packaging November 2015

packaging@lists.fedoraproject.org

25 participants
16 discussions

[Fedora-php-devel-list] Change proposal: drop runtime dependency on PEAR
by Remi Collet 25 Jan '16

25 Jan '16

Hi, For now all our PECL extensions have a dependency on php-pear. PEAR is mostly dead and most user don't want to install it. (The project is not dead, it even work with PHP 7, but most users have switched to composer) In fact, the pecl command is only used in scriptlets for extension (un)registration (to allow manual pecl/pear usage). Really nothing is required at runtime. So I'd like to remove dependency on /usr/bin/pecl for PECL extension. Here is a proposal for new scriptlets # when pear installed alone, after us %triggerin -- %{?scl_prefix}php-pear if [ -x %{__pecl} ] ; then %{pecl_install} %{pecl_xmldir}/%{name}.xml >/dev/null || : fi # posttrans as pear can be installed after us %posttrans if [ -x %{__pecl} ] ; then %{pecl_install} %{pecl_xmldir}/%{name}.xml >/dev/null || : fi %postun if [ $1 -eq 0 -a -x %{__pecl} ] ; then %{pecl_uninstall} %{pecl_name} >/dev/null || : fi Notice: pear still required at build time for macro definition. Notice; this is already used to also fix a circular dep issue (php-pecl-jsonc => php-pear => php-cli => php-common => php-pecl-jsonc) Minor issue, /var/lib/pear/pkgxml (%{pecl_xmldir}) is own by php-pear. So we also need to move this directory to php-common to avoid unowned dir when pear is not installed. Comments, before opening a FPC ticket ? Remi.

2 3

Fwd: tex dependencies on EPEL
by Gianluca Sforna 04 Dec '15

04 Dec '15

Hi all, I am packaging a largish rpm here: https://bugzilla.redhat.com/show_bug.cgi?id=804125 and I've got to build its documentation from sources using tex. In Fedora land, I managed to find a simple way to specify which packages I need as Buildrequries, like this: BuildRequires: tex(titlesec.sty) BuildRequires: tex(framed.sty) BuildRequires: tex(threeparttable.sty) BuildRequires: tex(wrapfig.sty) BuildRequires: tex(upquote.sty) BuildRequires: tex(multirow.sty) But this apparently does not work in EPEL 6 (no deps found) or EPEL 7 (tex(upquote.sty) not found). So, any suggestion about how to proceed here? -- Gianluca Sforna http://morefedora.blogspot.com http://plus.google.com/+gianlucasforna - http://twitter.com/giallu

3 5

Re: (fix typos) Re: Packaging:NamingGuidelines Re: DNF is completly unable to act with local packages
by Sérgio Basto 30 Nov '15

