Re: Port's OpenSSL 1.1.1k, 1 on FreeBSD 11-STABLE shows strange error (Inappropriate ioctl for device) on handshake (Server-side)
On 01/04/2021 00:10, Lev Serebryakov wrote: System is 11.4-STABLE r366937. I've ports built with OpenSSL from ports (DEFAULT_VERSIONS+=ssl=openssl). After upgrade to latest 1.1.1k,1 my nginx reject HTTPS connections with cryptic SSL_do_handshake() failed (SSL: error:02015019:system library:ioctl:Inappropriate ioctl for device error:1427D044:SSL routines:construct_stateless_ticket:internal error error:02015019:system library:ioctl:Inappropriate ioctl for device) What happens? What did I do wrong? Previous OpenSSL version worked without problem. We are using FreeBSD 11.4 with port's OpenSSL too. Already deployed newer version of OpenSSL on all machines without any troubles. (Apache / Nginx / Lighttpd / Postfix / ProFTPd and some other openssl library consumers) Do you have any special SSL / TLS configuration in nginx.conf? (protocol version, ciphers etc.) Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Python 2.7 removal outline
On 25/03/2021 16:03, Baptiste Daroussin wrote: I will only here answer about the quality of the communication of portmgr, yes there is room of improvement in general in the current portmgr team as of how we do communicate about plans and policy and we are working on it. "There is room of improvement" are too kind words. It happened in the past and it is back again. As explained by Olivier and the others in this thread there are no clear policy written and explained to the community, there are mixed terms "all" / "but some exceptions" chosen by what criteria, defined by what policy? It is really annoying for maintainers like Olivier to spend some time to provide solution for port useful for others (Pale Moon and Tauthon in this case) and have it removed from the tree after 4 hours without prior discussion or notice. Who will benefit from this behaviors? It all seems more like witch hunting than any rational moves for community profit. Telling users that they can maintain it locally is like p***ing them in face. And until overlays are not fully supported with poudriere options and easily defined exceptions for MOVED entries it is really not for everybody to use overlays in current state (overlays are poor documented at least). Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Python 2.7 removal outline
On 25/03/2021 07:26, Dewayne Geraghty wrote: On 25/03/2021 4:01 am, Miroslav Lachman wrote: I really appreciate the work of ports team, committers and maintainers but I dislike double standards. All ports requiring Python 2.7 were marked deprecated the last year almost all of them removed according to expiration date 2020-12-31 but some of them are still there. If there is Python 2.7, if there is Chromium then any of removed ports can be there. If "we" want to get rid of them then "we" should remove all of them and not just some by sentiment. For example Iridium browser was removed because of Python 2.7 but Chromium is still there. They are both based on the same source with the same dependencies but Iridium cares more about privacy, yet it was slaughtered instead of Chromium. I really would like to see some policies for things like this next time. Miroslav Lachman Thanks Miroslav, I have the same view. Though I agree with Rene about the need to remove vulnerable ports and the interests of the FreeBSD community, its worth considering those with both a need and an understanding of the ramifications of using python2.7. From the security point of view I can agree with removing ports requiring Python 2.7 as run dependency but if I have it right, Iridium nor Chromium have it as run dependency. Python is needed for build only so users of Chromium, Iridium and many other ports / packages do not need to have vulnerable Python 2.7 installed. But these ports were removed anyway even if there is not proper replacement. Or in case of Chromium vs Iridium the better one was removed. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Python 2.7 removal outline
On 24/03/2021 14:03, Rene Ladan wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA512 Hi, below is an outline continuing the Python 2.7 cleanup: - - all affected ports are now marked as deprecated, with an expiration date of either 2020-12-31 or 2021-06-23. - - we will have to wait for Chromium to fully switch to Python 3 before we can fully remove Python 2.7. This is work in progress on their side. Not waiting would imply removing www/chromium (obviously), editors/vscode (it escaped the recursive-deprecation dance of devel/electron*), but most importantly www/qt5-webengine which would drag half of KDE with it. However, lang/python27 will be marked as RESTRICTED so that all ports mentioned above can still be built and run, but Python 2.7 itself will not be available as a package. [...] I really appreciate the work of ports team, committers and maintainers but I dislike double standards. All ports requiring Python 2.7 were marked deprecated the last year almost all of them removed according to expiration date 2020-12-31 but some of them are still there. If there is Python 2.7, if there is Chromium then any of removed ports can be there. If "we" want to get rid of them then "we" should remove all of them and not just some by sentiment. For example Iridium browser was removed because of Python 2.7 but Chromium is still there. They are both based on the same source with the same dependencies but Iridium cares more about privacy, yet it was slaughtered instead of Chromium. I really would like to see some policies for things like this next time. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: poudriere merging multiple ports trees
On 28/01/2021 17:47, Henrik Rosenke wrote:
I tried this with a patched MOVED file, copied from the master
portstree, deleted the entry and tried to build but this File seems to
be ignored. I dont really understand where the MOVED file is handled, it
seems to be bsd.port.subdir.mk but i am not sure.
MOVED is handled by bsd.port.subdir.mk but overlays do not account with
it. Overlays are handled on more places
Scripts/do-depends.sh
Scripts/depends-list.sh
Next is bsd.port.mk:
.if defined(USE_LOCAL_MK)
.include "${PORTSDIR}/Mk/bsd.local.mk"
.endif
.for odir in ${OVERLAYS}
.sinclude "${odir}/Mk/bsd.overlay.mk"
.endfor
But there is no bsd.overlay.mk and I don't know what should be in it
(file in the overlay ports tree)
The AWK code in bsd.port.subdir.mk can be changed to use MOVED from
overlay if it exists insead of the default one.
I think the best way
to handle this would be a NOT_MOVED file in the overlay portstree with
the ports to ignore from the MOVED file with entries like "sysutils/sge62".
This can be possible too. NOT_MOVED can be used as additional INDEXFILE
Or maybe start with PORTSEARCH_MOVED=0 (not tried yet)
Kind regards
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: poudriere merging multiple ports trees
On 28/01/2021 10:55, Henrik Rosenke wrote: Greetings, UIDs and GIDs are also not resprected, i could solve this via this patch: I didn't tried it but what if you have patched Mk/bsd.port.mk and MOVED in overlay tree? Does it work or not? Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: poudriere merging multiple ports trees
On 25/01/2021 15:10, Baptiste Daroussin wrote: On Sun, Jan 24, 2021 at 10:23:45PM +0100, Guido Falsi via freebsd-ports wrote: On 24/01/21 20:35, Russell L. Carter wrote: Greetings, I am completely ignorant here and am looking for up to date advice on how to get poudriere to build and make available package sets from multiple ports trees. I see there is a port "portshaker" that seems to do much of what I want. [...] BTW I noticed poudriere performs shallow clones for git repos, so it should not use up a lot of disk space. Why not using directly overlays, it will simplify everything ;) I don't know if you read me reply or not - I am using poudriere with ports overlay but have a problem with it. Poudriere options does not take overlay in to account so ports options cannot be configured for overlayed ports which do not exist in the base three. Is there a way to fix it / should I file a PR for it? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: poudriere merging multiple ports trees
On 24/01/2021 20:35, Russell L. Carter wrote:
Greetings,
I am completely ignorant here and am looking for up to
date advice on how to get poudriere to build and make
available package sets from multiple ports trees. I
see there is a port "portshaker" that seems to do much
of what I want.
[...]
Ports and poudriere supports "overlay" for some time (it was added in
the last year)
So you can use default ports tree plus your own tree with own ports or
modified existing ports and build packages from both of them.
Both ports trees must be registred in poudriere. In this example it will
be "default" and "mfh_overlay". The second contains my modifications.
# poudriere ports -l
PORTSTREE METHODTIMESTAMP PATH
default svn+https 2021-01-15 20:23:10 /vol0/poudriere/ports/default
headtestportsnap 2020-03-10 12:37:55 /vol0/poudriere/ports/headtest
mfh_overlay svn+https 2019-11-04 16:33:04 /vol0/poudriere/ports/mfh_overlay
And the you can run "poudriere bulk" command with -O and the name of
your "overlay" tree - "mfh_overlay" in this case ("monitoring" is the
name of my packages set used only on machines doing monitoring stuff)
# poudriere bulk -v -j 11_4_amd64 -z monitoring -p default -O
mfh_overlay -f /usr/local/etc/poudriere.d/pkglists/monitoring
But there is one problem. "poudriere bulk" supports overlay with -O but
"poudriere options" not. It means you cannot run "poudriere options" to
configure options for both trees. Poudriere ends with error about
nonexistent ports origin if it exists only in overlay tree and not in
the default tree.
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
FreeBSD Port: net-im/signal-cli
Hello, I read your comment on revision 549450 - BROKEN FreeBSD 11 amd64. I really would like to build it and run it on my desktop FreeBSD 11.4 amd64. Is there some way to debug it and fix it for 11.4? I am building all my packages in poudriere, I can test something but I don't know what and how. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: MariaDB 1.3.27 overwrites customized my.cnf
On 06/12/2020 15:43, Bernard Spil wrote: On 2020-12-04 12:29, Miroslav Lachman wrote: Am I the only one who sees this (critical) problem after upgrade from MariaDB 10.3.23 to newer version (10.3.25 or 10.3.27)? There is our customized fine-tuned /usr/local/etc/my.cnf for years untouched by pkg install or pkg upgrade. After the last pkg upgrade MariaDB cannot (re)start because my.cnf was replaced with some generic file which contains this: [..] Fixed in mariadb103-client-10.3.27_1 Very sorry for this, that was very bad. I hope you have a backup of your my.cnf! Fortunately yes, we have backup of all config files. Thank you for the fix! Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
MariaDB 1.3.27 overwrites customized my.cnf
Am I the only one who sees this (critical) problem after upgrade from MariaDB 10.3.23 to newer version (10.3.25 or 10.3.27)? There is our customized fine-tuned /usr/local/etc/my.cnf for years untouched by pkg install or pkg upgrade. After the last pkg upgrade MariaDB cannot (re)start because my.cnf was replaced with some generic file which contains this: # # This group is read both by the client and the server # use it for options that affect everything # [client-server] # # include *.cnf from the config directory # !includedir /usr/local/etc/mysql/conf.d But the directory /usr/local/etc/mysql/conf.d is empty. If something silently replaces my config file I would expect it to move my file to proper location which is not the case. My file is simply replaced and configuration of MariaDB is lost and daemon cannot be (re)started any more. I think this is POLA, should be mentioned in UPDATING and pkg-message. The only way to make it work again is restore my.cnf from backup. I filled bug report as PR 251550 https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=251550 It seems really critical to me but I am surprised nobody else reported this. This is on FreeBSD 11.4 amd64 with packages built in our Poudriere Options: CONNECT_EXTRA : off DOCS : off GSSAPI_BASE: off GSSAPI_HEIMDAL : off GSSAPI_MIT : off GSSAPI_NONE: on INNOBASE : on LZ4: on LZO: on MROONGA: off MSGPACK: off OQGRAPH: off ROCKSDB: off SNAPPY : off SPHINX : on SPIDER : on TOKUDB : off WSREP : on ZMQ: off ZSTD : off Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
sysutils/zfsnap2 fails to build - committer needed
zfsnap2 is unbuildable if Bash or Zsh option is swtiched off. There is a patch for pkg-plist to fix this. Can somebody commit this fix, please? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: poudriere overlays build failed with xargs: illegal option -- e
On 2020-05-15 19:36, Bryan Drewery wrote:
===> Patching for python27-2.7.18
xargs: illegal option -- e
usage: xargs [-0opt] [-E eofstr] [-I replstr [-R replacements] [-S
replsize]]
[-J replstr] [-L number] [-n number [-x]] [-P maxprocs]
[-s size] [utility [argument ...]]
*** Error code 1
I suspect somewhere ${REINPLACE_CMD} or ${SED} or
${FRAMEWORK_REINPLACE_CMD} is misspelled or not set.
Perhaps your overlay is outdated and one of these is not set.
You could try this in your poudriere.conf to get more debugging (untested):
PORT_FLAGS="-dlx"
I completely removed Mk directory from overlay and now it works. I tried
to diff old Mk with new Mk but didn't find the difference in REINPLACE
commands.
Now I am building 200 ports. I will try PORT_FLAGS next time.
Thank you!
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
poudriere overlays build failed with xargs: illegal option -- e
I am trying to use port overlays with poudriere-devel. I have main tree with old quaterly branch from December 2019 (default). The overlay contains just a bunch of ports from current tree with updated versions of Apache, Python etc. and Mk from current ports tree. This is called mfh_overlay. I need this because I need to maintain some old SW not existing in current ports tree (phantomjs with old openssl for example) Some ports are building fine but when port needs to be patched it is failing: =>> Status lang/python27 | python27-2.7.18: extract-depends === === =>> Status lang/python27 | python27-2.7.18: extract === ===> NOTICE: This port is deprecated; you may wish to reconsider installing it: EOLed upstream. It is scheduled to be removed on or after 2020-12-31. ===> python27-2.7.18 depends on file: /usr/local/sbin/pkg - found ===> Fetching all distfiles required by python27-2.7.18 for building ===> Extracting for python27-2.7.18 => SHA256 Checksum OK for python/Python-2.7.18.tar.xz. === =>> Status lang/python27 | python27-2.7.18: patch-depends === === =>> Status lang/python27 | python27-2.7.18: patch === ===> Patching for python27-2.7.18 xargs: illegal option -- e usage: xargs [-0opt] [-E eofstr] [-I replstr [-R replacements] [-S replsize]] [-J replstr] [-L number] [-n number [-x]] [-P maxprocs] [-s size] [utility [argument ...]] *** Error code 1 Stop. make: stopped in /overlays/mfh_overlay/lang/python27 build of lang/python27 | python27-2.7.18 ended at Fri May 15 15:14:08 CEST 2020 build time: 00:00:03 !!! build failure encountered !!! [00:00:13] Error: Build failed in phase: patch [00:00:13] Logs: /vol0/poudriere/data/logs/bulk/11_3_amd64-default-php71m103/2020-05-15_15h13m56s [00:00:13] WWW: http://pkg.codelab.cz/build.html?mastername=11_3_amd64-default-php71m103=2020-05-15_15h13m56s [00:00:13] Cleaning up Does anybody know what is wrong with xargs and illegal option -e? I guess it is something in Mk but I don't know what is different in the current tree and the old one from December. I did my last successfull build few weeks ago where the the ports trees were manually merged in to one. But is has some other problems so I am trying to use overlays. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Using pkg in documentation
On 2020-04-27 09:10, Baptiste Daroussin wrote: On Mon, Apr 27, 2020 at 08:23:09AM +0200, Miroslav Lachman wrote: On 2020-04-26 21:37, Muhammad Moinur Rahman wrote: What is the way of mentioning about installing a py-package in documentation? Let’s say now the default version of python is 3.7 so in most of the cases we can write in our documentation that do the following: # pkg install py37-babel You can use following form too: pkg install devel/py-babel Which would install all flavor of py*-babel imho thta is wrong You are right. My bad. I am using it with private repo where only one flavor exists. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Using pkg in documentation
On 2020-04-26 21:37, Muhammad Moinur Rahman wrote: What is the way of mentioning about installing a py-package in documentation? Let’s say now the default version of python is 3.7 so in most of the cases we can write in our documentation that do the following: # pkg install py37-babel You can use following form too: pkg install devel/py-babel Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Poudriere error: cannot rollback ..., there is a snapshot ... more recent ....
