt3/lib to the LDFLAGS variable, but that seems wrong since it's never been necessary before. Well, adding LDFLAGS helps... Now compilation works... Thanks! Yeah, that's definitely a workaround, though. It would be nice to figure out why this isn't working out of the box. I don't have the KDE sources handy, though. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: booting LFS from usb-external disk
On Nov 27, 2007 1:29 PM, Dr. Edgar Alwers [EMAIL PROTECTED] wrote: On Tuesday 27 November 2007 09:39, Heinrich Tomanek wrote: calling grub with a disk option is not the right way, i hope it was a typo. In grub, every disk is a HD, therefore is your disk a hd0. Not at all a typo. Lack of knowledge ! The correct command sequence for sda5 (hd0,4) is: # sh grub # grub root (hd0,4) # grub setup (hd0,4) # grub quit ... # sh reboot # and enjoy Well, this for sure will install in the main /dev/hda5, which I can not hide during this process. The point is, I am trying to install really to /dev/sda5, which should be something like grub root ( sda0,4). Any way to perform this ? Thank you very much for the help, as well as to Lauri's comment. Look in /boot/grub/device.map to see how grub interprets your drives. If it doesn't exist yet, run this command: echo quit | grub --batch --device-map=/boot/grub/device.map -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Can't compile libusb
On Nov 15, 2007 9:48 AM, john q public [EMAIL PROTECTED] wrote: I have tried google searchs in vain but noone seems to know why something like: ../.libs/libusbpp.so: undefined reference to `usb_find_busses' keeps happening when I try to build libusb. That's strange. That symbol should be in libusb.so, which I would assume that libusbpp is linking to. Can you show the whole error stream, from the beginning of the error messages? -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Keyboard Debug messages won't go away
On Nov 13, 2007 5:32 PM, Walter Barnes [EMAIL PROTECTED] wrote: From: Dan Nicholson [EMAIL PROTECTED] FYI, init.d/consolelog is configured from /etc/sysconfig/console. You can add LOGLEVEL=1 there and not have to edit the init script. That was how I intended it and how it works on my system. Thanks, I did not notice that before. However, adding the console file to /etc/sysconfig will also activate the console init script which I don't need right now. Eventually I plan to toy around with UTF-8 but until then I'll keep the LOGLEVEL set to 3 in consolelog. Not quite. Just don't set any of the variables that the console script uses. It will continue to use the defaults it's using now. So: echo 'LOGLEVEL=1' /etc/sysconfig/console will do what you want and leave the other console settings alone. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Keyboard Debug messages won't go away
On Nov 9, 2007 5:48 PM, Walter Barnes [EMAIL PROTECTED] wrote: Open /etc/rc.d/init.d/consolelog and look for the line at the top that sets LOGLEVEL. By default it's set to 7, if you set it to a lower value less kernel messages will be sent to the console. I set my mine to 3 to get rid of iptable log messages, a hancheck message that comes up everytime the laptop returns from standby and lost synchronization messages from the mouse pad. IMO, I think it should be set to a low value to start with. Before editing consolelog you can test different values by executing dmesg -n $LOGLEVEL. FYI, init.d/consolelog is configured from /etc/sysconfig/console. You can add LOGLEVEL=1 there and not have to edit the init script. That was how I intended it and how it works on my system. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: libm/glibc-2.3.6 issue (was - Re: libstdc++ issue?)