30 Nov '15

Switching to packaging(a)lists.fedoraproject.org On Ter, 2015-11-24 at 01:47 +0000, Sérgio Basto wrote: > On Seg, 2015-11-23 at 09:39 +0100, Dominik 'Rathann' Mierzejewski > wrote: > > On Sunday, 22 November 2015 at 00:46, Sérgio Basto wrote: > > > On Sex, 2015-11-20 at 15:18 +0100, Tomas Mraz wrote: > > > > On Čt, 2015-11-19 at 20:59 +0000, Sérgio Basto wrote: > > > > > On Qua, 2015-11-18 at 17:11 -0600, Jason L Tibbitts III > > > > > wrote: > > > > > > > > > > > "SB" == Sérgio Basto <sergio(a)serjux.com> writes: > > > > > > > > > > > > SB> When we fix the .spec and don't change the source, we > > > > > > bump > > > > > > rightmost > > > > > > SB> version, when we change the source, we bump the left > > > > > > version, > > > > > > so > > > > > > we > > > > > > SB> can distinguish when we update the source and when we > > > > > > updated > > > > > > the > > > > > > SB> .spec, this contrast for me is important. > > > > > > > > > > > > For me, the simple rule that a Release: tag less than 1 > > > > > > implies > > > > > > prerelease software, while a Release: tag of 1 or greater > > > > > > implies > > > > > > a > > > > > > post-release package, is important. So far the proponents > > > > > > of > > > > > > this > > > > > > change haven't shown what things would actually look like > > > > > > after > > > > > > this > > > > > > change, so it's hard for me to come up with a reason to > > > > > > change my > > > > > > opinion. > > > > > > > > > > prerelease numbering can't begin with 0 and increased to 0.1 > > > > > because : > > > > > > > > > > next version of foo-0.b would be foo-0.1.b and "b">1 > > > > > > > > Nope, 1>"b" in rpm version compare. > > > > Even so, we shouldn't depend on upstream preserving sorting order > > in > > their pre-release suffixes. Numerical sorting is always monotonous. > > > > > If so, we could begging numeration with 0 for pre-release: > > > > > > foo-0.c -> foo-0.c.1 -> foo-0.1.b -> foo-0.1.b.1 -> foo-0.2.a -> > > > foo- > > > 0.2.a.1 > > > > I don't understand why you want to introduce another level of > > numbering. > > What's wrong with the current guideline? > > I'd like improve for cases that upstream doesn't make a release and > the package stays forever in a pre-release, this happens a lot with > old projects that are half dead upstream, instead of have just one > counter, we have two counters, one when upstream change the source > other when we rebuild the package, it will be better readable, to > understand if the upstream had updates or not. In this link http://fedoraproject.org/wiki/Packaging:NamingGuidelines#Pre-Release_packag… where we read : Release Tag for Pre-Release Packages: 0.%{X}.%{alphatag}%{?dist} And I'm proposing : 0[.%{X}].%{alphatag}[.%{Y}]%{?dist} is just better IMHO . > Best regards, > -- > Sérgio M. B. > > -- Sérgio M. B.

2 2

Naming of Python 3 binaries
by Juan Orti Alcaine 26 Nov '15

26 Nov '15

Hi, I'm reading the guidelines about the naming of conflicting python2/python3 binaries [1], and I have doubts about if it is up to date. It says: "The unversioned executable must be the python2 version.", is this still true? In the example shown some lines before, the unversioned binary goes to the python3 package. With python3 as the default I guess the unversioned binary should be the python3 one. [1] https://fedoraproject.org/wiki/Packaging:Python#Naming -- Juan Orti https://apuntesderootblog.wordpress.com/

5 6

How to package systemd user instance services?
by Tadej Janež 23 Nov '15

23 Nov '15

Hi! First a little context. I'm in the process of packaging bup [1] (review request [2]), a backup system based on the git packfile format. One of its sub-commands/functionalities is "bup web" [3], a web server for browsing through bup repositories. By default, "bup web" will show the bup repository stored in user's ~/.bup directory. Thus, it seems most appropriate to provide a systemd user instance service which a user can enable and start. Since the current systemd packaging guidelines [4] don't mention user instance services at all, I've researched this a bit and found that the systemd package ships /usr/lib/rpm/macros.d/macros.systemd file, which contains various "user" macros: - %_userunitdir - %systemd_user_post - %systemd_user_preun - %systemd_user_postun - %systemd_user_postun_with_restart My naive attempt at using them was to follow [4] and just replace all % systemd_* calls with appropriate %systemd_user_* calls (and %_unitdir with %_userunitdir). The SPEC file can be viewed here: https://github.com/tjanez/bup-package/blob/5f4abed/bup.spec However, this doesn't work as expected. This is the output of rpmlint: Rpmlint ------- Checking: bup-0.27-0.3.fc24.x86_64.rpm bup-web-0.27-0.3.fc24.x86_64.rpm bup-0.27-0.3.fc24.src.rpm bup.x86_64: W: spelling-error Summary(en_US) packfile -> pack file, pack-file, packsaddle [... trimmed spelling warnings ...] bup-web.x86_64: W: empty-%postun bup.src: W: spelling-error Summary(en_US) packfile -> pack file, pack-file, packsaddle [... trimmed spelling warnings ...] bup.src: E: specfile-error systemd_post: invalid option -- '-' bup.src: E: specfile-error error: Unknown option - in systemd_post() 3 packages and 0 specfiles checked; 2 errors, 15 warnings. Rpmlint (installed packages) ---------------------------- bup-web.x86_64: W: empty-%postun 3 packages and 0 specfiles checked; 0 errors, 1 warnings. I would appreciate help and feedback on this topic. Thanks and regards, Tadej [1] https://bup.github.io/ [2] https://bugzilla.redhat.com/show_bug.cgi?id=1222334 [3] http://bup.github.io/man/bup-web.html [4] https://fedoraproject.org/wiki/Packaging:Systemd

