DevHeads.net

Does anybody care about gettext?

Well, it was retired because it did not built since F30 mass rebuild…

* petersen (Jens Petersen)
* ueno (Daiki Ueno)
* praiskup (Pavel Raiskup)
* suanand (Sundeep Anand)
* nphilipp (Nils Philippsen)
* jjanco (Jakub Janco)

6 maintainers could not fix FTBFS for half a year? Anybody is
interested to maintain gettext?

Comments

Re: Does anybody care about gettext?

By Jens Petersen at 08/09/2019 - 07:28

On Fri, Aug 9, 2019 at 9:27 AM Igor Gnatenko <

I went ahead and built it with the testsuite disabled for now: I suppose
any Proven Packager could also have done, but yeah normally it should be
done by the maintainers.
I admit the ball was dropped on this by various people (myself included),
and sorry about that. [1]
The new major upstream release was also a long time coming...

But to me the deeper question is still "why are we proactively breaking the
distro" in this way with package retirements by non-maintainers?

Sure FTBFS is bad but there is no need to proactively remove core packages
which are still working okay.
I really really wish could stop this... causing more busy work and stress.

Jens

[1] short version is: there was a partial package owner handover which was
not properly completed unfortunately.

Re: Does anybody care about gettext?

By =?UTF-8?B?TWlyb... at 08/09/2019 - 08:50

On 09. 08. 19 14:28, Jens-Ulrik Petersen wrote:
Because if we don't, people just gonna ignore FTBFS forever.

Could we have known that gettext will be unretired right away anyway? Probably
yes, but nobody got time / energy / resources to do any kind of analysis of the
FTBFS packages.

Next time, I hope that FTBFS bugs for critical component are actually actively
solved sooner than the retirement happens. We can try to be more aggressive with
the reminders, but I don't know if that helps, because even currently, packages
just switch the Bugzilla to ASSIGNED to stop them.

Re: Does anybody care about gettext?

By Kevin Kofler at 08/10/2019 - 17:27

Miro Hrončok wrote:
And this would be a problem why exactly?

Packages built for older Fedora releases tend to run on newer Fedora
releases just fine. If the package:
* has no broken dependencies, and
* is not reported as completely broken in Bugzilla,
it should be assumed to be working by default.

If you encounter a package that does not actually work, it is your job as a
user to report it. This can happen independently of whether the package
still compiles or not. It can even still be broken after a successful
rebuild.

So why are we wasting packagers' time on fixing FTBFS issues (which are
typically NOT caused by their package, but by changes in dependencies such
as your python-unversioned-command change, or such as yet another
incompatible change in GCC for the sake of compliance with some obscure
subparagraph of a language standard, etc.) when not actually needed? I
actually NEED to fix the FTBFS if the package has broken dependencies or if
I need to make some other change to it. Otherwise, the FTBFS fix is just
churn that Fedora forces me to waste my time on.

Kevin Kofler

Re: Does anybody care about gettext?

By Dridi Boukelmoune at 08/10/2019 - 18:19

On Sat, Aug 10, 2019 at 10:28 PM Kevin Kofler <kevin. ... at chello dot at> wrote:
Hi,

I understand where you are coming from, but I still disagree. I think
there has been an unfair hostility towards Miro on this topic.

Your package suddenly FTBFS because of dependencies or system-wide
changes but the latest package build is still functional? I agree that
there is no urgency to fix this, but I disagree that status quo is
fine X releases later.

For starters may miss out on system wide changes (and whether someone
agrees with proposed changes is not the question) and in the case you
made about bug reports that mandate action from the maintainer, not
taking care of FTBFS timely means that once shit hits the fan you have
to both solve the FTBFS situation and the user-facing bug report.

So yes, it sucks when someone's package fails because someone else
screwed up by not coordinating an soname bump or whatever, but it
doesn't mean that we can keep the latest successful build around and
let the source repository bit-rot forever because there are no bug
reports.

Now there's certainly room for improvement, but I don't have a
solution to offer and hammering Miro because he's been (very) active
lately retiring FTBFS or orphan packages (as per the normal process)
is helping. Here we had an acknowledgement from a couple maintainers
and someone who stepped in to help, a very positive outcome.

Dridi

Re: Does anybody care about gettext?

By Michael Cronenworth at 08/09/2019 - 13:07

On 8/9/19 8:50 AM, Miro Hrončok wrote:
Instead of spending your time on better nag e-mails why don't you look at
contributing patches?

Re: Does anybody care about gettext?