George Hartzell wrote on 2020/03/20 18:13: I have a FreeBSD 12.1p3 (just updated) root-on-zfs system that's having trouble doing Poudriere builds. It started off life as a FreeBSD 12.1 system and has been building it's own set of packages using Poudriere successfully since its inception. I few updates ago I started seeing errors in the Poudriere builds. I'm trying to run one now and am seeing the errors again. Here's an example: [...] [00:30:24] [02] [00:00:00] Building devel/py-more-itertools@py37 | py37-more-itertools-8.2.0 cannot rollback 'zroot/poudriere/jails/12_1-ports_and_hartzell-ref/02': there is a snapshot or bookmark more recent than 'zroot/poudriere/jails/12_1-ports_and_hartzell-ref/02@prepkg' [00:30:26] [02] [00:00:02] Error: Unable to rollback zroot/poudriere/jails/12_1-ports_and_hartzell-ref/02 to prepkg =>> Error: Unable to rollback zroot/poudriere/jails/12_1-ports_and_hartzell-ref/02 to prepkg [...] After crashing, I can hit ^T and see: [...] load: 1.01 cmd: sh 36519 [piperd] 1994.36r 0.08u 0.13s 0% 3188k [12_1-ports_and_hartzell] [2020-03-20_09h30m34s] [parallel_build:] Queued: 149 Built: 72 Failed: 1 Skipped: 4 Ignored: 0 Tobuild: 72 Time: 00:33:31 [01]: crashed [02]: devel/p5-IO-stringy | p5-IO-stringy-2.111 package (00:00:00 / 00:00:06) [00:33:33] Logs: /usr/local/poudriere/data/logs/bulk/12_1-ports_and_hartzell/2020-03-20_09h30m34s [...] a bit later, that builder seems to recover and continue sucessfully with new jobs: [...] load: 3.40 cmd: sh 36519 [piperd] 2282.32r 0.09u 0.15s 0% 3188k [12_1-ports_and_hartzell] [2020-03-20_09h30m34s] [parallel_build:] Queued: 149 Built: 80 Failed: 3 Skipped: 10 Ignored: 0 Tobuild: 56 Time: 00:38:19 [01]: lang/cython@py27 | py27-cython-0.29.15 build (00:01:00 / 00:01:12) [02]: devel/icu | icu-66.1,1build (00:03:10 / 00:03:24) [00:38:21] Logs: /usr/local/poudriere/data/logs/bulk/12_1-ports_and_hartzell/2020-03-20_09h30m34s [...] I'd like to get it working normally again. Any ideas what might be going wrong? Any possibility that you have some kind of automatic snapshoting utility run from crontab recursively on filesystem where poudriere have builder jails mounted? I saw this error when I setup zfsnap in crontab with resursive snapshot of /vol0 (each 5 minutes) and my poudriere is in /vol0/poudriere. The workaround was to not create recursive snapshots on /vol0 but on each /vol0 sub-filesystem except /vol0/poudriere. I don't know it this is your problem too. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: [RFC] Adding a Rados block driver to bhyve
Willem Jan Withagen wrote on 2020/03/09 11:31: [...] 3) Create a bhyve-rbd port. Problem with that is that it will require the FreeBSD source tree for the bhyve sources, but there is no Ports option for that? Or bhyve sources are manually copied into the port. And then try to keep these sources up to date. Then compile and install a bhyve-rbd into /usr/local/sbin There are some ports (for example sysutils/lsof) which need kernel sources to build. So this can be a way too. I cannot say if this is the best way or not. [...] Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: ports-mgmt/poudriere-devel does not follow options
Martin Neubauer wrote on 2020/03/01 02:00: On 01/03/2020 00:29, Miroslav Lachman wrote: Martin Neubauer wrote on 2020/02/29 23:44: [...] But I am talking about "poudriere options" taking different saved options than "poudriere bulk". The "poudriere options" does not show me any dialog because they are all set in "default-php71m103-options" and then "poudriere bulk" ignore settings in "default-php71m103-options". That is a bug from my point of view. Both commands should work with the same set of stored options. Otherwise if "poudriere bulk" wants options from "11_3_amd64-default-php71m103-options" I am not able to set those options by calling "poudriere options". If the issued command lines were in fact those from your earlier mail, then you did in fact request two different option sets. If you also add "-j 11_3_amd64" to the "poudriere options" call, you should access the same option set the "poudriere bulk" run did. I migrated from "poudriere" to "poudriere-devel" about week ago that's when I noticed this problem. That's why I suspect poudriere-devel. Everything worked for me for years with poudriere. I search my command history and I never called poudriere options with -j It was like poudriere options -z php71m103 -p default -c mail/dovecot graphics/py-pillow www/nginx poudriere options -z php71m103 -p default -c ports-mgmt/poudriere-devel poudriere options -z php71m103 -p default -f /usr/local/etc/poudriere.d/pkglists/php71m103 poudriere bulk -j 11_3_amd64 -z php71m103 -p default -c -f /usr/local/etc/poudriere.d/pkglists/php71m103 I don't know when / what command exactly created 11_3_amd64-default-php71m103-options in disk. You are right that anything after that is "expected behaviour", a bit unexpected to me. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: ports-mgmt/poudriere-devel does not follow options
Martin Neubauer wrote on 2020/02/29 23:44: On 28/02/2020 17:20, Miroslav Lachman wrote: I am using Poudriere for a long time. I switched to poudriere-devel few days ago because I want to test ports overlay. I run poudriere options -z php71m103 -p default -f /usr/local/etc/poudriere.d/pkglists/php71m103 just to be sure everything is set. No options dialog appeared because all options was set few week ago on regular update. Now I am running poudriere bulk -j 11_3_amd64 -z php71m103 -p default -c -f /usr/local/etc/poudriere.d/pkglists/php71m103 [snip] I tried to find what is going on and I found that all ports are built with default options instead of what I have stored in /usr/local/etc/poudriere.d/default-php71m103-options Just a wild guess... "poudriere options" read options from default-php71m103-options but "poudriere bulk" are trying to read 11_3_amd64-default-php71m103-options. Is it possible? It isn't such a wild guess at all, considering that documentation does state that the most specific set of options gets used with higher priority. More details at: But I am talking about "poudriere options" taking different saved options than "poudriere bulk". The "poudriere options" does not show me any dialog because they are all set in "default-php71m103-options" and then "poudriere bulk" ignore settings in "default-php71m103-options". That is a bug from my point of view. Both commands should work with the same set of stored options. Otherwise if "poudriere bulk" wants options from "11_3_amd64-default-php71m103-options" I am not able to set those options by calling "poudriere options". Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: good gui bit-torrent client?
Robert Huff wrote on 2020/02/29 00:49: I used to use azureus/vuze, but it hasn't been maintained is quite a while. So I changed to deluge ... which now has a dependency semi-permanently BROKEN. What can people recommend as a replacement? I used uTurrent in Windows times. When I switched to FreeBSD on my desktop I used Vuze / Azureus. But it was resource hungry and at some point in time no longer works for me. Then I tried qBittorrent and I am very happy with it. Simple, stable, good looking with my KDE. net-p2p/qbittorrent is my choice Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
ports-mgmt/poudriere-devel does not follow options
I am using Poudriere for a long time. I switched to poudriere-devel few days ago because I want to test ports overlay. I run poudriere options -z php71m103 -p default -f /usr/local/etc/poudriere.d/pkglists/php71m103 just to be sure everything is set. No options dialog appeared because all options was set few week ago on regular update. Now I am running poudriere bulk -j 11_3_amd64 -z php71m103 -p default -c -f /usr/local/etc/poudriere.d/pkglists/php71m103 There were some error on ports like "You are using OpenSSL from ports and have selected GSSAPI from base, please select another GSSAPI value" It is strange because all ports have GSSAPI set to NONE from the beginning. No changes done. I tried to find what is going on and I found that all ports are built with default options instead of what I have stored in /usr/local/etc/poudriere.d/default-php71m103-options Just a wild guess... "poudriere options" read options from default-php71m103-options but "poudriere bulk" are trying to read 11_3_amd64-default-php71m103-options. Is it possible? For example From the log of cyrus-sasl built few weeks ago: ---Begin OPTIONS List--- ===> The following configuration options are available for cyrus-sasl-2.1.27: ALWAYSTRUE=off: Alwaystrue password verifier (discouraged) AUTHDAEMOND=on: Use of authdaemon DOCS=off: Build and/or install documentation KEEP_DB_OPEN=off: Keep handle to Berkeley DB open OBSOLETE_CRAM_ATTR=off: cmusaslsecretCRAM-MD5 auxprop property OBSOLETE_DIGEST_ATTR=on: cmusaslsecretDIGEST-MD5 auxprop property > Options available for the group PLUGIN ANONYMOUS=off: ANONYMOUS authentication CRAM=off: CRAM-MD5 authentication DIGEST=off: DIGEST-MD5 authentication LOGIN=on: LOGIN authentication NTLM=off: NTLM authentication OTP=off: OTP authentication PLAIN=on: PLAIN authentication SCRAM=off: SCRAM authentication > SASLdb auxprop plugin: you can only select none or one of them BDB1=on: Berkeley DB 1.85 support BDB=off: Berkeley DB support GDBM=off: GNU dbm library support LMDB=off: OpenLDAP Lightning Memory-Mapped Database support ===> Use 'make config' to modify these settings ---End OPTIONS List--- From the log of cyrus-sasl built today ---Begin OPTIONS List--- ===> The following configuration options are available for cyrus-sasl-2.1.27: ALWAYSTRUE=off: Alwaystrue password verifier (discouraged) AUTHDAEMOND=on: Use of authdaemon DOCS=off: Build and/or install documentation KEEP_DB_OPEN=off: Keep handle to Berkeley DB open OBSOLETE_CRAM_ATTR=on: cmusaslsecretCRAM-MD5 auxprop property OBSOLETE_DIGEST_ATTR=on: cmusaslsecretDIGEST-MD5 auxprop property > Options available for the group PLUGIN ANONYMOUS=on: ANONYMOUS authentication CRAM=on: CRAM-MD5 authentication DIGEST=on: DIGEST-MD5 authentication LOGIN=on: LOGIN authentication NTLM=on: NTLM authentication OTP=on: OTP authentication PLAIN=on: PLAIN authentication SCRAM=on: SCRAM authentication > SASLdb auxprop plugin: you can only select none or one of them BDB1=on: Berkeley DB 1.85 support BDB=off: Berkeley DB support GDBM=off: GNU dbm library support LMDB=off: OpenLDAP Lightning Memory-Mapped Database support ===> Use 'make config' to modify these settings ---End OPTIONS List--- Both built with the same commands (the same poudriere options & poudriere bulk) Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: About protocols in openssl
Willem Jan Withagen wrote on 2020/02/27 20:00: Hi, My ceph ports uses all kinds of python stuff, and now the trouble is that I'm getting an error on missing: SSLv3_client_method Which i guess, is because in the current openssl libs SSLv3 is disabled. And I sort of get this, SSLv3 is unsafe. But I need it to be able to run parts of the ceph port. So how do I get a openssl lib dependancy that has SSLv3 enabled. You can build OpenSSL 1.1.1 from the ports where you can enable SSLv3 in the options dialog. https://www.freshports.org/security/openssl/ The defaults are: > Protocol Support NEXTPROTONEG=on: Next Protocol Negotiation (SPDY) SCTP=on: SCTP (Stream Control Transmission) SSL3=off: SSLv3 (unsafe) TLS1=on: TLSv1.0 (requires TLS1_1, TLS1_2) TLS1_1=on: TLSv1.1 (requires TLS1_2) TLS1_2=on: TLSv1.2 Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Cannot build qt5-webkit with debug
Marcin Cieslak wrote on 2020/02/26 13:50:
On Wed, 26 Feb 2020, Miroslav Lachman wrote:
And what is not working now with the line fontFamilyList->append...
removed()?
Should I report it upstream or is it FreeBSD specific problem?
I think it is a bit too early to tell. I think it is not really
FreeBSD-specific
but there might be some quirks.
I am rebuilding qt5-webkit now with debug symbols using your phantomfs
port (thanks!) and I'll poke around a bit with debugger.
Just the small update.
phantomjs with modified version of qt5-webkit does not render any text /
fonts on the web page screenshots taken by
page.render('/tmp/page.png');
So... it builds, it does not crashes but rendered pages are "empty"
(pictures, background colors etc are rendered OK, but no fonts)
Kind regards
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Cannot build qt5-webkit with debug
Michael Osipov wrote on 2020/02/25 11:16: Am 2020-02-24 um 22:52 schrieb Miroslav Lachman: [...] I tried it on the simplest web page without any fonts, css etc. but it is still crashing. I removed the charset from the page so it is now not using ICU. https://hastebin.com/iyonapasuz.bash Let me know if I can try anything else. OK, it is getting closer. The crash comes from this call: https://github.com/qtwebkit/qtwebkit/blob/qtwebkit-5.212.0-alpha3/Source/WebCore/css/CSSParser.cpp#L6907 Can you commet comment this line, recompile qtwebkit and see what happens? I am sorry for my late reply. Rebuilding of qt5-webkit is very strange. The build crashes with out of memory 3 times but finally I have qt5-webkit rebuilt with commented out line you suggested and it works now! The example from my first attempt works: (root@testjail) ~/# phantomjs /tmp/phantom.2.js [blocked] The page at https://phantomjs.org/api/phantom/ was not allowed to run insecure content from http://fonts.googleapis.com/css?family=Droid+Sans:400,700. https://phantomjs.org/api/phantom/ element is [object Object] https://phantomjs.org/ I tried few more examples without crash so far. This is the exact modification patch for qt5-webkit # cat /vol0/poudriere/ports/headtest/www/qt5-webkit/files/patch-Source_WebCore_css_CSSParser.cpp --- Source/WebCore/css/CSSParser.cpp.orig 2019-06-26 16:25:02 UTC +++ Source/WebCore/css/CSSParser.cpp @@ -6904,7 +6904,7 @@ void CSSParser::parseSystemFont(bool imp addProperty(CSSPropertyFontWeight, CSSValuePool::singleton().createValue(fontDescription.weight()), important); addProperty(CSSPropertyFontSize, CSSValuePool::singleton().createValue(fontDescription.specifiedSize(), CSSPrimitiveValue::CSS_PX), important); Ref fontFamilyList = CSSValueList::createCommaSeparated(); - fontFamilyList->append(CSSValuePool::singleton().createFontFamilyValue(fontDescription.familyAt(0), FromSystemFontID::Yes)); + //fontFamilyList->append(CSSValuePool::singleton().createFontFamilyValue(fontDescription.familyAt(0), FromSystemFontID::Yes)); addProperty(CSSPropertyFontFamily, WTFMove(fontFamilyList), important); addProperty(CSSPropertyFontVariantCaps, CSSValuePool::singleton().createIdentifierValue(CSSValueNormal), important); addProperty(CSSPropertyLineHeight, CSSValuePool::singleton().createIdentifierValue(CSSValueNormal), important); I am very curious what is wrong here? qt5-webkit is used by 60+ ports as a dependency so why only phantomjs crashes with it? And what is not working now with the line fontFamilyList->append... removed()? Should I report it upstream or is it FreeBSD specific problem? Again, thank you so much for your help! Without it I will not be able to fix it. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Cannot build qt5-webkit with debug
Michael Osipov wrote on 2020/02/25 11:16:
Am 2020-02-24 um 22:52 schrieb Miroslav Lachman:
Michael Osipov wrote on 2020/02/24 21:40:
Am 2020-02-24 um 20:56 schrieb Miroslav Lachman:
[..]