2 3

Self Introduction : Dhanesh Sabane
by Dhanesh Bhalchandra Sabane 22 Nov '15

22 Nov '15

Hello everyone, I am Dhanesh Sabane, a computer science student from Pune, Maharashtra, India. I've been a Linux user since 2011 and a Fedora fan boy since only a couple of months. It didn't take long for the Fedora community to woo me there after. :P I'm genuinely interested in being a integral part of the community and would love to establish liaison with other contributors. I've a FAS account [1] which, I'm sure, will fill up with a huge list of contributions with the help of the community. Fedora packaging has especially caught my eye and right after I'm done with my exams, I'll be looking forward to learning more about Packaging and contributing to the Fedora Project. I'll also be looking for mentors in the package maintainers field and I would really appreciate it if anyone is willing to take up that role. :P I'm currently associated with Mozilla as a*Firefox Student Ambassador* and as a*Support Volunteer*. You'll find my contribution story on my blog [2]. Looking forward to working in collaboration and helping the Fedora Project grow further. Until then, think open! 1.https://fedoraproject.org/wiki/User:Dhanesh95 2.https://openctrl.wordpress.com/category/mozilla/ -- Dhanesh Bhalchandra Sabane IRC : dhanesh95

2 2

Python C extensions permisions and rpmlint
by Sergio Pascual 20 Nov '15

20 Nov '15

Hello, I'm starting to use the new Python macros in my packages. Very good job, specfiles are much more cleaner now. One thing that doesn't change is the rpmlint error about .so files having mode 775 instead of 755 For example: E non-standard-executable-perm /usr/lib64/python2.7/site-packages/sep.so 775 I routinely do a chmod 755 after %install, but now that I'm using %py{2.3}_install, could this step be part of the macro? Or if the rpmlint error is inaccurate and 775 permisions are OK, fix rpmlint? Best, Sergio

3 2

Summary/Minutes from today's FPC Meeting (2015-11-19 17:00 - 17:50 UTC)
by James Antill 19 Nov '15

19 Nov '15

====================== #fedora-meeting-1: fpc ====================== Meeting started by geppetto at 17:00:22 UTC. The full logs are available at http://meetbot.fedoraproject.org/fedora-meeting-1/2015-11-19/fpc.2015-11-19… . Meeting summary --------------- * Roll Call (geppetto, 17:00:22) * Schedule (geppetto, 17:04:03) * LINK: https://lists.fedoraproject.org/archives/list/packaging% 40lists.fedoraproject.org/thread/L3QD2JPSIIFN2M6BHLJJVLRHTYNR2LFN/ (geppetto, 17:04:07) * #579 Mention abipkgdiff and abidiff instead of abi-compliance-checker (geppetto, 17:04:40) * LINK: https://fedorahosted.org/fpc/ticket/579 (geppetto, 17:04:54) * ACTION: Dodji Will tweak the draft for the week after next. (geppetto, 17:36:24) * #558 Application/Library distinction and package splitting (geppetto, 17:39:49) * LINK: https://fedorahosted.org/fpc/ticket/558 (geppetto, 17:40:00) * Open Floor (geppetto, 17:45:46) Meeting ended at 17:51:29 UTC. Action Items ------------ * Dodji Will tweak the draft for the week after next. Action Items, by person ----------------------- * Dodji * Dodji Will tweak the draft for the week after next. * **UNASSIGNED** * (none) People Present (lines said) --------------------------- * Dodji (45) * geppetto (45) * Rathann (22) * orionp (13) * zodbot (10) * racor (7) * mbooth (2) Generated by `MeetBot`_ 0.1.4 .. _`MeetBot`: http://wiki.debian.org/MeetBot