On Nov 13, 2007 10:42 AM, juras [EMAIL PROTECTED] wrote: Further investigation of the problem led me to the conclusion that there must be a bug inside the /lib/libm.so.6, which is a part of glibc-2.3.6 (in my system) That probably means that I'll have to upgrade glibc. But I am not sure if the upgrade - (a rather adventurous task) solves the problem. LFS-6.3 contains now the glibc-2.5.1. I would like to ask someone who has the libc-2.5.1 installed to try to compile, run and send the results of the following simple program: $ cat tanh.c EOF #include stdio.h #include complex.h #include math.h int main() { complex z=I*M_PI*0.5; complex th=ctanh(z); printf(z = (%g, %g), ctanh(z) = (%g,%g)\n, creal(z),cimag(z),creal(th),cimag(th)); return 0; } EOF $gcc tanh.c -o tanh -lm $./tanh Please send the result for me. In my system (glibc-2.3.6 according LFS-6.2) the output of the program is: z = (0, 1.5708), ctanh(z) = (nan, inf) Which is wrong! (both with gcc-3.3.6 and gcc-4.2.2) It should be: z = (0, 1.5708), ctanh(z) = (0, inf) Yeah, it works on glibc-2.5.1. $ ./tanh z = (0, 1.5708), ctanh(z) = (0,1.63318e+16) $ /lib/libc.so.6 GNU C Library stable release version 2.5.1, by Roland McGrath et al. Copyright (C) 2006 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Compiled by GNU CC version 4.1.2. Compiled on a Linux 2.6.22.1-2 system on 2007-08-01. Available extensions: crypt add-on version 2.1 by Michael Glad and others GNU Libidn by Simon Josefsson GNU libio by Per Bothner NIS(YP)/NIS+ NSS modules 0.19 by Thorsten Kukuk Native POSIX Threads Library by Ulrich Drepper et al BIND-8.2.3-T5B Thread-local storage support included. For bug reporting instructions, please see: http://www.gnu.org/software/libc/bugs.html. It's possible to look around and find a patch, but it might be difficult for glibc-2.3.6. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: libm/glibc-2.3.6 issue (was - Re: libstdc++ issue?)
On Tue, Nov 13, 2007 at 08:12:42PM +0100, juras wrote: Dan Nicholson napisał(a): Yeah, it works on glibc-2.5.1. Thank you. So I'll have to upgrade the glibc. I know, that it may make my system unusable, but I will try. Just a sec. I found some bug reports and the upstream commits. https://bugzilla.redhat.com/show_bug.cgi?id=160759 http://www.mail-archive.com/[EMAIL PROTECTED]/msg28193.html http://sources.redhat.com/cgi-bin/cvsweb.cgi/libc/ChangeLog.diff?cvsroot=glibcr1=1.9409r2=1.9410 Try the attached patch. -- Dan diff -pNur glibc-2.3.6.orig/sysdeps/generic/s_ctan.c glibc-2.3.6/sysdeps/generic/s_ctan.c --- glibc-2.3.6.orig/sysdeps/generic/s_ctan.c 2001-07-05 21:55:49.0 -0700 +++ glibc-2.3.6/sysdeps/generic/s_ctan.c2007-11-13 11:25:12.0 -0800 @@ -1,5 +1,5 @@ /* Complex tangent function for double. - Copyright (C) 1997 Free Software Foundation, Inc. + Copyright (C) 1997, 2005 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper [EMAIL PROTECTED], 1997. @@ -61,8 +61,18 @@ __ctan (__complex__ double x) den = cos2rx + __ieee754_cosh (2.0 * __imag__ x); - __real__ res = sin2rx / den; - __imag__ res = __ieee754_sinh (2.0 * __imag__ x) / den; + if (den == 0.0) + { + __complex__ double ez = __cexp (1.0i * x); + __complex__ double emz = __cexp (-1.0i * x); + + res = (ez - emz) / (ez + emz) * -1.0i; + } + else + { + __real__ res = sin2rx / den; + __imag__ res = __ieee754_sinh (2.0 * __imag__ x) / den; + } } return res; diff -pNur glibc-2.3.6.orig/sysdeps/generic/s_ctanf.c glibc-2.3.6/sysdeps/generic/s_ctanf.c --- glibc-2.3.6.orig/sysdeps/generic/s_ctanf.c 2004-01-13 01:08:04.0 -0800 +++ glibc-2.3.6/sysdeps/generic/s_ctanf.c 2007-11-13 11:25:20.0 -0800 @@ -1,5 +1,5 @@ /* Complex tangent function for float. - Copyright (C) 1997 Free Software Foundation, Inc. + Copyright (C) 1997, 2005 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper [EMAIL PROTECTED], 1997. @@ -61,8 +61,19 @@ __ctanf (__complex__ float x) den = cos2rx + __ieee754_coshf (2.0 * __imag__ x); - __real__ res = sin2rx / den; - __imag__ res = __ieee754_sinhf (2.0 * __imag__ x) / den; + + if (den == 0.0) + { + __complex__ float ez = __cexpf (1.0i * x); + __complex__ float emz = __cexpf (-1.0i * x); + + res = (ez - emz) / (ez + emz) * -1.0i; + } + else + { + __real__ res = sin2rx / den; + __imag__ res = __ieee754_sinhf (2.0 * __imag__ x) / den; + } } return res; diff -pNur glibc-2.3.6.orig/sysdeps/generic/s_ctanh.c glibc-2.3.6/sysdeps/generic/s_ctanh.c --- glibc-2.3.6.orig/sysdeps/generic/s_ctanh.c 2001-07-05 21:55:49.0 -0700 +++ glibc-2.3.6/sysdeps/generic/s_ctanh.c 2007-11-13 11:25:26.0 -0800 @@ -1,5 +1,5 @@ /* Complex hyperbole tangent for double. - Copyright (C) 1997 Free Software Foundation, Inc. + Copyright (C) 1997, 2005 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper [EMAIL PROTECTED], 1997. @@ -61,8 +61,18 @@ __ctanh (__complex__ double x) den = (__ieee754_cosh (2.0 * __real__ x) + cos2ix); - __real__ res = __ieee754_sinh (2.0 * __real__ x) / den; - __imag__ res = sin2ix / den; + if (den == 0.0) + { + __complex__ double ez = __cexp (x); + __complex__ double emz = __cexp (-x); + + res = (ez - emz) / (ez + emz); + } + else + { + __real__ res = __ieee754_sinh (2.0 * __real__ x) / den; + __imag__ res = sin2ix / den; + } } return res; diff -pNur glibc-2.3.6.orig/sysdeps/generic/s_ctanhf.c glibc-2.3.6/sysdeps/generic/s_ctanhf.c --- glibc-2.3.6.orig/sysdeps/generic/s_ctanhf.c 2004-01-13 01:08:04.0 -0800 +++ glibc-2.3.6/sysdeps/generic/s_ctanhf.c 2007-11-13 11:25:29.0 -0800 @@ -1,5 +1,5 @@ /* Complex hyperbole tangent for float. - Copyright (C) 1997 Free Software Foundation, Inc. + Copyright (C) 1997, 2005 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper [EMAIL PROTECTED], 1997. @@ -61,8 +61,18 @@ __ctanhf (__complex__ float x) den = (__ieee754_coshf (2.0 * __real__ x) + cos2ix); - __real__ res = __ieee754_sinhf (2.0 * __real__ x) / den; - __imag__ res = sin2ix / den; + if (den == 0.0f) + { + __complex__ float ez = __cexpf (x); + __complex__ float emz = __cexpf (-x); + + res = (ez - emz) / (ez + emz); + } + else + { + __real__ res = __ieee754_sinhf (2.0 * __real__ x) / den; + __imag__ res = sin2ix / den; + } } return res; diff -pNur glibc-2.3.6.orig/sysdeps/generic
Re: unable to compile xorg libraries
On Nov 9, 2007 4:22 AM, Allan Lavell [EMAIL PROTECTED] wrote: I don't think this email got through for some reason or another. Here it is in its entirety. I have been unable to successfully copmile the xorg 7 libraries. I have been following the development BLFS tutorial on installing xorg, step by step. xproto-7.2 seems to have installed fine, so i don't know why xorg-lib isnt compiling. the problem seems to lie with xproto. grep xext xorg-lib-compile.log generates teh following output: checking for XXF86MISC... configure: error: Package requirements (xproto x11 xextproto xext xf86miscproto) were not met: No package 'xext' found checking for XXF86VM... configure: error: Package requirements (xproto x11 xextproto xext xf86vidmodeproto) were not met: No package 'xext' found Actually, the output is much longer than that, but it's basically the same thing for every package. I don't get it: xproto installed fine. When I check /usr/lib/pkgconfig, xextproto.pc and xproto.pc are there. I installed xorg to the standard prefix. The pkgconfig path env variable is set up correctly (/etc/profile.d/extrapaths.sh has pathappend /usr/local/lib/pkgconfig PKG_CONFIG_PATH in it). What it's bombing on is xext, which is provided by libXext. Do you have that installed? Are you following the order listed in the lib-7.2.wget file? -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: BLFS 6.2 and LFS 6.3