I tried it on simpler website on HTTP without external fonts etc. but
it is still crashing
Example JS code and truss output is on hastebin
https://hastebin.com/sizefupiki.pl
No fopen found there :(
Can you also please provide the backtrace from lldb for this core dump?
lldb backtrace
https://hastebin.com/ovufejevil.bash
Surprisingly, the last call is:
stat("/usr/local/share/icu/65.1/icudt65l/iso-8859_16-2001.cnv",0x7fffc998)
which comes from ICU. The entire website has been read from the
socket, as you can see from the read in 2246.
Maybe it to convert the "charset=iso-8859-2" to LC_ALL value.
I would recommend to have a simple server built in PHP or Python and
try the simplest site possible w/o CSS and then add one w/o fonts first.
I tried it on the simplest web page without any fonts, css etc. but it
is still crashing. I removed the charset from the page so it is now not
using ICU.
https://hastebin.com/iyonapasuz.bash
Let me know if I can try anything else.
OK, it is getting closer.
The crash comes from this call:
https://github.com/qtwebkit/qtwebkit/blob/qtwebkit-5.212.0-alpha3/Source/WebCore/css/CSSParser.cpp#L6907
Can you commet comment this line, recompile qtwebkit and see what happens?
I commented out the line number 6907 and rebuilding qt5-webkit now but
it will take about 3 hours.
Either this line:
https://github.com/qtwebkit/qtwebkit/blob/qtwebkit-5.212.0-alpha3/Source/WebCore/css/CSSParser.cpp#L6897
does not work properly or this is broken:
https://github.com/qtwebkit/qtwebkit/blob/qtwebkit-5.212.0-alpha3/Source/WebCore/css/CSSValuePool.cpp#L121-L134
I'd like to see to see the content of familyName when passed. Either via
breakpoint or with std::cout. Unfortunately, my C++ knowledge is
minimal. I recommend to write a minimal C++ app with a function call and
signature "const String& str" and then pass the string [1]. Since this
is a custom type, I don't know whether cout will work. You have to try.
Michael
[1]
https://github.com/adobe/webkit/blob/master/Source/WTF/wtf/text/WTFString.h
It is out of my knowledge.
I will try it with newly rebuilt qt5-webkit and report soon. Then I will
try to find how to track it with breakpoints in debuger.
Thank you very much. I really appreciate your help.
Kind regards
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Cannot build qt5-webkit with debug
Michael Osipov wrote on 2020/02/24 21:40:
Am 2020-02-24 um 20:56 schrieb Miroslav Lachman:
[..]
I tried it on simpler website on HTTP without external fonts etc. but
it is still crashing
Example JS code and truss output is on hastebin
https://hastebin.com/sizefupiki.pl
No fopen found there :(
Can you also please provide the backtrace from lldb for this core dump?
lldb backtrace
https://hastebin.com/ovufejevil.bash
Surprisingly, the last call is:
stat("/usr/local/share/icu/65.1/icudt65l/iso-8859_16-2001.cnv",0x7fffc998)
which comes from ICU. The entire website has been read from the socket,
as you can see from the read in 2246.
Maybe it to convert the "charset=iso-8859-2" to LC_ALL value.
I would recommend to have a simple server built in PHP or Python and try
the simplest site possible w/o CSS and then add one w/o fonts first.
I tried it on the simplest web page without any fonts, css etc. but it
is still crashing. I removed the charset from the page so it is now not
using ICU.
https://hastebin.com/iyonapasuz.bash
Let me know if I can try anything else.
Kind regards
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Cannot build qt5-webkit with debug
Michael Osipov wrote on 2020/02/24 20:35:
Am 2020-02-24 um 19:10 schrieb Miroslav Lachman:
Looking at your sample code and the website, there are several issues:
* The resource employs mixed site content HTTPS loads HTTP. Firefox
blocks this, I guess WebKit blocks it too.
* Result: Droid Sans is not loaded [1], like the CSS
* The CSS has this:
body {
font-family: 'Droid Sans','Lucida Grande',sans-serif;
line-height: 1.125em;
color: #444; }
Lucida Grande isn't there either. It must choke somewhere.
See also
frame #10: 0x0008039cf92c
libQt5WebKit.so.5`WebCore::CSSParser::parseSystemFont
I assume it does some fopen() and chokes on when looking for sans-serif
replacement.
Can you run the application with truss? We should see the fopen() and
shortly after the crash. Upload the truss output.
I tried it on simpler website on HTTP without external fonts etc. but it
is still crashing
Example JS code and truss output is on hastebin
https://hastebin.com/sizefupiki.pl
No fopen found there :(
I installed webfonts in the meantime without any effect.
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Cannot build qt5-webkit with debug
Marcin Cieslak wrote on 2020/02/24 18:51:
On Mon, 24 Feb 2020, Marcin Cieslak wrote:
On Mon, 24 Feb 2020, Miroslav Lachman wrote:
frame #13: 0x0008039aa3ed
libQt5WebKit.so.5`WebCore::CSSParser::parseSheet(this=0x7fffa490,
sheet=0x0008155f5e40, string=0x7fffb888,
textPosition=0x7fffb650,
ruleSourceDataResult=0x, logErrors=false) at
CSSParser.cpp:423:5
Looks like it has some trouble parsing font declarations in the
HTML/CSS code you have been testing
on. This can be a webkit bug...
Would be good to have a minimal HTML/CSS testcase to reproduce this
and this should
probably be reported upstream.
There is this report:
https://github.com/qtwebkit/qtwebkit/issues/933
It says it crashes if you have no fonts installed (no TTF).
Could this "fix" it?
Thank you for the tip, I will look at it.
The following fonts related packages are installed:
(root@testjail) ~/# pkg info | grep -i font
dejavu-2.37_1 Bitstream Vera Fonts clone with a wider
range of characters
encodings-1.0.5,1 X.Org Encoding fonts
font-bh-ttf-1.0.3_4X.Org Bigelow & Holmes TTF font
font-misc-ethiopic-1.0.3_4 X.Org miscellaneous Ethiopic font
font-misc-meltho-1.0.3_4 X.Org miscellaneous Meltho font
fontconfig-2.12.6,1XML-based font configuration API for X
Windows
freetype2-2.10.1 Free and portable TrueType font rendering
engine
libfontenc-1.1.4 The fontenc Library
mkfontscale-1.2.1 Creates an index of scalable font files for X
woff2-1.0.2_4 Library and converter tools for the WOFF
2.0 web font format
xorg-fonts-truetype-7.7_1 X.Org TrueType fonts
This is the code of the tested example - fetching from
https://phantomjs.org/
(root@testjail) ~/# cat /tmp/phantom.2.js
var page = require('webpage').create();
page.onConsoleMessage = function(str) {
console.log(str);
}
page.open('https://phantomjs.org/api/phantom/', function(status) {
page.render('/tmp/beforeclick.png');
console.log(page.url);
var element = page.evaluate(function() {
return document.querySelector('img[src =
"https://phantomjs.org/img/phantomjs-logo.png;]');
});
page.sendEvent('click', element.offsetLeft, element.offsetTop, 'left');
window.setTimeout(function () {
console.log(page.url);
page.render('/tmp/afterclick.png');
phantom.exit();
}, 5000);
console.log('element is ' + element);
});
Kind regards
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Cannot build qt5-webkit with debug
Michael Osipov wrote on 2020/02/24 16:47: Am 2020-02-24 um 16:35 schrieb Miroslav Lachman: Loaded symbols for /usr/lib/libdl.so.1 Reading symbols from /usr/local/lib/qt5/libQt5WebKitWidgets.so.5...done. Loaded symbols for /usr/local/lib/qt5/libQt5WebKitWidgets.so.5 Reading symbols from /usr/local/lib/qt5/libQt5Widgets.so.5...done. Loaded symbols for /usr/local/lib/qt5/libQt5Widgets.so.5 Reading symbols from /usr/local/lib/qt5/libQt5WebKit.so.5...Segmentation fault (core dumped) From my POV, something is fishy with the shared object libQt5WebKit.so.5. Try objdump/readelf of similiar whether the thing can be read at all. I have the feeling that either the objects are invalid somehow or the linker failed here. You might have found a bug with LLVM infra. Try the latest version of llvm-c++ and lld from ports. Which LLVM version do you use? You might want also try g++ and GNU ld. Michael PS: I am not a C/C++ expect either, but have dealt with many crashes before. It was build on top of ports tree updated at 2020-02-18 with llvm80 i guess. PORTSTREE METHODTIMESTAMP PATH headtestportsnap 2020-02-18 19:21:17 /vol0/poudriere/ports/headtest readelf and objdump can read the libQt5WebKit.so.5 I tired it with some options where the dumped text was too long to paste here so I paste just this short one (root@testjail) ~/# readelf -d /usr/local/lib/qt5/libQt5WebKit.so.5 Dynamic section at offset 0x54ba6c8 contains 53 entries: TagType Name/Value 0x0001 NEEDED Shared library: [libQt5Quick.so.5] 0x0001 NEEDED Shared library: [libQt5WebChannel.so.5] 0x0001 NEEDED Shared library: [libicuuc.so.65] 0x0001 NEEDED Shared library: [libQt5Positioning.so.5] 0x0001 NEEDED Shared library: [libX11.so.6] 0x0001 NEEDED Shared library: [libXcomposite.so.1] 0x0001 NEEDED Shared library: [libXrender.so.1] 0x0001 NEEDED Shared library: [libicui18n.so.65] 0x0001 NEEDED Shared library: [libQt5Sensors.so.5] 0x0001 NEEDED Shared library: [libwoff2dec.so.1.0.2] 0x0001 NEEDED Shared library: [libjpeg.so.8] 0x0001 NEEDED Shared library: [libpng16.so.16] 0x0001 NEEDED Shared library: [libz.so.6] 0x0001 NEEDED Shared library: [libwebp.so.7] 0x0001 NEEDED Shared library: [libhyphen.so.0] 0x0001 NEEDED Shared library: [libxml2.so.2] 0x0001 NEEDED Shared library: [libxslt.so.1] 0x0001 NEEDED Shared library: [libsqlite3.so.0] 0x0001 NEEDED Shared library: [libQt5Gui.so.5] 0x0001 NEEDED Shared library: [libQt5Qml.so.5] 0x0001 NEEDED Shared library: [libQt5Network.so.5] 0x0001 NEEDED Shared library: [libQt5Core.so.5] 0x0001 NEEDED Shared library: [libc++.so.1] 0x0001 NEEDED Shared library: [libcxxrt.so.1] 0x0001 NEEDED Shared library: [libm.so.5] 0x0001 NEEDED Shared library: [libgcc_s.so.1] 0x0001 NEEDED Shared library: [libthr.so.3] 0x0001 NEEDED Shared library: [libc.so.7] 0x000e SONAME Library soname: [libQt5WebKit.so.5] 0x000f RPATHLibrary rpath: [/usr/local/lib/qt5:/usr/local/lib] 0x001d RUNPATH Library runpath: [/usr/local/lib/qt5:/usr/local/lib] 0x000c INIT 0x5adfa8 0x000d FINI 0x3eb1498 0x0004 HASH 0x190 0x6ef5 GNU_HASH 0x12d30 0x0005 STRTAB 0x65b98 0x0006 SYMTAB 0x257a0 0x000a STRSZ518712 (bytes) 0x000b SYMENT 24 (bytes) 0x0003 PLTGOT 0x56c11d8 0x0002 PLTRELSZ 148872 (bytes) 0x0014 PLTREL RELA 0x0017 JMPREL 0x589a20 0x0007 RELA 0xea0a0 0x0008 RELASZ 4848000 (bytes) 0x0009 RELAENT 24 (bytes) 0x6ffc VERDEF 0xe9b80 0x6ffd VERDEFNUM17 0x6ffe VERNEED 0xe9dd0 0x6fff VERNEEDNUM 18 0x6ff0 VERSYM 0xe45d0 0x6ff9 RELACOUNT154470 0x NULL 0x0 I tried lldb
Re: Cannot build qt5-webkit with debug
Michael Osipov wrote on 2020/02/24 16:00: Am 2020-02-24 um 15:50 schrieb Miroslav Lachman: Miroslav Lachman wrote on 2020/02/24 12:48: [..] There are 2 core dumps -rw--- 1 root wheel 327M Feb 24 14:36 gdb.core -rw--- 1 root wheel 4.9M Feb 24 14:36 phantomjs.core What more can I try? I am not a C / C++ skilled person. All these things are new to me. You can load the coredump into GDB and say "where" [1]. It will tell you the last frame it executed before it crashed. Make sure that your objects contains debug symbols. This mostly helped me to get a grasp of failures on FreeBSD and HP-UX. Thank you! I tried to debug gdb.core with this result (root@testjail) ~/# gdb /usr/bin/gdb gdb.core GNU gdb 6.1.1 [FreeBSD] Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "amd64-marcel-freebsd"... Core was generated by 'gdb /usr/local/bin/phantomjs'. Program terminated with signal 11, Segmentation fault. Reading symbols from /lib/libm.so.5...Reading symbols from /usr/lib/debug//lib/libm.so.5.debug...done. done. Loaded symbols for /lib/libm.so.5 Reading symbols from /lib/libncursesw.so.8...Reading symbols from /usr/lib/debug//lib/libncursesw.so.8.debug...done. done. Loaded symbols for /lib/libncursesw.so.8 Reading symbols from /usr/lib/libgnuregex.so.5...Reading symbols from /usr/lib/debug//usr/lib/libgnuregex.so.5.debug...done. done. Loaded symbols for /usr/lib/libgnuregex.so.5 Reading symbols from /lib/libc.so.7...Reading symbols from /usr/lib/debug//lib/libc.so.7.debug...done. done. Loaded symbols for /lib/libc.so.7 Reading symbols from /usr/lib/libthread_db.so...Reading symbols from /usr/lib/debug//usr/lib/libthread_db.so.3.debug...done. done. Loaded symbols for /usr/lib/libthread_db.so Reading symbols from /libexec/ld-elf.so.1...Reading symbols from /usr/lib/debug//libexec/ld-elf.so.1.debug...done. done. Loaded symbols for /libexec/ld-elf.so.1 #0 0x000801327fd6 in memcpy () from /lib/libc.so.7 (gdb) where #0 0x000801327fd6 in memcpy () from /lib/libc.so.7 #1 0x0008012fcfab in __fread (buf=0x8163f6704, size=1, count=2391484440, fp=0x80158a698) at /usr/src/lib/libc/stdio/fread.c:101 #2 0x0008012fceb3 in fread (buf=0x815d32940, size=1, count=2391484440, fp=0x80158a698) at /usr/src/lib/libc/stdio/fread.c:58 #3 0x005671c0 in cache_bread () #4 0x0059d8f8 in bfd_bread () #5 0x005356c7 in dwarf2_read_section () #6 0x00535080 in dwarf2_build_psymtabs () #7 0x005297a2 in elf_symfile_read () #8 0x004d233a in syms_from_objfile () #9 0x004d267b in symbol_file_add_with_addrs_or_offsets () #10 0x004d2479 in symbol_file_add () #11 0x0048f9f4 in symbol_add_stub () #12 0x004ac696 in catcher () #13 0x004ac830 in catch_errors () #14 0x0048f6c2 in solib_add () #15 0x0052041b in handle_inferior_event () #16 0x0051f588 in wait_for_inferior () #17 0x0051f3c8 in proceed () #18 0x004acda9 in execute_command () #19 0x00466725 in command_handler () #20 0x0046645a in command_line_handler () #21 0x005e4f0f in rl_callback_read_char () #22 0x004660c9 in rl_callback_read_char_wrapper () #23 0x00466f22 in gdb_do_one_event () #24 0x004ac696 in catcher () #25 0x004ac830 in catch_errors () #26 0x005537e9 in tui_command_loop () #27 0x00448609 in captured_command_loop () #28 0x004ac696 in catcher () #29 0x004ac830 in catch_errors () #30 0x00448446 in captured_main () #31 0x004ac696 in catcher () #32 0x004ac830 in catch_errors () #33 0x00447974 in gdb_main () #34 0x00447931 in main (argc=, argv=optimized out>) at /usr/src/contrib/gdb/gdb/gdb.c:35 When I tried to debug phantomjs.core it crashed again (root@testjail) ~/# gdb /usr/local/bin/phantomjs phantomjs.core GNU gdb 6.1.1 [FreeBSD] Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "amd64-marcel-freebsd"...(no debugging symbols found)... Core was generated by `/usr/local/bin/phantomjs /tmp/phantom.2.js'. Program terminated with signal 5, Trace/breakpoint trap. Reading symbols from /usr/lib/libdl.so.1...Reading symbols from /usr/lib/debug//usr/lib/libdl.so.1.debug...done. done. Loade