1 0

Schedule for Thursday's FPC Meeting (2015-11-19 16:00 UTC)
by James Antill 18 Nov '15

18 Nov '15

Following is the list of topics that will be discussed in the FPC meeting Thursday at 2015-11-19 16:00 UTC in #fedora-meeting-1 on irc.freenode.net. Local time information (via. rktime): 2015-11-19 09:00 Thu US/Pacific PST 2015-11-19 12:00 Thu US/Eastern EST 2015-11-19 17:00 Thu UTC <- 2015-11-19 17:00 Thu Europe/London <- 2015-11-19 18:00 Thu Europe/Paris CET 2015-11-19 18:00 Thu Europe/Berlin CET 2015-11-19 22:30 Thu Asia/Calcutta IST ------------------new day---------------------- 2015-11-20 01:00 Fri Asia/Singapore SGT 2015-11-20 01:00 Fri Asia/Hong_Kong HKT 2015-11-20 02:00 Fri Asia/Tokyo JST 2015-11-20 03:00 Fri Australia/Brisbane EST Links to all tickets below can be found at: https://fedorahosted.org/fpc/report/13 = Followups = #topic #558 Application/Library distinction and package splitting .fpc 558 https://fedorahosted.org/fpc/ticket/558 #topic #566 RPM file triggers .fpc 566 https://fedorahosted.org/fpc/ticket/566 #topic #567 Packaging Python 3 applications and modules for EPEL 7+ .fpc 567 https://fedorahosted.org/fpc/ticket/567 = New business = #topic #579 Mention abipkgdiff and abidiff instead of abi-compliance-checker .fpc NNN https://fedorahosted.org/fpc/ticket/NNN = Open Floor = For more complete details, please visit each individual ticket. The report of the agenda items can be found at: https://fedorahosted.org/fpc/report/13 If you would like to add something to this agenda, you can reply to this e-mail, file a new ticket at https://fedorahosted.org/fpc, e-mail me directly, or bring it up at the end of the meeting, during the open floor topic. Note that added topics may be deferred until the following meeting.

2 2

For a cross-compiling toolchain, does it matter if /usr/bin/* links are symbolic or hard?
by Jonathan Dieter 16 Nov '15

16 Nov '15

I'm reviewing nacl-binutils (https://bugzilla.redhat.com/show_bug.cgi?i d=1270355), which has hard links from /usr/x86_64-nacl/* to /usr/bin/x86_64-nacl-*. According to https://fedoraproject.org/wiki/Pa ckaging_Cross_Compiling_Toolchains, these should be symlinks, and rpmlint complains about cross-directory-hard-links. Is there any reason to convert these to symlinks or can we just leave them as hard links? Thanks, Jonathan

4 8

Summary/Minutes from today's FPC Meeting (2015-11-12 17:00 - 18:05 UTC)
by James Antill 12 Nov '15