By Stephen John Smoogen at 08/09/2019 - 14:02

Miro does contribute a lot of patches. There are still a lot of packages
which need more than just a patch but someone to take ownership. If you
have a problem with these actions please bring them up with FESCO or the
Council as this was brought up previously in both places as things that
needed to be fixed. They can then work out how this is to be dealt with in
the future.

Re: Does anybody care about gettext?

By Adam Williamson at 08/09/2019 - 13:38

On Fri, 2019-08-09 at 13:07 -0500, Michael Cronenworth wrote:
Do you know he doesn't? But regardless, one person can run an FTBFS
notification system. One person cannot fix all the FTBFS bugs.

It's similar to the blocker bug process: a couple of us in QA can
handle discovering, reporting and tracking blocker bugs. We can't *fix*
them all (though I try to fix as many as I have time and ability to).

Re: Does anybody care about gettext?

By Simo Sorce at 08/09/2019 - 09:29

On Fri, 2019-08-09 at 15:50 +0200, Miro Hrončok wrote:
If you send direct emails to package owner specifying what package is
broken maybe yes.

If you send multikilobyte emails generically stating FTBFS and parse
yourself this list of 1000 packages to see if one is yours, then no,
they go to trash directly.

Simo.

Re: Does anybody care about gettext?

By Adam Williamson at 08/09/2019 - 10:10

On Fri, 2019-08-09 at 10:29 -0400, Simo Sorce wrote:
The bug reports do that.

Re: Does anybody care about gettext?

By =?UTF-8?B?TWlyb... at 08/09/2019 - 09:52

On 09. 08. 19 16:29, Simo Sorce wrote:
Patches welcome :D Even now, the reminders are semiautomatical and painful.
Sending direct e-mail is impossible without proper automation. And it seems all
our automation from years ago is now nonfunctional.

Re: Does anybody care about gettext?

By Brian (bex) Exe... at 08/09/2019 - 10:16

On Fri, Aug 9, 2019 at 5:06 PM Miro Hrončok < ... at redhat dot com> wrote:
FWIW, python3-mailmerge might be useful here as the email would be
direct, personalized, and not subject to filters around BZ mail.

regards,

bex

Re: Does anybody care about gettext?

By Daniel P. Berrange at 08/09/2019 - 08:00

On Fri, Aug 09, 2019 at 02:28:55PM +0200, Jens-Ulrik Petersen wrote:
When a FTBFS hits, we don't know whether the package is still working
ok or not as there are many possible reasons for the failure. Filing
the FTBFS BZs informs the maintainer(s) & allows them to investigate,
figure what has gone wrong & decide what changes are needed. Missing
on 2 mass rebuilds means the package is still build with F29 toolchain,
and thus lacking desired improvements Fedora is introducing, so this
has a cost for the rest of the distro. Somewhere there's a balance
between cost for the maintainer in work & cost for the distro in the
package being outdated.

There was no acknowlegement on the BZ that anyone was actively working
on fixing it in 6 months. This is true for so many of the FTBFS BZs that
get filed. If the packages don't get orphaned after 6+ months of being
ignored, when would they ever be fixed ?

Having said that, I think in the case of packages which are deps of
so much of the distro, it could have been useful to have a warning of
imminent orphaning on fedora-devel. There was a warning that orphaning
was starting, but no list of affected packages included.

If a package list was included, *non-maintainers* might have noticed that
gettext was at risk & would massively impact the distro & could have
stepped up to help solve it, where the original maintainer drops the ball.

Overall though, despite the disruption, orphaning has had the intended
effect of getting the long standing FTBFS problem resolved.

Regards,
Daniel

Re: Does anybody care about gettext?

By Martin Kolman at 08/09/2019 - 08:13

On Fri, 2019-08-09 at 14:00 +0100, Daniel P. Berrangé wrote:
Before dropping a batch of packages, do a test compose without them and postpone
the drop if the compose run crashes and burns.

Sounds really like something doable which could save a lot of everyones time once in place.

Re: Does anybody care about gettext?

By Daniel P. Berrange at 08/09/2019 - 09:14

On Fri, Aug 09, 2019 at 03:13:07PM +0200, Martin Kolman wrote:
The gating CI is a pretty low bar right now. As CI is made stronger,
it could well actually make FTBFS *more* common, as CI is introducing
more scope for things to be classed as a failure. So be careful what
you wish for :-)

We can't carry on postponing things indefinitely though - at some point we
have to say enough, and expect a maintainer to actually do some maintaining.