Re: Cannot build qt5-webkit with debug
Miroslav Lachman wrote on 2020/02/24 12:48: Short story: I am trying to build qt5-webkit with WITH_DEBUG=yes in make.conf on our E3 Xeon machine with FreeBSD 11.3, poudriere-devel, 16GB of RAM and 10GB of swap. The build always hangs, machine is unresponsive on SSH / HTTP, only ping is responding. I track it down to build eats all memory (RAM + Swap). Did anybody tried to build (successfully?) qt5-webkit with debug option? Is there a bug in build process with WITH_DEBUG or is 16GB of RAM and 10GB of swap not enough in these days? Any help appreciated. Long story: As I didn't found replacement for removed port lang/phantomjs nor found how to build firefox or chromium as static binary I am trying to create new port of lang/phantomjs. I posted few weeks ago - phantomjs github is active again https://lists.freebsd.org/pipermail/freebsd-ports/2020-January/117556.html The new development version of phantomjs has different building process, newer dependencies, can be built with current OpenSSL 1.1.1 etc. My (ugly) temporary version of this port is buildable, resulting binary "phantomjs" can run simple "Hello world" example https://github.com/ariya/phantomjs/blob/master/examples/hello.js But when I try to run some real work (fetching web page) it segfaulted. [...] I guess I need debug symbols for qt5-webkit to debug it further but I am not able to build it with WITH_DEBUG. Are there any ways / tuning options to build it without eating all memory and swap space? Meanwhile I added 20GB more of a swap (total of 30GB) and the build succeeded. (according to Zabbix graphs it used 22GB of swap for this build) Phantomjs is still segfaulting and I am not able to debug it with gdb because it is interrupted by segfault to: (root@testjail) ~/# phantomjs /tmp/phantom.2.js PhantomJS has crashed. Please read the bug reporting guide at <http://phantomjs.org/bug-reporting.html> and file a bug report. Segmentation fault (root@testjail) ~/# gdb /usr/local/bin/phantomjs GNU gdb 6.1.1 [FreeBSD] Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "amd64-marcel-freebsd"...(no debugging symbols found)... (gdb) run /tmp/phantom.2.js Starting program: /usr/local/bin/phantomjs /tmp/phantom.2.js Segmentation fault (core dumped) There are 2 core dumps -rw--- 1 root wheel 327M Feb 24 14:36 gdb.core -rw--- 1 root wheel 4.9M Feb 24 14:36 phantomjs.core What more can I try? I am not a C / C++ skilled person. All these things are new to me. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Cannot build qt5-webkit with debug
() from /usr/local/lib/qt5/libQt5WebKit.so.5 #25 0x000802cc9319 in WebCore::FileIconLoader::notifyFinished () from /usr/local/lib/qt5/libQt5WebKit.so.5 #26 0x0008029fb7d3 in WebCore::Element::setAttribute () from /usr/local/lib/qt5/libQt5WebKit.so.5 #27 0x000802cfffa3 in WebCore::DocumentLoader::commitData () from /usr/local/lib/qt5/libQt5WebKit.so.5 #28 0x000801d66a74 in QWebSettings::enablePersistentStorage () from /usr/local/lib/qt5/libQt5WebKit.so.5 #29 0x000802d01fa3 in WebCore::DocumentLoader::responseReceived () from /usr/local/lib/qt5/libQt5WebKit.so.5 #30 0x000802d4c8f2 in WebCore::CachedImage::imageForRenderer () from /usr/local/lib/qt5/libQt5WebKit.so.5 #31 0x000802d4c7d4 in WebCore::CachedImage::imageForRenderer () from /usr/local/lib/qt5/libQt5WebKit.so.5 #32 0x000802d3d7d5 in WebCore::SubresourceLoader::create () from /usr/local/lib/qt5/libQt5WebKit.so.5 #33 0x000802d3d8aa in WebCore::SubresourceLoader::create () from /usr/local/lib/qt5/libQt5WebKit.so.5 #34 0x000802d3995c in WebCore::ResourceLoader::cancelledError () from /usr/local/lib/qt5/libQt5WebKit.so.5 #35 0x000802f4a1f1 in WebCore::proxyServersForURL () from /usr/local/lib/qt5/libQt5WebKit.so.5 #36 0x000802f48b7f in WebCore::proxyServersForURL () from /usr/local/lib/qt5/libQt5WebKit.so.5 #37 0x000802f4a0e6 in WebCore::proxyServersForURL () from /usr/local/lib/qt5/libQt5WebKit.so.5 #38 0x000802f492b6 in WebCore::proxyServersForURL () from /usr/local/lib/qt5/libQt5WebKit.so.5 #39 0x0008052bd241 in QMetaObject::activate () from /usr/local/lib/qt5/libQt5Core.so.5 #40 0x00080441c4a4 in QHttpMultiPart::setBoundary () from /usr/local/lib/qt5/libQt5Network.so.5 #41 0x0008044b5c91 in QHttpMultiPart::qt_metacall () from /usr/local/lib/qt5/libQt5Network.so.5 #42 0x0008052b6348 in QObject::event () from /usr/local/lib/qt5/libQt5Core.so.5 #43 0x000800f79f03 in QApplicationPrivate::notify_helper () from /usr/local/lib/qt5/libQt5Widgets.so.5 #44 0x000800f7b44c in QApplication::notify () from /usr/local/lib/qt5/libQt5Widgets.so.5 #45 0x00080528a212 in QCoreApplication::notifyInternal2 () from /usr/local/lib/qt5/libQt5Core.so.5 #46 0x00080528b2fb in QCoreApplicationPrivate::sendPostedEvents () from /usr/local/lib/qt5/libQt5Core.so.5 #47 0x0008052e32a8 in QEventDispatcherGlib::qt_metacall () from /usr/local/lib/qt5/libQt5Core.so.5 #48 0x00080b772837 in g_main_context_dispatch () from /usr/local/lib/libglib-2.0.so.0 #49 0x00080b772bfa in g_main_context_pending () from /usr/local/lib/libglib-2.0.so.0 #50 0x00080b772cb4 in g_main_context_iteration () from /usr/local/lib/libglib-2.0.so.0 #51 0x0008052e2cb6 in QEventDispatcherGlib::processEvents () from /usr/local/lib/qt5/libQt5Core.so.5 #52 0x0008052858be in QEventLoop::exec () from /usr/local/lib/qt5/libQt5Core.so.5 #53 0x00080528a97e in QCoreApplication::exec () from /usr/local/lib/qt5/libQt5Core.so.5 ---Type to continue, or q to quit--- #54 0x0043d73a in inner_main (argc=2, argv=0x7fffe9c8) at /var/ports/usr/ports/headtest/lang/phantomjs/work/phantomjs-cb1b04b/src/main.cpp:74 #55 0x0043d38d in main (argc=2, argv=0x7fffe9c8) at /var/ports/usr/ports/headtest/lang/phantomjs/work/phantomjs-cb1b04b/src/main.cpp:95 I guess I need debug symbols for qt5-webkit to debug it further but I am not able to build it with WITH_DEBUG. Are there any ways / tuning options to build it without eating all memory and swap space? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Starting with poudriere
Jose Quinteiro wrote on 2020/02/19 20:35: On 2020-02-15 20:43, Dan McGrath wrote: Probably not ideal since you generally want to disable the FreeBSD repository, and use only your poudriere repo, instead. You would need to build everything you intend to install in the jail, however. While I believe that you can enable multiple repositories (FreeBSD's, and your own poudriere one), I am not sure about repo priorities, or how you would deal with conflicts with build options that pull in common ports. It is something I have been meaning to look into, sorry! Perhaps someone else here can give some advice? You can set priorities for remote package repositories in the repo's conf file. The default priority is 0, which is lowest. I plan to set up a repo with the ports I build with poudriere at a higher priority. Are there concerns with this approach? I am using 2 or 3 repositories (all built with our poudriere) for some time. It works but the problem can be if some repo has different default versions. For example one repo has PostgreSQL 9.6 as default version for all dependencies and another repo has PostgreSQL 10. Then you need to carefully choose what to install from what repo. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: package +MANIFEST file - look inside
Yuet-nan Wong via freebsd-ports wrote on 2020/02/11 01:04: Is there a command, or short script that enables the +MANIFEST to be examined. There are cases where we want to review what the scripts section of the +MANIFEST contains. In particular the post-install script. What we're trying to achieve is something like pkg view scripts post-install -f $OUTPUT so we can modify or run it at a better time, like first boot. If you want to modify or run in in a better time then you probably need to know it before running pkg install / pkg upgrade command. Then you probably need to fetch a package alone (pkg upgrade with -f or --fetch-only) and then you can unpack +MANIFEST from it and use some JSON / YAML tool to examine content (e.g. textproc/jq). After that you can run pkg upgrade (or pkg install) again with -I or --no-install-scripts. As a follow-on from this, we would like to use the "files" list to set the MAC mls/ settings as a post-install task, so getting the first, filename element (of filename:signature,) from the "files" list is important. Thank-you. You can get the list of the installed files by "pkg info -l pkgname" or "pkg query '%Fp' pkgname" Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: replacement for lang/phantomjs / minimalistic chrome headless build
Miroslav Lachman wrote on 2019/05/22 14:29: Steven Hartland wrote on 2019/05/22 13:56: Pretty sure chrome has a headless mode now, perhaps you could use that. The last time I tried chrome headless it was so big beast (for headless servers) that I abandoned it. Build time in my poudriere was about 33 hours and so many dependencies I don't want to install in headless machines. So I hope there is still some chance to find something lighter than headless chrome. I would like to continue this old thread. We are still using phantomjs because there is no lightweight replacement. It does not build with OpenSSL 1.1.1, it was removed from the ports tree... so it is time to replace it with something as soon as possible. The question is - Is it possible to build Chromium (or Firefox?) as static with minimum run-time dependencies? Possibly without X11? As we need it on headless server with chromedriver + selenium I really would like to avoid installing tens of dependencies / libraries. Kind regards Miroslav Lachman PS: out of curiosity there is ongoing activity on PhantomJS github in recent months https://github.com/ariya/phantomjs ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
FreeBSD Port: www/iridium
There is version 2019.11.78 available on https://iridiumbrowser.de/ Are there any plans to update port www/iridium to the more current version? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
some ports not building with openssl111 / expiration of openssl
We use self built packages from quaterly ports tree (built with
poudriere). OpenSSL 1.0 is deprecated with expiration date 2019-12-31 so
I tried to switch to openssl111 but some ports cannot be build with it.
lang/erlang-runtime19 (required by couchdb) is marked as broken: does
not build with DEFAULT_VERSIONS+=ssl=openssl111
lang/phantomjs is marked as broken: does not build with
DEFAULT_VERSIONS+=ssl=openssl111 (I know this port expired but we do not
have suitable alternative)
security/sslscan is marked as broken: Does not build with openssl111
They are known to not build, are there any plans to fix sslscan and
erlang-runtime19?
And more to this list - ports not marked as broken but building with
openssl111 failed:
sysutils/rsyncrypto
mail/heirloom-mailx
Tail of a log for sysutils/rsyncrypto
gmake[2]: Entering directory
'/wrkdirs/usr/ports/sysutils/rsyncrypto/work/rsyncrypto-1.12'
c++ -DHAVE_CONFIG_H -I. -I/usr/local/include/ -O2 -pipe
-fstack-protector-strong -fno-strict-aliasing -MT main.o -MD -MP -MF
.deps/main.Tpo -c -o main.o main.cpp
mv -f .deps/main.Tpo .deps/main.Po
c++ -DHAVE_CONFIG_H -I. -I/usr/local/include/ -O2 -pipe
-fstack-protector-strong -fno-strict-aliasing -MT crypto.o -MD -MP -MF
.deps/crypto.Tpo -c -o crypto.o crypto.cpp
crypto.cpp:158:13: error: member access into incomplete type 'RSA' (aka
'rsa_st')
if( (prv->p==0 || prv->q==0) ) {
^
/usr/local/include/openssl/ossl_typ.h:110:16: note: forward declaration
of 'rsa_st'
typedef struct rsa_st RSA;
^
1 error generated.
gmake[2]: *** [Makefile:293: crypto.o] Error 1
gmake[2]: Leaving directory
'/wrkdirs/usr/ports/sysutils/rsyncrypto/work/rsyncrypto-1.12'
Tail of a log for mail/heirloom-mailx
-DHAVE_IPv6_FUNCS -c version.c
cc -Wl,-rpath,/usr/local/lib -fstack-protector-strong aux.o base64.o
cache.o cmd1.o cmd2.o cmd3.o cmdtab.o collect.o dotlock.o edit.o fio.o
getname.o getopt.o head.o hmac.o imap.o imap_search.o junk.o lex.o
list.o lzw.o macro.o maildir.o main.o md5.o mime.o names.o nss.o
openssl.o pop3.o popen.o quit.o send.o sendout.o smtp.o ssl.o strings.o
temp.o thread.o tty.o v7.local.o vars.o version.o `grep '^[^#]' LIBS`
-L/usr/local/lib -o mailx
openssl.o: In function `smime_cipher':
openssl.c:(.text+0x19e2): undefined reference to `EVP_rc2_40_cbc'
openssl.c:(.text+0x19eb): undefined reference to `EVP_rc2_64_cbc'
cc: error: linker command failed with exit code 1 (use -v to see invocation)
*** Error code 1
Stop.
make[1]: stopped in /wrkdirs/usr/ports/mail/heirloom-mailx/work/mailx-12.4
*** Error code 1
If these two are not marked as broken do they build for somebody else?