12 Nov '15

====================== #fedora-meeting-1: fpc ====================== Meeting started by geppetto at 17:00:33 UTC. The full logs are available at http://meetbot.fedoraproject.org/fedora-meeting-1/2015-11-12/fpc.2015-11-12… . Meeting summary --------------- * Roll Call (geppetto, 17:00:34) * Schedule (geppetto, 17:09:14) * LINK: https://lists.fedoraproject.org/pipermail/packaging/2015-November/011086.ht… (geppetto, 17:09:19) * #540 Define gcc and gcc-c++ as providing minimal C and C++ (geppetto, 17:09:38) * LINK: https://fedorahosted.org/fpc/ticket/540 (geppetto, 17:09:40) * #540 Define gcc and gcc-c++ as providing minimal C and C++ compilation environments. (geppetto, 17:09:56) * LINK: https://fedorahosted.org/fpc/ticket/540 (geppetto, 17:10:09) * ACTION: (geppetto, 17:23:46) * ACTION: Promote C and C++ page and link to it from main guidelines (+1:6, 0:0, -1:0) (geppetto, 17:24:18) * Open Floor (geppetto, 17:36:05) Meeting ended at 18:05:16 UTC. Action Items ------------ * * Promote C and C++ page and link to it from main guidelines (+1:6, 0:0, -1:0) Action Items, by person ----------------------- * **UNASSIGNED** * * Promote C and C++ page and link to it from main guidelines (+1:6, 0:0, -1:0) People Present (lines said) --------------------------- * geppetto (73) * tibbs|w (65) * nirik (13) * zodbot (11) * racor (8) * Rathann (8) * tomspur (7) * mbooth (5) * orionp (5) * gbcox (1) * tibbs (0) Generated by `MeetBot`_ 0.1.4 .. _`MeetBot`: http://wiki.debian.org/MeetBot

1 0

Schedule for Thursday's FPC Meeting (2015-11-11 17:00 UTC)
by James Antill 11 Nov '15

11 Nov '15

Following is the list of topics that will be discussed in the FPC meeting Thursday at 2015-11-11 16:00 UTC in #fedora-meeting-1 on irc.freenode.net. Local time information (via. rktime): 2015-11-11 09:00 Wed US/Pacific PST 2015-11-11 12:00 Wed US/Eastern EST 2015-11-11 17:00 Wed UTC <- 2015-11-11 17:00 Wed Europe/London <- 2015-11-11 18:00 Wed Europe/Paris CET 2015-11-11 18:00 Wed Europe/Berlin CET 2015-11-11 22:30 Wed Asia/Calcutta IST ------------------new day---------------------- 2015-11-12 01:00 Thu Asia/Singapore SGT 2015-11-12 01:00 Thu Asia/Hong_Kong HKT 2015-11-12 02:00 Thu Asia/Tokyo JST 2015-11-12 03:00 Thu Australia/Brisbane EST Links to all tickets below can be found at: https://fedorahosted.org/fpc/report/13 = Followups = #topic #540 Define gcc and gcc-c++ as providing minimal C and C++ compilation environments. .fpc 540 https://fedorahosted.org/fpc/ticket/540 #topic #558 Application/Library distinction and package splitting .fpc 558 https://fedorahosted.org/fpc/ticket/558 #topic #566 RPM file triggers .fpc 566 https://fedorahosted.org/fpc/ticket/566 #topic #567 Packaging Python 3 applications and modules for EPEL 7+ .fpc 567 https://fedorahosted.org/fpc/ticket/567 = Open Floor = For more complete details, please visit each individual ticket. The report of the agenda items can be found at: https://fedorahosted.org/fpc/report/13 If you would like to add something to this agenda, you can reply to this e-mail, file a new ticket at https://fedorahosted.org/fpc, e-mail me directly, or bring it up at the end of the meeting, during the open floor topic. Note that added topics may be deferred until the following meeting.

1 0

Force fedora-review to redownload repo metadata
by Julien Enselme 10 Nov '15

10 Nov '15

Hi, I am trying to do a review of a package which requires dependencies that landed in rawhide yesterday. My local fedora-review/mock doesn't know about it yet and so it fails to setup the environment: No matching package to install: 'python2-h5io' No matching package to install: 'python2-jdcal' No matching package to install: 'python3-h5io' No matching package to install: 'python3-jdcal' To get rid off the cache manually, I tried: - rm -rf /var/cache/yum/* - rm -rf /var/cache/dnf/* - dnf update --refresh - rm -rf /var/cache/mock/* - rm -rf /var/lib/mock/fedora-rawhide-x86_64 - mock -r fedora-rawhide-x86_64 --init To make fedora-review to it, I tried: - fedora-review -m fedora-rawhide-x86_64 -b 1276880 --mock-options '--clean --scrub=all ' fedora-review -m fedora-rawhide-x86_64 -b 1276880 --mock-options '--scrub=all ' Is there anyway do to it? It is getting very frustrating. Regards, -- Julien Enselme aka Jujens http://www.jujens.eu/