On 11/6/07, Chris [EMAIL PROTECTED] wrote: First of all. great stuff! This is the most interesting experience i have had with Linux. Question. Is the BLFS 6.2 book ok to use with LFS 6.3 or should I wait for the BLFS 6.3 book. Use the development BLFS book. 6.2 is really long in the tooth and will have some slight incompatibilities with LFS-6.3. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Numlockx can't find X
On 11/6/07, Nicolas FRANCOIS [EMAIL PROTECTED] wrote: While installing numlockx (a little tool to have numlock on just after boot under xdl or kdm), I uncouter problems in the configure process : checking for X... configure: error: Can't find X includes. Please check your installation and add the correct paths! That is just a horrible use of autoconf. The offender is the K_PATH_X macro in acinclude.m4 which is totally unnecessary because there's already a default autoconf macro for finding X. Try adding x_includes=/usr/include x_libraries=/usr/lib to the end of your ./configure command. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Problem with ftp access
On 11/6/07, Nicolas FRANCOIS [EMAIL PROTECTED] wrote: I have an annoying problem with my new LFS box (LFS SVN post 6.3, BLFS SVN) : Everything is working quite OK...except ftp. When I want to connect to an ftp server, everything goes OK until I want to dir a directory : ftp ls 200 PORT command successful. Consider using PASV. 425 Failed to establish connection. Same problem with the ftp program from cpan, I guess, for I can only retreive packages with wget or lynx. With Firefox and Konqueror, ftp is fine. Ncftp works fine, so I think it's a specific problem from the core ftp program installed in LFS. How can I trace this problem to it's origin ? I don't have any ideas, but I would try strace. http://sourceforge.net/projects/strace/ strace -f -o ftp.log ftp Then just do what you normally do. There will be a lot of output in ftp.log, hopefully something will point you in the right direction. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: xterm locale détail
On 11/3/07, Nicolas FRANCOIS [EMAIL PROTECTED] wrote: Hi. I'm going threw the compilation of my brand new LFS (hope I'lll end before end of vacations wednesday !). LFS post 6.3, BLFS SVN. When I launch X the first time after compilation, everyhting works... except xterm says : couldn't find charset data for locale [EMAIL PROTECTED]; using ISO 8859-1 What do I have to set for this message to disappear ? I think it's because luit can't find the locale.alias file. I finally fixed this bug in BLFS a couple days ago. If you do `strings $XORG_PREFIX/bin/luit | grep locale.alias', you'll probably see it pointing to the wrong spot. Rebuild luit with the switch I added a couple days ago. If that's not the issue, use strace. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: Discussion of BLFS, LFS, etc. current and new versions
On 11/2/07, randd [EMAIL PROTECTED] wrote: I would very much like to discuss XULRunner / embedded mozilla - embedded gecko engine apps [snip] ...or it would be really nice to have dedicated forum for discussing the stuff like this? or both? The blfs wiki is intended for just this kind of thing. http://wiki.linuxfromscratch.org/blfs/wiki/BlfsNotes If you have an account, I think you can create new pages. Just try to keep them organized. -- Dan -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page