Kind regards
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
committer needed for security/vuxml SA entries
As always after new base Security Advisories are published I am fighting to have them committed to vuln.xml soon. Those SAs were published in the last week and are still missing in vuln.xml. https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=241931 The patch is really simple. Please commit it soon. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Omg upgrade wants to install MySQL
@lbutlr wrote on 11/02/2019 14:48: I have mariadb101 (client and server) installed, but when looking at what pkg update will do, one of the things it will do is install MySQL. Is there something I need to do to tell it to use MariaDB instead? Since I am here, it also wants to install a bunch of X11 packages, but I neither want nor have any desktop environment installed, so similar question, how can I tell pkg to run in “don’t enable GUIs” mode? Build your own packages wit poudriere and change default options to use MariaDB instead of MySQL for all ports, do not use X11 / GUI / DOC / NLS and whatever options you like / don't like. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: poudriere and ports overlay support
Andrea Venturoli wrote on 2019/10/25 11:54: On 2019-10-25 11:35, Miroslav Lachman wrote: I would like to try Overlay support in ports tree (announced few weeks ago in 2019Q4 branch). The question is will it work with Poudriere or not? We are building all our packages with poudriere. I tried adding OVERLAYS= /vol0/poudriere/ports/myports in to myports-make.conf but "poudriere options -z myports -p default -f /usr/local/etc/poudriere.d/pkglists/myports" failed when it reaches the / existing only in myports and not in default ports tree. Am I doing something wrong or Poudriere does not support Overlays? Sorry, I might have missed the original announcement. What is this Overlay support? I obviosly tried searching, but came up empty. Is that something which should achieve what ports-mgmt/portshaker (which I'm using) does? Yes, it should be something like portshaker but implemented directly in the ports tree Mk files. I didn't dig deep in it yet. Original announcement https://lists.freebsd.org/pipermail/freebsd-ports/2019-October/117066.html and in ports/CHANGES 20190903: AUTHOR: b...@freebsd.org The ports tree now supports overlays overlays are a way to help users to integrate their own ports tree with the official ports tree without having to maintain clone of the official tree and remerge on regular basis. The ports tree will lookup in the overlays (in the order the are listed in OVERLAY variable) for the dependencies and the USES. It will use the first found. in order to use it the user have to declare his overlays that way in their make.conf: OVERLAYS= overlay1 overlay2 overlay3 As I understand it should be possible to have default ports tree synchronised by SVN or Git with official tree and the second tree with private ports only, called for example "myports" (/usr/myports) Then in make.conf define OVERLAYS= /usr/myports and the ports framework should lookup first at the /usr/myports then /usr/ports. If /usr/myports contains some changes in Mk/Uses files, they should be applied too. But it didn't work with poudriere. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
poudriere and ports overlay support
I would like to try Overlay support in ports tree (announced few weeks ago in 2019Q4 branch). The question is will it work with Poudriere or not? We are building all our packages with poudriere. I tried adding OVERLAYS= /vol0/poudriere/ports/myports in to myports-make.conf but "poudriere options -z myports -p default -f /usr/local/etc/poudriere.d/pkglists/myports" failed when it reaches the / existing only in myports and not in default ports tree. Am I doing something wrong or Poudriere does not support Overlays? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Bug report commit request security/vuxml
Yasuhiro KIMURA wrote on 2019/10/24 02:54: Dear Committers, Would someone please commit following bug report? security/vuxml: Document potential bypass of Runas user restrictions in sudo https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=241244 If somebody will touch vuxml, please commit this too: security/vuxml: Add August FreeBSD Security Advisories https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=240322 I submitted it 2 months ago. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Is IPV6 option still necessary?
LuKreme wrote on 2019/10/11 00:23: On Oct 10, 2019, at 10:01, Lars Liedtke wrote: Why not just make building in IPv6 support the default, and introduce a flag if someone really needs or wants to build without that support? Because it adds to the load of testing. If you really need it, build from source. Building official packages with IPv6 is OK. Removing existing options from ports and saying "build from source" is ... stupid. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
committer needed for PR 240322 vuxml entries
Do we really need to wait more than month to get known SAs added to vuln.xml... https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=240322 Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: PHP version retirement
Martin Waschbüsch wrote on 2019/08/16 09:27: Thank you for your input. While I agree that PHP, in general, has been and still is a source of lots of security issues, I do not think this is the central point in this debate. There might be a high probability of security issues that are PHP related for all I know, but again, the real question is: Why drop a package that has just had recent security updates after a couple of weeks? I pointed out that I do not think lack of upstream development is in and of itself sufficient grounds for doing so. At the very least, while it may be unwise to use a now obsolete version of PHP, I doubt if an argument along the lines of 'We removed this from ports. It's for your own good' is a very good one. (For a number of reasons). +1 The only other arguments I got so far seem to be about resources. I can understand that. With limited resources you have to prioritize and something will have to give. Now, in a reply to Adam, I asked specifically if there were pointers that would help me evaluate how much effort is really involved. (My working theory being that I so far underestimate the work required to do this.) The effort to keep 5.6 in a tree for a few more months would be ... very little. It was done in quaterly branch after 5.6 was removed from head branch. I did my own updated version of the port (and extensions) from 5.6.39 to 5.6.40 without any issues - running on couple of machines till this day. Also, I asked if people were open to letting a group of people interested in doing so continue to maintain an old version of php so that it does not have to be removed from ports. Kurt suggested that as a feasible way forward and I agree. Earlier, Adam seemed open to discussing a way forward as well, but I am not sure that still is the case. Since I do not yet feel comfortable that I correctly estimate the amount of work, if enough people can be found to volunteer for this, but I remain hopeful. All this notwithstanding, would you be willing to exchange hints & ideas about securing (as far as possible) PHP setups some more, off-list? I'd like to ask some more about your approach. You can put webserver, or just php-fpm inside jail and then just nullfs mount the directory tree with websites on partition with noexec mount flag .. to name a few. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: firefox crash on start
moridin wrote on 2019/08/16 05:30: FF crashing immediately on start sounds like dbus being disabled (or rather, not being enabled). Yes, it says "Bus error", but dbus is running and all other browsers start without any problem. > ps axuww | grep dbus messagebus 895 0.0 0.0 256282440 - Is Wed07PM0:01.98 /usr/local/bin/dbus-daemon --system root1496 0.0 0.0 371642116 - IWed07PM0:00.00 dbus-launch --autolaunch ae4280dca0bb31b02b4dbd2057d0dd6d --binary-syntax --close-stderr root1497 0.0 0.0 256281696 - Is Wed07PM0:00.00 /usr/local/bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session quip1544 0.0 0.0 371642168 - IWed07PM0:00.00 dbus-launch --exit-with-session /usr/local/bin/startkde quip1545 0.0 0.0 256282904 - Ss Wed07PM0:17.99 /usr/local/bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session quip1986 0.0 0.0 256281960 - SWed07PM0:00.74 /usr/local/bin/dbus-daemon --config-file=/usr/local/share/defaults/at-spi2/accessibility.conf --nofork --print-address 3 Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: firefox
@lbutlr wrote on 2019/08/12 07:08: On 11 Aug 2019, at 20:29, bruce wrote: I have tried firefox. It crashes regularly That doesn’t sound right. If Firefox is crashing a lot there is something not quite right with your system or install. I am using Seamonkey on everyday basis. Sometimes my computer is running 20+ days without reboot and Seamonkey running all the time with 5 windows and total count of 200+ openned tabs. No crashes at all. Firefox crashes instantly after start (even if I delete its profile) Palemoon, Qupzilla, Iridium crashes few times a day. LibreOffice, Gimp and other apps are running fine. Only browsers are so unstable these days. And as you can read I am not alone with this kind of experience so I don't think it's just my computer problem. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: PHP version retirement
Franco Fichtner wrote on 2019/08/12 08:20: That „while“ is debatable, but it’s neither indefinitely nor immediately. The people responsible for FreeBSD ports and packages would be wise to enrich their policies with a more graceful way of dealing with legacy software, especially if it relates to more than a handful of ports in a single deprecation decision. TL;DR: don’t remove PHP ports prematurely and you’ll have less work reading mails like these. Part of the contract in providing packages includes providing support for them. Other OSes provide packages for software that they can never support, and there are definitely consequences for that paradigm. This is doubly true for PHP, which is extremely common and for which security fixes can be vitally important. Well, you are arguing against a grace period for obsolete software which is quite pointless because the software is not bad per se. It will be eventually and it should be removed and nobody argued against that. In the case of PHP 5.6 a clear error of judgement was made based on a reasonable decision at the time. It should give enough incentive to not let this happen again so quickly and try to learn from how it negatively impacts users. +1 from me. Removing PHP 5.6 before the last version was released by upstream was very inconvenient and we end up doing 5.6.40 version ourself. Then deploy on servers which cannot be updated to 7.x at that time. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Cleaning up pkg-message
Grzegorz Junka wrote on 2019/06/09 16:12: On 08/06/2019 19:11, Adam Weinberger wrote: Hello everyone, I want to get some stakeholder input on our pkg-message files. I think we need to have a clear policy about what does and doesn't belong in them, and I'd like to get your input. pkg-message is shown to every user on every install. UPDATING is only shown when users run `pkg updating` *and* /usr/ports/UPDATING exists. I suspect that only a small proportion of users do that. pkg-message needs to contain only highly relevant information. Many, many ports have messages with irrelevant information that users are likely to get message fatigue and ignore them entirely. I don't want to pick on Joe Barbish, because his work is absolutely fantastic, but dns/dns2blackhole/pkg-message is an example of a giant message that tells users to do the same thing they always do for any port: dns2blackhole Malware Prevention through Domain Blocking (Black Hole) Issue "man dns2blackhole" For configuration and usage information We now have the ability to specify messages that appear on initial install, or on upgrades from/to specific version. So here is what I propose as policy: pkg-message must contain only information that is vital to setup and operation, and that is unique to the port in question. Setup information should only be shown on initial install, and upgrade instructions should be shown only when upgrading to the relevant version. All committers have blanket approval to constrain existing messages to install/upgrade ranges using the UCL format specifications. Message pruning falls under the blanket approval as well, but committers are encouraged to get maintainer input beforehand. <<< What are your thoughts? # Adam I don't like the approach of separating install from update messages. It only works in the ideal scenario, which is almost never. Two reasons: 1. Very rarely I have time to configure all package requirements when installing a bunch of packages. I usually configure a few most important ones and leave the rest for later. Then I need to remember to re-read whatever requirements they might have had. 2. Very rarely just adding packages to the system works. From adding flavours, to removing KDE4, to renaming packages, etc. There is always something going on and almost every time I try to upgrade all packages in the system because of various problems I end up reinstalling all of them anyway (pkg upgrade -f). In either case update messages don't matter. In my opinion there should be just one short message shown when either upgrading or installing. If there are any specific instructions applicable when only installing or upgrading then it's safer to show in both cases with info in what condition they are applicable. When installing packages with many dependencies a typical user isn't even aware which packages have been added / installed and which have been updated. Why make the life more complicated than it needs to be? I disagree. The more the general messages the more noise to users. If something is useful only on the first install why should user read it on each pkg upgrade for many years in a lifetime of a machine? Then some useful info on upgrade will be missed between many useless messages. I remember change in PHP extensions which caused printing of useless notice on every pkg upgrade of every PHP extension. Average webserver has 10 - 20 of them (or more). This was so annoying that I patched our ports/Mk to not print those messages. If new UCL pkg-message format allows us to print only useful information in specific event I am glad it is finally here! The current state of pkg-message is very bad. Info in it is something I totally ignore on each upgrade because it contains useless informations which are printed to me on all machines on each pkg upgrade once or twice a month... Why if the info is useful only for the first install. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: PR 236751 security/base-audit committer needed
Kurt Jaeger wrote on 2019/06/03 22:00: Hi! There is a new version of security/base-audit waiting for commit almost 2 months. Could somebody commit it, please? PR 236751 Done. Thank you for all your work on ports! Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
PR 236751 security/base-audit committer needed
There is a new version of security/base-audit waiting for commit almost 2 months. Could somebody commit it, please? PR 236751 Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
net-mgmt/zabbix4-server ipmi poller eats all CPU
I upgraded Zabbix server from 4.0.4 to 4.0.7 and right after "service zabbix_server restart" zabbix starts to eat all CPU last pid: 4696; load averages: 2.11, 2.39, 2.24up 0+00:35:43 14:00:17 84 processes: 2 running, 82 sleeping CPU: 32.7% user, 0.0% nice, 67.3% system, 0.0% interrupt, 0.0% idle Mem: 312M Active, 126M Inact, 281M Wired, 200M Buf, 1243M Free Swap: 1024M Total, 1024M Free PID USERNAMETHR PRI NICE SIZERES STATETIMEWCPU COMMAND 899 zabbix1 810 63268K 14008K RUN 10:40 98.85% zabbix_server: ipmi poller #2 [polled 5 values, idle 5.055918 sec during 5.080817 sec] (zabbix_server) 769 mysql57 200 540M 202M select 0:09 0.25% [mysqld] I tried to restart Zabbix and reboot of the machine (VM in VirtualBox) but no change. Are there somebody running Zabbix server 4.0.7 with IPMI pollers without this CPU issues? The CPU was almost idle before upgrade. More details can be found in this PR 238224 https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=238224 Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Dovecot packages
Doug Hardie wrote on 2019/05/23 00:41: [...] Unless you notice this interesting anomaly, you can get easily burned like I did. I would recommend that either the dovecot2 package be deleted, or at least kept current with the dovecot package. There is no dovecot2 package in the official FreeBSD repository / ports tree: https://www.freshports.org/search.php?query=dovecot You have very stale packages on your machines and you didn't read instructions in UPDATING file of "pkg updating" command: 20170807: AFFECTS: users of mail/dovecot2 and mail/dovecot2-pigeonhole AUTHOR: ad...@freebsd.org Now that dovecot1 has been removed from the ports tree, dovecot2 and dovecot2-pigeonhole have been renamed to simply dovecot and dovecot-pigeonhole. pkg should handle the rename automatically, but if you run into trouble, you can point pkg at the new origin via: # pkg set -o mail/dovecot2:mail/dovecot # pkg set -o mail/dovecot2-pigeonhole:mail/dovecot-pigeonhole Similarly, you can point portmaster at the new location via: # portmaster -o mail/dovecot mail/dovecot2 If all else fails, just remove and reinstall the package: # pkg delete dovecot2 # pkg install dovecot Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Reinstalling with dependencies
Grzegorz Junka wrote on 2019/05/22 14:11: [...] Are you saying that even if elinks was reinstalled with dependencies that wouldn't help? We have two issues here: 1. How to reinstall a package with dependencies (as stated in the subject) You can try something like this pkg install -f `pkg info -d elinks | tr -d :` pkg info -d will list direct dependencies of the port and then this list is given to command pkg install -f. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: replacement for lang/phantomjs [deprecated]
Steven Hartland wrote on 2019/05/22 13:56: Pretty sure chrome has a headless mode now, perhaps you could use that. The last time I tried chrome headless it was so big beast (for headless servers) that I abandoned it. Build time in my poudriere was about 33 hours and so many dependencies I don't want to install in headless machines. So I hope there is still some chance to find something lighter than headless chrome. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: graphics/php*-gd requires X11 by default