4 6

review swap
by Antonio Trande 10 Nov '15

10 Nov '15

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Hi all, i have a couple of review requests in waiting: https://bugzilla.redhat.com/show_bug.cgi?id=1215354 https://bugzilla.redhat.com/show_bug.cgi?id=1267189 I'm willing to exchange them for two other. Thanks. - -- Antonio Trande mailto: sagitter 'at' fedoraproject 'dot' org http://fedoraos.wordpress.com/ https://fedoraproject.org/wiki/User:Sagitter GPG Key: 0x565E653C Check on https://keys.fedoraproject.org/ -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAEBCAAGBQJWQefRAAoJEF5tK7VWXmU80kAH/AkPOhcK/NTp1D5uVaSeA1/4 tfrDAnjfBOj3YIyNyLlU1NU8PmRs48wqjdC0CBsiiyHIYyJ3rbpQS2xlgRdjAory DP5llmefFRw8VqHJo4WP9TRSH3I21NjOAfV1X5U6pvGhSiI2QSK+LazhiSZjEg92 RrdqucJAZtoDgoIZeXhGrGItlF0pTr58om7MzaSm0wMXAkqvYRxSKnKd9BgpokCW SodGWARvD5CQe3SsHA58boPCdiGadfCMBX7Js+Viti+GtxLcizt/FLMFKp5ZUPy1 VWEPIpcteJqRUUwLnjBNZC0zt8WDOynowasdEXU2efrHI0ulAH+8tQ3vLN30USA= =UT0W -----END PGP SIGNATURE-----

1 0

SELinux packing page example results in fedora-review error
by myk 07 Nov '15

07 Nov '15

Hi, According to https://fedoraproject.org/wiki/SELinux_Policy_Modules_Packaging _Draft in the "Building the Policy Modules" section the example approach to build a SELinux policy is: %build ... cd SELinux for selinuxvariant in %{selinux_variants} do make NAME=${selinuxvariant} -f /usr/share/selinux/devel/Makefile mv mymodule.pp mymodule.pp.${selinuxvariant} make NAME=${selinuxvariant} -f /usr/share/selinux/devel/Makefile clean done So for https://bugzilla.redhat.com/show_bug.cgi?id=1209166, in palo.spec in the %build section, I now looking to use: #Use cmake to build %cmake ../%{name}-%{version}/ make %{?_smp_mflags} cd - #Build the SELinux policy cd SELinux for selinuxvariant in %{selinux_variants} do make NAME=${selinuxvariant} -f /usr/share/selinux/devel/Makefile mv %{name}.pp %{name}.pp.${selinuxvariant} make NAME=${selinuxvariant} -f /usr/share/selinux/devel/Makefile clean done But, when running fedora-review, I get the error message: [!]: Uses parallel make %{?_smp_mflags} macro. I tried to include %{?_smp_mflags} in the SELinux's Makefile make call, but this always resulted in an error during rpmbuild. Since the %{?_smp_mflags} is useful, either 'make' needs to accept -j against the SELinux Makefile (and the SELinux Packaging page updated), or fedora-review needs to exclude SELinux Makefile lines when testing for the %{?_smp_mflags}. As I don't know what would involved in either approach, would like some guidance on whether filing a bug report against fedora-review is the most appropriate next step. Kind regards, Michael

2 1

FPC meeting for 2015-11-05 canceled
by Jason L Tibbitts III 04 Nov '15

04 Nov '15

Many folks on the committee will be unavailable tomorrow and there's only one item of business which needs consideration: https://fedorahosted.org/fpc/ticket/540 Define gcc and gcc-c++ as providing minimal C and C++ compilation environments Since our chances of reaching quorum are low, I'll go ahead and cancel the meeting. I would urge FPC folks to comment on that ticket if you have anything to add so they're not waiting too long for input. - J<

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

packaging November 2015