Greg Kennedy wrote on 2019/05/22 05:07: Never mind! I just saw that this was fixed in October for php73-gd. Need to update my PHP version. I think all supported versions of PHP should have the same options enabled. And I support X11=off by default. On Thu, May 16, 2019 at 8:54 PM Greg Kennedy wrote: I'm raising this again as the last time I tried to check with the port maintainer, he had not made the change, and wasn't available to do so. Can someone change the default config for graphics/php72-gd from X11=on to X11=off? It will save a large number of dependencies on X11, at the cost of not supporting this (very rarely used) image format. On Sat, Sep 15, 2018 at 12:14 PM Greg Kennedy wrote: I sent an email to the maintainer of these ports (t...@freebsd.org) a few days ago, but didn't hear back - so I thought I'd raise the same question to the mailing list instead. --- Hey there! I was wondering something about the php-gd ports (specifically 7.2 but earlier as well). Looking at the config options I see this: X11=on: Enable XPM support With this option enabled php-gd can read XPixMap images, but in turn this pulls in a host of additional X11 dependencies. That's an awful lot for (usually) headless servers, especially for an image format that is (relatively) unused. I don't mind the option being there but because it is defaulted to "on", that means official FreeBSD packages are built with this support and all the dependencies. On my system I have to manually build this one port just to turn off the option. Would it be possible to modify the defaults for this port so that X11=off by default? Or, why is the default to "on"? (I have a guess, that this respects WITHOUT_X11 in /etc/make.conf, but that's useless for pre-built binary packages) ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
replacement for lang/phantomjs [deprecated]
The port lang/phantomjs is marked as deprecated / expired 2019-05-15. Is there any alternative in the ports tree? I searched the net and found https://slimerjs.org but it is not in the ports tree. (I don't have a skill to port it and maintain it myself) One of my clients is using phantomjs thru some webapplication so I need to find some replacement. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: CFT: FreeBSD Package Base
Cy Schubert wrote on 2019/05/01 05:56: In message <292eadc6-3662-ec43-1175-53fc25248...@quip.cz>, Miroslav Lachman wri tes: David Chisnall wrote on 2019/04/30 10:22: On 29/04/2019 21:12, Joe Maloney wrote: With CFT version you chose to build, and package individual components such as sendmail with a port option. That does entirely solve the problem of being able to reinstall sendmail after the fact without a rebuild of the userland (base) port but perhaps base flavors could solve that problem assuming flavors could extend beyond python. This sounds very much like local optimisation. It's now easy to create a custom base image. Great. But how do I express dependencies in ports on a specific base configuration? This is easy if I depend on a specific base package, but how does this work in your model? For example, if I have a package that depends on a library that is an optional part of the base system, how do I express that pkg needs to either refuse to install it, or install a userland pkg that includes that library in place of my existing version as part of the install process? More importantly for the container use case, if I want to take a completely empty jail and do pkg ins nginx (for example), what does the maintainer of the nginx port need to do to express the minimum set of the base system that needs to be installed to allow nginx to work? One of the goals for the pkg base concept was to allow this kind of use case, easily creating a minimal environment required to run a single service. With a monolithic base package set, you're going to need some mechanism other than packages to express the specific base subset package that you need and I think that you need to justify why this mechanism is better than using small individual packages. Will it not be maintainer's nightmare to take care of all the dependencies on the base packages for each port we have in the ports tree? No more than it is today. Remember, people have been doing this sort of thing for decades. If the folks at Red Hat, Oracle (formerly Sun), and IBM can do it, I'm sure we can too. The dependency lists will be longer. We may require dependency lists that allow the choice of one of many prereqs or coreqs. They are experts and they are paid for their work. I am not. I am maintaining a few packages and the reality is I don't know what they need in base. Till these days I don't care about this kind of dependency. I am not system developer or programmer and I think there are more than just me who see this as a kind of problem. So in this case, pkg base gives me nothing but more work on those packages. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: CFT: FreeBSD Package Base
David Chisnall wrote on 2019/04/30 10:22: On 29/04/2019 21:12, Joe Maloney wrote: With CFT version you chose to build, and package individual components such as sendmail with a port option. That does entirely solve the problem of being able to reinstall sendmail after the fact without a rebuild of the userland (base) port but perhaps base flavors could solve that problem assuming flavors could extend beyond python. This sounds very much like local optimisation. It's now easy to create a custom base image. Great. But how do I express dependencies in ports on a specific base configuration? This is easy if I depend on a specific base package, but how does this work in your model? For example, if I have a package that depends on a library that is an optional part of the base system, how do I express that pkg needs to either refuse to install it, or install a userland pkg that includes that library in place of my existing version as part of the install process? More importantly for the container use case, if I want to take a completely empty jail and do pkg ins nginx (for example), what does the maintainer of the nginx port need to do to express the minimum set of the base system that needs to be installed to allow nginx to work? One of the goals for the pkg base concept was to allow this kind of use case, easily creating a minimal environment required to run a single service. With a monolithic base package set, you're going to need some mechanism other than packages to express the specific base subset package that you need and I think that you need to justify why this mechanism is better than using small individual packages. Will it not be maintainer's nightmare to take care of all the dependencies on the base packages for each port we have in the ports tree? Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: How to use @preexec to test for installed packages
Matthew Seaman wrote on 2019/04/06 19:00: On 06/04/2019 14:58, Matthias Fechner wrote: as pkg cannot handle CONFLICTS_INSTALL I tried now to implement this as a preinstall command using @preexec in pkg-plist. The command should check if a package is installed and stop the installation or continue if the package is not installed. I tried it with the following command: @preexec `/usr/sbin/pkg -N info -e gogs`; if [ $? -eq 0 ]; then echo "Gitlab cannot be installed together with gogs as both of them modify .ssh/authorized_keys" && exit 1; else echo "Gogs not installed, continue."; fi But it does not work. Now matter if gogs is installed or not. If I execute the same line in a shell script, it works fine. Anyone an idea? pkg(8) does handle the most common reason for packages conflicting at install time -- file name clashes. Indeed, it does this automatically with no need of input from porters or maintainers, although it is usual to add CONFLICTS_INSTALL lines to port Makefiles to document clashes discovered this way. However, where there are other reasons for packages to conflict at install time, then you are correct that pkg doesn't handle this. There's simply no mechanism to include information about package conflicts into pkg metadata. Then you can add some fake (empty) file to the plist of package A which will conflict with package B and then pkg conflict will work as expected. But I am not sure we should prevent installation of some package just because it uses authorized_keys too. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
commiter needed - mail/ratelimit-policyd update to version 1.0.3
Hi, can somebody commit PR 236501? https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=236501 It is my own port and this is just a simple update to a newer version. I see Steve Wills assigned to this PR but nothing happened. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
hp-ams for HP servers
I have HP Microserver G8, many options in iLO 4 says "iLO did not detect the Agentless Management Service when this page was loaded" I know HP has AMS for Windows and Linux but is there any port of this to FreeBSD? I just found this https://github.com/marker55/hp-ams but it does not contains anything about FreeBSD. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Wire port
Mahdi Mokhtari wrote on 2019/03/08 12:03: On Fri, Mar 8, 2019 at 11:28 AM Miroslav Lachman <000.f...@quip.cz> wrote: Hello, Hi Miroslav, I am interested in using Wire on FreeBSD and I found the information about Wire on https://wiki.freebsd.org/WantedPorts where it states WIP and you are listed as a person working on porting. Can I ask you how it goes with porting Wire to FreeBSD and if it will be possible to run it on FreeBSD? The WIP on my side goes back in 2017 works on Electron with pizzamig@. So far the wire requires that as a dependency, and that's why I personally use the webapp now `:) After that once I tried to write a similar-to-linux section in the package.json[1] of the wire with the assumption "IF we had electron ready". But still that's not even useful now I guess. If you want to start it I encourage you to take a look at this repo[2] and this open issue on upstream[3]. [1] https://github.com/wireapp/wire-desktop/blob/master/package.json [2] https://github.com/yzgyyang/freebsd-ports-electron [3] https://github.com/electron/electron/issues/3797 Thank you for your quick reply. I'll take a look on it soon. I am not a developer so I think I will not be able to push it forward by my-self. But I will try it. :) Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Wire port
Hello, I am interested in using Wire on FreeBSD and I found the information about Wire on https://wiki.freebsd.org/WantedPorts where it states WIP and you are listed as a person working on porting. Can I ask you how it goes with porting Wire to FreeBSD and if it will be possible to run it on FreeBSD? Kind Regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: how to patch in port with NO_BUILD?
Walter Schwarzenfeld wrote on 2019/02/24 16:02: Make a patch that includes the whole changed file and change the do-install target. In the meantime I came up with this https://bz-attachments.freebsd.org/attachment.cgi?id=202298 https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=235968 Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: FreeBSD Port: graphics/xpdf4 without X11 GUI
Cy Schubert wrote on 2019/02/22 15:45: On February 22, 2019 1:42:05 AM PST, Miroslav Lachman <000.f...@quip.cz> wrote: Hi, I used xpdf for many years on headless servers without X11, now I noticed that this options is nolonger there. Was it removed by upstream or is it still some way to have working xpdf without dependency on Qt5? It was removed upstream. Xpdf3 still has the no X11 option. Ideally people should use poppler. Poppler was designed for headless use and it is also a prerequisite for many other ports which require some of the headless utilities. Poppler contains some command line utilities not found in other pdf ports such as the xpdf's. I was curious what changed between 4.00 and 4.01, because 4.00 was able to build without X11 xpdf4-4.00_1,1 Depends on : gsfonts-8.11_8 freetype2-2.9.1 png-1.6.35 Anyway, thank you for the hint to poppler. We'll test it and if it'll work, we replace xpdf with it! Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
FreeBSD Port: graphics/xpdf4 without X11 GUI
Hi, I used xpdf for many years on headless servers without X11, now I noticed that this options is nolonger there. Was it removed by upstream or is it still some way to have working xpdf without dependency on Qt5? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
cannot build Zabbix with MariaDB 10.3
I am trying to build Zabbix 4 and Zabbix 3.4 in poudriere on FreeBSD 11.2 with MariaDB 10.3. Both versions (4.0 and 3.4) failed with the same error: checking for architecture... freebsd (freebsd11.2) checking size of void *... 8 checking for mysql_config... /usr/local/bin/mysql_config checking for main in -lmariadb... no configure: error: Not found mariadb library ===> Script "configure" failed unexpectedly. This is make.conf for Poudriere OPTIONS_UNSET= X11 GUI CUPS DOCS EXAMPLES NLS HAL DEFAULT_VERSIONS=apache=2.4 perl5=5.28 mysql=10.3m php=7.1 python=2.7 python3=3.6 pgsql=9.6 ssl=openssl bdb=5 And Zabbix options ---Begin OPTIONS List--- ===> The following configuration options are available for zabbix34-server-3.4.15: CURL=on: Support for web monitoring FPING=on: Build/install fping for ping checks IPMI=on: Support for IPMI checks IPV6=on: IPv6 protocol support JABBER=off: Support for Jabber media type LDAP=off: Support for LDAP server checks LIBXML2=off: Support for libxml2 (required by monitoring VMware) NMAP=on: Build/install nmap for o/s detection SSH=on: Support for SSH-based checks > Options available for the single DB: you have to select exactly one of them MYSQL=on: MySQL database support PGSQL=off: PostgreSQL database support SQLITE=off: SQLite database support ORACLE=off: Oracle database support > Support for database checks via ODBC: you have to select exactly one of them IODBC=off: ODBC backend via iODBC UNIXODBC=on: ODBC backend via unixODBC > SSL protocol support: you have to select exactly one of them OPENSSL=on: SSL/TLS support via OpenSSL GNUTLS=off: SSL/TLS support via GnuTLS ===> Use 'make config' to modify these settings ---End OPTIONS List--- I found similar reports on the net but they were for older versions, marked as fixed more then year ago. I am using / building Zabbix with the same setting with MariaDB 10.1 without problem. Can anybody help to debug and fix this issue? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Why does pkg want to install emby-server?
Dave Cottlehuber wrote on 2019/01/19 01:08: On Fri, 18 Jan 2019, at 22:27, Stefan Bethke wrote: No matter what package I try to install on my 12-stable machine, pkg always adds emby-server. When I check after the install, nothing requires emby-server. What’s going on? you probably have a further package that is missing the emby-server as a dependency. sudo pkg check -d -v should tell you what package that is, either delete it or accept the corrective action I had this problem in the past too and it was not related to any required dependencies. After pkg install finished the installation of wanted package and unwanted unrelated packages next call to "pkg autoremove" deinstalled those unwanted packages (it means nothing needs them). But next pkg install put them back. This problem disappeared after a few month after complete reinstall of all packages by pkg upgrade -f (after system upgrade). Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: php56
Randy Bush wrote on 2019/01/13 21:19: I have a mission critical app server running an old PHP 5.6 application which will not work on PHP 7+. Wordpress 5 is working fine with PHP 7.1. If you have some problem with Wordpress it is probably caused by some unmaintained incompatible plugin. Try to disable plugins and then WP will work and you can try to install some newer versions of the plugins or replace them with someting else. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: vim - GTK2 or GTK3?
Adriaan de Groot wrote on 2019/01/03 14:28: Niclas wrote: On Thursday, 3 January 2019 13:00:02 CET freebsd-ports-requ...@freebsd.org wrote: Firefox and Chromium both depend on GTK3, so it's highly likely that a typical desktop user has GTK3 installed. +1, GTK3 is probably the best choice. As a side note, it looks like libreoffice defaults to GTK2 as well, perhaps it should be switched to GTK3 also? As a not-really-GTK-using person, I still have both GTK2 and GTK3 installed on my system running KDE Plasma. Not for vim though: Installed packages to be REMOVED: gtk2-2.24.32 fontforge-20170731 mftrace-1.2.18_1 Installed packages to be REMOVED: gtk3-3.22.30_4 gpsd-3.17 I'm a fan of pushing for toolkit migration, so reducing the number of things that pull in GTK2 is a good thing. So if we're expressing hopes that ports might be made GTK2-free (by porting to GTK3 for instance) then I'd hope that fontforge gets that treatment, too. From looking at the source repo, I don't think the GTK2 option actually works (and the comments suggest it's not all that good anyway). In the configure.ac it looks like there are spelling-inconsistencies between fontforge_can_use_gtk=yes and, e.g,, FONTFORGE_ARG_ENABLE_GDK (mtrace depends on fontforge, so fixing fontforge would clean GTK2 off my system) It is not so simple. What works for you doesn't work for somebody else. I am running KDE4 with QT4 but some applications use GTK2 as the only option or I set them to use GTK2 because GTK2 has better theme integration to KDE4 / QT style than GTK3 (in my personal case). And, for example, Total Commander has two choices - QT or GTK2. I tried it to build with QT but it failed so I switched to GTK2 and it builds fine. So in my current situation I am glad we have GTK2 and GTK3 and all my applications are working (better with GTK2 than GTK3). And having both GTK versions installed is not a big problem. Their size is relatively small compared to apps like browsers, libreoffice etc. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Maia Mailgaurd / PHP 7 support
Janketh Jay wrote on 2018/12/30 01:03: Hi All, On 03/12/2018 2:17 pm, Reko Turja via freebsd-ports wrote: Hello Janky Jay! -Original Message- Sent: Monday, December 3, 2018 4:17 PM Subject: Re: Maia Mailgaurd Okay. No sweat. I'll work on getting a port patch going as that seems it would be the easiest way to push this out. I'll provide a download as soon as I have something available for testing. Sounds good, thank you for your work on ports! -Reko So, after a bit of testing (and not being able to update the port for PHP7 due to Smarty issues), below is a simple patch that will keep the current port version of Maia (1.0.4_3) working with PHP7 (testing with PHP 7.2). Please let me know if this does not work or if there is anything else that needs to be added or changed. Hopefully, I'll get the Smarty issue(s) sorted out soon and have a new port submitted. [...] Also, if you're interested in seeing/testing the new port (from the latest Github commit for technion's maia_mailguard), you can download the patch from the link below [1] and let me know if it's successful for you. Everything seems to work except for the WebUI due to a fatal call to Smarty which I cannot figure out (I'm not a PHP developer). However, once that is sorted, I should be ready to submit the new port so Maia can be updated to the latest version. I forgot about this in my previous message - I guess there is an old version of Smarty not supporting PHP 7. It would be probably better to replace it with newer version of Smarty with PHP 7 support but there can be some syntax changes in Smarty config / initialisation. I remember something about it about a year ago when I was upgrading one of our own (very old) project to newer Smarty templates. There are some differences between Smarty 2 and Smarty 3. But as I read now on their website, the latest Smarty 2 version 2.6.31 supports PHP 7.2 so it should be painless to replace old noncompatible version with this one. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Maia Mailgaurd
Janketh Jay wrote on 2018/12/30 01:03:
[...]
diff -Naur maia.orig/cache.php maia/cache.php
--- maia.orig/cache.php 2015-02-15 15:19:45.0 -0700
+++ maia/cache.php 2018-10-14 20:25:30.27896 -0600
@@ -554,7 +554,7 @@
$rectmp = "";
foreach ($to_list as $recipient) {
if (isset($personal_addresses[$recipient]) ||
$domain_default) {
- $rectmp[] = $recipient;
+ $rectmp = $recipient;
}
}
$rows[$count]['recipient_email'] = $rectmp;
Essentially, you just need to remove the "[]" from "rectmp" on
line 558 in your /usr/local/www/maia/cache.php file.
I don't use Maia Mailguard and I didn't read the source code but I think
your patch is wrong. It changed the function. Original code assigned all
recipient addresses (appending) in to an array (hash) $rectmp in a
foreach loop and then assign this array to $rows[$count]['recipient_email'].
But now you are using it as variable so if there are more than one
recipient this variable is overwritten on each iteration and then just
the last recipient is assigned to $rows[$count]['recipient_email'].
My very wild guess is that it should be like this
-$rectmp = "";
+$rectmp = array();
foreach ($to_list as $recipient) {
if (isset($personal_addresses[$recipient]) ||
$domain_default) {
$rectmp[] = $recipient;
}
}
$rows[$count]['recipient_email'] = $rectmp;
I guess you want to fix some PHP 7 warning / syntax error with $rectmp
created ass plain variable and later used as an array so I defined as an
array first.
But maybe I am totally wrong :)
I just made similar fix few days ago in an old version of PostfixAdmin
after upgrade from PHP 5.6 to 7.1.
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: vulnerabilities bogus error
Ernie Luzar wrote on 2018/12/15 15:22: Walter Schwarzenfeld wrote: https://vuxml.freebsd.org/freebsd/vuln.xml.bz2 What path is this file suppose to be uncompressed into? /var/db/pkg/vuln.xml But isn't "pkg audit -F" enough? Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: How much memory to compile www/chromium?
Kurt Jaeger wrote on 2018/12/12 19:46: Hi! The system in question is an RPI3 running r341643. That will hurt. A regular chromium build takes about 16 CPU-hours on my fairly modern i7 CPU, SSDs and large RAM. Parallel compilation helps a lot, keeping 16 CPU threads busy and my feet warm (building chromium alone takes 1 hour of wall clock time on my system - if that's the only thing I'm running at that time). How can this parallel compilation be configured for chromium in poudriere ? I have a builder host that would be able to handle this 8-} Something like this can help ALLOW_MAKE_JOBS_PACKAGES="pkg py* cmake openjdk* openjdk8 phantomjs gcc6 gcc7 gcc* webkit2-* webkit-* llvm39 llvm60 llvm* rust iridium-browser chromium" But also can hurt if you run in to parallel build of multiple parallel packages (for example iridium with chromium at the same time) Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: make makepatch does not work (only 2 of 5 expected patches created)
Miroslav Lachman wrote on 2018/12/04 12:44: I am trying to create patch for pandorafms_console to make it work with PHP 7.1. There are 5 file that need to be patched. I created *.orig for all of them, make changes in them and then "make makepatch" created patches only for 2 out of 5 files. Patches are crated for: install.php include/auth/mysql.php Not created for: include/functions.php include/functions_agents.php include/functions_db.php I tried it many times but makepatch always ignores functions files in include dir. Has it something to do with filenames? Sorry for the noise. It was mistake on my side. functions files did not contain changes. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
make makepatch does not work (only 2 of 5 expected patches created)
I am trying to create patch for pandorafms_console to make it work with PHP 7.1. There are 5 file that need to be patched. I created *.orig for all of them, make changes in them and then "make makepatch" created patches only for 2 out of 5 files. Patches are crated for: install.php include/auth/mysql.php Not created for: include/functions.php include/functions_agents.php include/functions_db.php I tried it many times but makepatch always ignores functions files in include dir. Has it something to do with filenames? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: packages and base jails
Eugene Grosbein wrote on 2018/11/27 00:42:
27.11.2018 3:24, Michael W. Lucas wrote:
Hi,
I'm writing a book on jails and am looking for BCP. I'd like to
present either "This is the approved solution and should work" or
"these are the gotchas with any of these, choose your pain."
Folks want base jails to include packages, but also want to install
additional packages--which won't happen if /usr/local is mounted
read-only in the base jail. Trawling around the Net I see a couple
options. Both involve the primary jail using a different package
repo. The overlay jail uses the standard package repo.
1) primary jail uses a repo with PREFIX=/usr/pkg or /opt. Works in my
simple use cases once I set ldconfig directories in rc.conf, but I'm
told programs like pkgconfig can go sideways.
2) base jail repo uses with PREFIX=/. Utterly violates separation of
base and pkg, but everything should find everything out of the
box. Again, seems to work in my wimpy use cases.
Is there an option that should work? Or is a matter of choosing
between horrors?
Not sure I understand the problem which I don't have using sysutils/ezjail
that uses base jail situated in /usr/local/j/basejail in my case.
For each distinct jail instance, it null-mounts it read-only
to /usr/local/j/${JAILNAME}/basejail and /usr/local/j/${JAILNAME} it jail's
root.
Inside this root, /bin is symlink to /basejail/bin, and /boot, /libexec, /rescue
and /sbin are similar symlinks, so are
/usr/{bin|include|lib|lib32|libdata|libexec|ports|sbin|share}
all symlinks to corresponding directories inside ro-mounted /basejail/usr/...
But not /usr/local nor /usr/{src|obj}, if that matters. So each jail have its
own
set of packages or even ports if I choose to null-mount host's /usr/ports
readonly
to /usr/local/j/${JAILNAME}/basejail/usr/ports and write to jail's
/etc/make.conf:
I guess Michael wants to have some packages installed in shared basejail
(packages common to all jails) and some packages later installed
separately in jails. And this is something that I would never do. :)
But you can try some union fs overlay on top of shared /usr/local. But
again - I will not do this in production environment.
Miroslav Lachman
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: packages and base jails
Ernie Luzar wrote on 2018/11/26 22:12: Michael W. Lucas wrote: Hi, I'm writing a book on jails and am looking for BCP. I'd like to present either "This is the approved solution and should work" or "these are the gotchas with any of these, choose your pain." Folks want base jails to include packages, but also want to install additional packages--which won't happen if /usr/local is mounted read-only in the base jail. Trawling around the Net I see a couple options. Both involve the primary jail using a different package repo. The overlay jail uses the standard package repo. 1) primary jail uses a repo with PREFIX=/usr/pkg or /opt. Works in my simple use cases once I set ldconfig directories in rc.conf, but I'm told programs like pkgconfig can go sideways. 2) base jail repo uses with PREFIX=/. Utterly violates separation of base and pkg, but everything should find everything out of the box. Again, seems to work in my wimpy use cases. Is there an option that should work? Or is a matter of choosing between horrors? Thanks, ==ml I use a common base jail mounted read only and the jail /usr/local & /etc mounted r/w. From the jail console bootstrap pkg and every thing works just like on the host. Now the ports tree is totally different, I create the ports tree normally on the host. And then if I need the ports tree in a jail I issue the mv command to move from host to jail and when its not needed any more I mv it back to the host. Only one ports tree for host and all jails. Haven't had the need to do that since new pkg works so good now. Saw this is how qjail does it so used that concept in my own manual jail system. You don't need to move ports tree in and out, you can use nullfs mount of a directory, probably read only in jail with some tweaks in make.conf in jail: WRKDIRPREFIX= /var/ports DISTDIR=/var/ports/distfiles PACKAGES= /var/ports/packages INDEXDIR= /var/ports Or you can share distfiles between host and jail. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: devel/llvm60 build failure in poudriere
Niclas Zeising wrote on 2018/11/03 19:14: On 11/2/18 12:39 PM, Miroslav Lachman wrote: Dimitry Andric wrote on 2018/11/02 08:26: You should first check whether your ports tree is cleanly checked out, revert any local changes, and then try again. If that fails too, delete the ports tree, and check it out from scratch. Update of ports/Mk fixed the problem with llvm60 but it breaks 40+ other ports dependent on old xorgproto. It seems I cannot update just iridium and firefox. xorgproto has replaced all the old *proto ports. You need to update that part of the tree if you are updating bsd.xorg.mk. Remember, howerver, that having different parts of the ports tree at different versions is a recipe for disaster. Yes, I know it. That's why I gave up and will stay on old versions of iridium and firefox until I will have time to prepare and test (in virtualbox) the update of all 1500+ installed packages on my desktop (PC-BSD based) Thanks anyway Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: devel/llvm60 build failure in poudriere
Dimitry Andric wrote on 2018/11/02 08:26: On 2 Nov 2018, at 01:56, Miroslav Lachman <000.f...@quip.cz> wrote: I am trying to build iridium-browser which dependes on llvm60, but my poudriere cannot build llvm60 with the following error: === ===> Fetching all distfiles required by llvm60-6.0.1_3 for building => SHA256 Checksum OK for llvm-6.0.1.src.tar.xz. => SHA256 Checksum OK for cfe-6.0.1.src.tar.xz. === === === === ===> Fetching all distfiles required by llvm60-6.0.1_3 for building ===> Extracting for llvm60-6.0.1_3 => SHA256 Checksum OK for llvm-6.0.1.src.tar.xz. => SHA256 Checksum OK for cfe-6.0.1.src.tar.xz. /bin/mv /wrkdirs/usr/ports/devel/llvm60/work/cfe-6.0.1.src /wrkdirs/usr/ports/devel/llvm60/work/llvm-6.0.1.src/tools/clang === === === === ===> Patching for llvm60-6.0.1_3 ===> Applying extra patch /usr/ports/devel/llvm60/files/clang I can't seem to find a patch in there anywhere. *** Error code 2 Stop. make: stopped in /usr/ports/devel/llvm60 =>> Cleaning up wrkdir ===> Cleaning for llvm60-6.0.1_3 build of devel/llvm60 | llvm60-6.0.1_3 ended at Fri Nov 2 01:50:44 CET 2018 build time: 00:00:10 !!! build failure encountered !!! OS on the host is 11.2 amd64, building jail is 10.4 amd64 I tried selec or unselect all options in dialog for llvm60 but it didn't help. What does the error "I can't seem to find a patch in there anywhere" mean? It means that one of the patch files in your port checkout has been corrupted, and the patch utility is not able to interpret it anymore. You should first check whether your ports tree is cleanly checked out, revert any local changes, and then try again. If that fails too, delete the ports tree, and check it out from scratch. Update of ports/Mk fixed the problem with llvm60 but it breaks 40+ other ports dependent on old xorgproto. It seems I cannot update just iridium and firefox. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: devel/llvm60 build failure in poudriere
Dimitry Andric wrote on 2018/11/02 08:26: On 2 Nov 2018, at 01:56, Miroslav Lachman <000.f...@quip.cz> wrote: ===> Patching for llvm60-6.0.1_3 ===> Applying extra patch /usr/ports/devel/llvm60/files/clang I can't seem to find a patch in there anywhere. *** Error code 2 Stop. make: stopped in /usr/ports/devel/llvm60 =>> Cleaning up wrkdir ===> Cleaning for llvm60-6.0.1_3 build of devel/llvm60 | llvm60-6.0.1_3 ended at Fri Nov 2 01:50:44 CET 2018 build time: 00:00:10 !!! build failure encountered !!! OS on the host is 11.2 amd64, building jail is 10.4 amd64 I tried selec or unselect all options in dialog for llvm60 but it didn't help. What does the error "I can't seem to find a patch in there anywhere" mean? It means that one of the patch files in your port checkout has been corrupted, and the patch utility is not able to interpret it anymore. You should first check whether your ports tree is cleanly checked out, revert any local changes, and then try again. If that fails too, delete the ports tree, and check it out from scratch. I freshly checked out devel/llvm60 but it end by the same error: root@sm-reserve ~/# svnlite checkout https://svn.freebsd.org/ports/head/devel/llvm60 /vol0/poudriere/ports/pcbsd-head/devel/llvm60 A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/files A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lldb A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/openmp A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-head-r331065.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-head-r336969.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-head-r337615.diff A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/openmp/patch-i386 A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/Makefile A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-head-r331066.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-head-r336227.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-head-r338697.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-head-r339019.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-fformat_extensions.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-fopenmp.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-tools_clang_lib_Headers_CMakeLists.txt A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-tools_clang_tools_clang-format_clang-format-sublime.py A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-tools_clang_tools_clang-format_clang-format.py A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-tools_clang_tools_clang-format_git-clang-format A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/clang/patch-tools_clang_tools_scan-build_libexec_ccc-analyzer A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld/patch-head-r331731.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld/patch-head-r333401.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld/patch-head-r336664.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld/patch-head-r336972.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-cmake_modules_AddSphinxTarget.cmake A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld/patch-head-r337282.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld/patch-head-r338251.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld/patch-head-r338682.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld/patch-head-r339013.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lld/patch-head-r339304.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lldb/patch-head-r332849.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/lldb/patch-head-r332965.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-head-r308867.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-head-r330686.diff A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-svn-r322325 A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/distinfo A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-head-r331366.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-head-r336970.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/patch-head-r338689.diff A /vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/openmp/patch-bug32279 A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/files/llvm-wrapper.sh.in A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/pkg-descr A/vol0/poudriere/ports/pcbsd-head/devel/llvm60/pkg-plist Checked out revision 483798. I cannot checkout the whole ports tree because I need to stay on older versions,
devel/llvm60 build failure in poudriere
I am trying to build iridium-browser which dependes on llvm60, but my poudriere cannot build llvm60 with the following error: === ===> Fetching all distfiles required by llvm60-6.0.1_3 for building => SHA256 Checksum OK for llvm-6.0.1.src.tar.xz. => SHA256 Checksum OK for cfe-6.0.1.src.tar.xz. === === === === ===> Fetching all distfiles required by llvm60-6.0.1_3 for building ===> Extracting for llvm60-6.0.1_3 => SHA256 Checksum OK for llvm-6.0.1.src.tar.xz. => SHA256 Checksum OK for cfe-6.0.1.src.tar.xz. /bin/mv /wrkdirs/usr/ports/devel/llvm60/work/cfe-6.0.1.src /wrkdirs/usr/ports/devel/llvm60/work/llvm-6.0.1.src/tools/clang === === === === ===> Patching for llvm60-6.0.1_3 ===> Applying extra patch /usr/ports/devel/llvm60/files/clang I can't seem to find a patch in there anywhere. *** Error code 2 Stop. make: stopped in /usr/ports/devel/llvm60 =>> Cleaning up wrkdir ===> Cleaning for llvm60-6.0.1_3 build of devel/llvm60 | llvm60-6.0.1_3 ended at Fri Nov 2 01:50:44 CET 2018 build time: 00:00:10 !!! build failure encountered !!! OS on the host is 11.2 amd64, building jail is 10.4 amd64 I tried selec or unselect all options in dialog for llvm60 but it didn't help. What does the error "I can't seem to find a patch in there anywhere" mean? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: amavisd-new failed to start after upgrade to FreeBSD 11.2
Mark Martinec wrote on 2018/10/06 00:39: Miroslav Lachman wrote on 2018/10/04 11:12: I upgraded mailserver from 10.4 to 11.2 (it is amd64 with GENERIC kernel). All packages were reinstalled from repository built in our poudriere for 11.2. Everything works except amavisd-new. Oct 4 10:21:17 roxy amavis[89209]: (!!)TROUBLE in child_init_hook: BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such process. at (eval 93) line 338. Oct 4 10:21:17 roxy amavis[89209]: (!)_DIE: Suicide in child_init_hook: BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such process. at (eval 93) line 338. for the records - 8 hours after upgrad and the same error happens on another mailserver. Amavisd is spamming log with errors about "BDB can't connect db env" and does not pass any e-mail messages. I found the same reports on the internet but no solution how to fix this. (enable_db=0 is temporary workaround but it disables some functions) Are there somebody with working setup of Amavisd-new on FreeBSD 11.2 with enable_db=1? 2018-10-05 22:18, je Florian Smeets via freebsd-ports napisal I've had to set enable_db = 0 on all my instances. Florian Probably not the answer you are looking for, but: ever since Redis database backend was introduced five years ago to amavis (see 2.8.1 release notes), and soon afterwards also as a SpamAssassin bayes back-end, I never looked back at BerkeleyDB again. It was all trouble - the sooner you switch to redis, the better - for reliability and speed, and also for added amavis functionality (pen pals, IP address reputations). OK, thanks for this answer. I'll try to configure Redis for Amavis. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: amavisd-new failed to start after upgrade to FreeBSD 11.2
Miroslav Lachman wrote on 2018/10/04 11:12: I upgraded mailserver from 10.4 to 11.2 (it is amd64 with GENERIC kernel). All packages were reinstalled from repository built in our poudriere for 11.2. Everything works except amavisd-new. for the records - 8 hours after upgrad and the same error happens on another mailserver. Amavisd is spamming log with errors about "BDB can't connect db env" and does not pass any e-mail messages. I have maillog full of errors: Oct 4 10:21:17 roxy amavis[89209]: (!!)TROUBLE in child_init_hook: BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such process. at (eval 93) line 338. Oct 4 10:21:17 roxy amavis[89209]: (!)_DIE: Suicide in child_init_hook: BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such process. at (eval 93) line 338. Oct 4 10:21:17 roxy amavis[89210]: (!!)TROUBLE in child_init_hook: BDB can''t connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such file or directory. at (eval 93) line 338. [...] I found the same reports on the internet but no solution how to fix this. (enable_db=0 is temporary workaround but it disables some functions) [...] Are there somebody with working setup of Amavisd-new on FreeBSD 11.2 with enable_db=1? ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
amavisd-new failed to start after upgrade to FreeBSD 11.2
I upgraded mailserver from 10.4 to 11.2 (it is amd64 with GENERIC kernel). All packages were reinstalled from repository built in our poudriere for 11.2. Everything works except amavisd-new. I have maillog full of errors: Oct 4 10:21:17 roxy amavis[89209]: (!!)TROUBLE in child_init_hook: BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such process. at (eval 93) line 338. Oct 4 10:21:17 roxy amavis[89209]: (!)_DIE: Suicide in child_init_hook: BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such process. at (eval 93) line 338. Oct 4 10:21:17 roxy amavis[89210]: (!!)TROUBLE in child_init_hook: BDB can''t connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such file or directory. at (eval 93) line 338. Oct 4 10:21:17 roxy amavis[89210]: (!)_DIE: Suicide in child_init_hook: BDB can''t connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such file or directory. at (eval 93) line 338. Oct 4 10:21:17 roxy amavis[89208]: (!!)TROUBLE in child_init_hook: register_proc: BDB N c_get: BDB0060 PANIC: fatal region error detected; run recovery, . at (eval 93) line 222. Oct 4 10:21:17 roxy amavis[89208]: (!)_DIE: Suicide in child_init_hook: register_proc: BDB N c_get: BDB0060 PANIC: fatal region error detected; run recovery, . at (eval 93) line 222. Oct 4 10:21:17 roxy amavis[89208]: (!)Amavis::END: DB unregistering failed: Oct 4 10:21:17 roxy amavis[89211]: (!!)TROUBLE in child_init_hook: BDB can''t connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such file or directory. at (eval 93) line 338. Oct 4 10:21:17 roxy amavis[89211]: (!)_DIE: Suicide in child_init_hook: BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such file or directory. at (eval 93) line 338. Oct 4 10:21:17 roxy amavis[89212]: (!!)TROUBLE in child_init_hook: BDB can't connect db env. at /var/amavis/db: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery, No such file or directory. at (eval 93) line 338. I tried to create empty /var/amavis/db with right permissions but it doesn't help. I found the same reports on the internet but no solution how to fix this. (enable_db=0 is temporary workaround but it disables some functions) https://lists.amavis.org/pipermail/amavis-users/2018-January/005182.html https://lists.amavis.org/pipermail/amavis-users/2018-April/005338.html https://forum.iredmail.org/topic12321-iredmail-support-amavisd-crashes-almost-daily-on-freebsd-11.html https://forum.iredmail.org/topic12034-iredmail-support-amavisd-suicide-in-childinithook-bdb-cant-connect-db-env-at-var.html Are there somebody with working setup of Amavisd-new on FreeBSD 11.2 with enable_db=1? Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: FreeBSD Port: databases/couchdb upgrade to 2.2
Kurt Jaeger wrote on 2018/09/10 12:06: Hi! are there any plans to create port for CouchDB 2.2? Yes, here's the PR about it: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=218844 Dave's a bit busy right now, so progress is slow. Great news. Thank you! Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
FreeBSD Port: databases/couchdb upgrade to 2.2
Hi, are there any plans to create port for CouchDB 2.2? According to latest vulnerability in 1.7.2 and statement on upstream website http://docs.couchdb.org/en/stable/cve/2018-11769.html there are no plans to fix it in 1.7, because this version is no longer supported. I am not able to create / maintain CouchDB 2.2 port by myself but I really would like to have not vulnerable version on our server. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: check_mssql_collection
Andrea Venturoli wrote on 2018/09/07 16:00: Hello. I was using a PHP based script to check MSSQL servers in my Nagios installations. However, it seems php56-mssql has no equivalent in he php71 world. So, while searching for an alternative, I found this: https://github.com/NagiosEnterprises/check_mssql_collection As anyone already tried porting it? If not I might give it a try. I cannot speak about Nagios but if you really want to use PHP to connect to MSSQL, you can use databases/php71-pdo_dblib Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
fix for vuln.xml / committer needed
Can somebody commit this easy fix, please? It is annoying to get false alarms every day in daily security reports. https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=231054 Kind Regards Miroslav Lachman Miroslav Lachman wrote on 2018/08/31 12:24: Miroslav Lachman wrote on 2018/08/28 00:20: Running pkg audit FreeBSD-10.4_11 gives me one vulnerability: # pkg audit FreeBSD-10.4_11 FreeBSD-10.4_11 is vulnerable: wpa_supplicant -- unauthenticated encrypted EAPOL-Key data CVE: CVE-2018-14526 WWW: https://vuxml.FreeBSD.org/freebsd/6bedc863-9fbe-11e8-945f-206a8a720317.html 1 problem(s) in the installed packages found. But information on the page shows it was fixed in 10.4-p10: Affected packages wpa_supplicant < 2.6_2 FreeBSD <= 10.4_10 FreeBSD <= 11.2_1 So... was it really fixed? Is there incorrect info in VuXML database file or on the web page? As noted privately by Dan Lukes, there is wrong entry in vuln.xml - missing < 10.4 and < 11.2 (start of the range) --- vuln.xml.orig 2018-08-30 03:02:57.656941000 +0200 +++ vuln.xml 2018-08-31 12:13:53.564345000 +0200 @@ -525,8 +525,8 @@ FreeBSD - 10.4_10 - 11.2_1 + 10.410.4_10 + 11.211.2_1 See PR 231054. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
PR 230393 - security/base-audit - committer needed
I am the author & maintainer of this port. Can somebody commit new version, please? https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230393 Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
pkg upgrade installs and then deinstalls package in one run
led: 9.4.19 -> 9.4.19 This is on FreeBSD 10.4-RELEASE-p9 amd64 GENERIC with pkg version 1.10.5 Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: how to make ports not install xorg or dependencies
tech-lists wrote on 2018/07/31 12:41: There used to be a way to enforce this no-xorg in make.conf but looking at /usr/share/examples/etc/make.conf I can find no reference to X Xorg x11 or xorg. I presume there's a new method. If there is, can anyone please tell me how? We are using OPTIONS_UNSET= X11 GUI CUPS DOCS EXAMPLES NLS for all of our packages (built with poudriere) As Guido Falsi already said it is not guaranteed that you will not have ports with some X libs, because some ports does not have option to disable X11 dependencies. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: libsmbclient broken after update
Ekkehard 'Ekki' Gehm wrote on 2018/07/25 11:05: Hey everybody, I do have an annoying issue here. I am not sure if I am just to stupit for that. In that case just give me a heads up ;-) After updating, it seems the smbclient is broken Situation: FreeBSD 11.1-RELEASE php72-7.2.8 php72-pecl-smbclient-0.9.0_3 nextcloud-php72-13.0.4 samba48-4.8.2_1 Problem: # su -m www -c "php ./occ status" PHP Warning: PHP Startup: Unable to load dynamic library 'smbclient.so' (tried: /usr/local/lib/php/20170718-zts/smbclient.so (Shared object "libsmbregistry-samba4.so" not found, required by "libsmbclient.so.0"), /usr/local/lib/php/20170718-zts/smbclient.so.so (Cannot open "/usr/local/lib/php/20170718-zts/smbclient.so.so")) in Unknown on line 0 The process control (PCNTL) extensions are required in case you want to interrupt long running commands - see http://php.net/manual/en/book.pcntl.php - installed: true - version: 13.0.4.0 - versionstring: 13.0.4 - edition: Can you find libsmbregistry-samba4.so somewhere in your system? Did you install packages from official FreeBSD repo or did you built them on your machine? It seems that you need Samba built with some proper options. I have Samba installed but I don't have libsmbregistry-samba4.so in my system. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: IBM MQ Client on FreeBSD
Alexander Leidinger wrote on 2018/07/17 16:29: [...] There's also an python MQ: - https://github.com/dsuch/pymqi - https://www.rabbitmq.com/devtools.html (there's a Python section) Note: I have used neither of them. https://github.com/dsuch/pymqi it still needs IBM MQ client installed cite: As a prerequisite, you first need to install an IBM MQ client in the system that PyMQI is about to be installed; it is a free library offered by IBM on top of which higher-level ones, such as PyMQI, can connect to queue managers. IBM MQ clients can be downloaded from IBM's website. Thank you anyway for your reply. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: IBM MQ Client on FreeBSD
Alexander Leidinger wrote on 2018/07/17 16:29: Quoting Miroslav Lachman <000.f...@quip.cz> (from Tue, 17 Jul 2018 16:00:52 +0200): Is there a way to install and use IBM MQ Client on FreeBSD? I cannot find any port or informations on IBM website. Our client needs to send some messages from Python powered project in to remote IBM MQ of another company. Not an answer to your question, but maybe a solution to your problem: You don't need an IBM MQ client to speak to an IBM MQ server (I would assume it's an IIB or WebSphere). There are other (even open source) clients around. One of them is RabbitMQ (java and erlang). There's also an python MQ: - https://github.com/dsuch/pymqi - https://www.rabbitmq.com/devtools.html (there's a Python section) Note: I have used neither of them. Are you sure they are compatible with IBM MQ on protocol level? App developers told me they need two versions of IBM MQ client, because they need to talk to v8 and v9 MQ. Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
IBM MQ Client on FreeBSD
Is there a way to install and use IBM MQ Client on FreeBSD? I cannot find any port or informations on IBM website. Our client needs to send some messages from Python powered project in to remote IBM MQ of another company. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Should a package restart on upgrade itself
Mathieu Arnold wrote on 2018/06/27 11:58: On Tue, Jun 26, 2018 at 01:27:49PM +0200, Miroslav Lachman wrote: Miroslav Lachman wrote on 2017/06/27 19:32: [...] Again and again and again... Can we have some written (or do we have?) policy to not stop/start/restart services from some @preunexec / @postexec targets? I really don't like that some packages are still shutting down or trying to restart in the middle of the pkg upgrade process. There is a policy, it MUST NOT be done. Do you have a link to docs where it is stated? I really would like to bookmark it so next time this discussion appear I will just post the link to docs ;) Services are stopped and started with the rc file they install, and pkg does it when its configuration variable HANDLE_RC_SCRIPTS is set to true. Please point out to ports doing this so that they can be fixed. I know about the open-vm-tools. I will let you know if I will find another. Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
Re: Should a package restart on upgrade itself
Josh Paetzel wrote on 2018/06/26 20:03: On Tue, Jun 26, 2018, at 6:27 AM, Miroslav Lachman wrote: Miroslav Lachman wrote on 2017/06/27 19:32: Matthias Fechner wrote on 2017/06/27 18:29: Dear all, it is always a pain if pkg upgrade a lot of packages to restart all services to make sure update/security fixes are applied to all running services. Is there an option in pkg that it restart services automatically or is it OK if I would add a post-install script to the packages (I maintain) that will include a "service foo restart"? What is best practice here? Please don't do this. Some ports did this in the past and this was really a pain during larger upgrades. It sometimes leave services stopped (hi MySQL). The same bad practice is disabling / enabling Apache modules on upgrade. pkg upgrade should just do it's work - upgrade packages on disk. But manipulating config files and restart of services is up to me - the Administrator (or my tools). It would be nice to have some kind of "hooks" in pkg, which can be used to notify deployment tools that some services should be (re)started, or do restart in some simpler environment if user allows this (setup hooks for service restart). But is must not be done automatically for individual ports / packages even if maintainer thinks it is Good Idea (tm) Again and again and again... Can we have some written (or do we have?) policy to not stop/start/restart services from some @preunexec / @postexec targets? I really don't like that some packages are still shutting down or trying to restart in the middle of the pkg upgrade process. One example from today upgrade: [87/96] Extracting open-vm-tools-nox11-10.2.5,2: .. done Stopping vmware_guestd. Waiting for PIDS: 516. Loading vmmemctl kernel module: already loaded. vmware_guestd not running? (check /var/run/vmware_guestd.pid). Starting vmware_guestd. Can committers take care of this bad behaviour and not commit things like this? https://svnweb.freebsd.org/ports/head/emulators/open-vm-tools/pkg-plist?revision=457485=markup @preunexec %%PREFIX%%/bin/vmware-rpctool 'tools.set.version 0' ; service vmware-guestd stop 2>/dev/null || /usr/bin/true @postexec service vmware-kmod restart && service vmware-guestd restart || /usr/bin/true Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org" Here's the diff for the commit you referenced: https://svnweb.freebsd.org/ports/head/emulators/open-vm-tools/pkg-plist?r1=457023=457485=457485 Which part are you objecting to? I don't really have any objections to changing open-vm-tools. I'll note that I inherited it in it's current state with regards to defaults and restarting, an it's probably worth fining out why it does those things before blatantly changing things. It's possible that open-vm-tools is a poor example of what you are talking about based on it providing services for the OS for running on VMWare versus running some application service or daemon , but I will have to think about this before taking a strongly held opinion. I am sorry, I was not talking about that revision. Only about @preunexec and @postexec. They were added in https://svnweb.freebsd.org/ports/head/emulators/open-vm-tools/pkg-plist?r1=388693=436703 https://svnweb.freebsd.org/ports/head/emulators/open-vm-tools/pkg-plist?r1=436816=444773 Stopping / Restarting any service in the midle of running pkg upgrade is Bad Thing (in my point of view). Note that "service vmware-kmod restart" is not doing what somebody may think it is - you end up with new version of vmware-guestd and old version of loaded kernel module, because it is not unloaded (and in situations like securelevel cannot be even loaded) Kind regards Miroslav Lachman ___ freebsd-ports@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