Regards,
Daniel

Re: Does anybody care about gettext?

By Paul Wouters at 08/09/2019 - 09:58

That is an argument to orphan, not an argument to remove the package.
Had gettext been orphaned, people would have noticed without the entire
OS breakage.

Paul

Re: Does anybody care about gettext?

By =?UTF-8?B?TWlyb... at 08/09/2019 - 10:16

On 09. 08. 19 16:58, Paul Wouters wrote:
The breakage took ~3 several hours to notice and a 5 minute fix.

I think it was not that catastrophic.

Re: Does anybody care about gettext?

By Martin Kolman at 08/09/2019 - 09:22

On Fri, 2019-08-09 at 15:14 +0100, Daniel P. Berrangé wrote:
I'm just trying to find ways that can sound the alarm bells &
prevent everyone impacting Rawhide breakage before it's too late
and things need to be fixed post-mortem. An this case really looks
like something that an automated check should be able to catch soon
enough. :)

Re: Does anybody care about gettext?

By =?UTF-8?B?TWlyb... at 08/09/2019 - 09:50

On 09. 08. 19 16:22, Martin Kolman wrote:
Tough question: Would the package get any attention if it was not retired?

Re: Does anybody care about gettext?

By Sundeep Anand at 08/09/2019 - 03:02

Hi,

most probably I’ll fix that by next week.

thanks,
sundeep

Re: Does anybody care about gettext?

By Leigh Scott at 08/09/2019 - 03:34

Maybe package the newer version 0.20.1 <a href="http://www.gnu.org/software/gettext/" title="http://www.gnu.org/software/gettext/">http://www.gnu.org/software/gettext/</a>

Re: Does anybody care about gettext?

By Alexander Ploumistos at 08/09/2019 - 03:09

Hello,

On Fri, Aug 9, 2019 at 11:03 AM Sundeep Anand < ... at redhat dot com> wrote:
All the patches we carried were merged back in the latest upstream
version (0.20.1), but when I took a stab at it, I got a lot of errors
about the variable types and I did not know how to proceed.

Best of luck

P.S.: Can anyone explain how the builds of other packages went back to
normal, even though there was no successful gettext build?

Re: Does anybody care about gettext?

By Jerry James at 08/09/2019 - 15:10

On Fri, Aug 9, 2019 at 2:12 AM Alexander Ploumistos
<alex. ... at gmail dot com> wrote:
Alexander, try the attached spec file. It leads to a successful build
for me. (Sorry for doing this in email, but I'm currently in a place
where I can't make git pull requests.) It may need a little tweaking
still, and feel free to trim my verbose changelog entry if you decide
to use any part of it. :-)

Regards,

Re: Does anybody care about gettext?

By Rafal Luzynski at 08/09/2019 - 18:12

9.08.2019 22:10 Jerry James < ... at gmail dot com> wrote:
I tried this as well and stumbled upon this bug:
<a href="https://savannah.gnu.org/bugs/?55356" title="https://savannah.gnu.org/bugs/?55356">https://savannah.gnu.org/bugs/?55356</a>

I am not Alexander but since I've already started working on this
I think I can take a look. Your spec file looks mostly good. I have
some doubts which may be completely wrong (I have never looked at the
internals of gettext before) but just in case here we go:

Before Jens' commit there was also a line:

%bcond_without check

If you remove it you enable the check phase unconditionally.

GFDL is a new thing here, I guess the public announce of the license
change will be needed.

Do we need to change ftp to https?

"BuildRequires: chrpath" removed here, really no longer needed?

It worked for me without these packages although the results
might have been different due of that.

Same here.

Good point.

A comment why GFDL has been added here would be helpful.
Maybe it is not needed here?

Looks great. Are the licenses correct?

Is this the reason why chrpath is no longer needed?

Previously the %check was conditional (with "%if %{with check}").
See also my comment on top.

Is it OK to list all these files explicitly?

In other subpackages the documentation (also in HTML format)
is moved to a directory htmldoc (although I am not sure why).

It would be nice to mention the bug report rhbz#1708013.

Also see this comment which suggests adding an upstream patch:
<a href="https://bugzilla.redhat.com/show_bug.cgi?id=1708013#c2" title="https://bugzilla.redhat.com/show_bug.cgi?id=1708013#c2">https://bugzilla.redhat.com/show_bug.cgi?id=1708013#c2</a>

Regards,

Rafal

Re: Does anybody care about gettext?

By =?iso-8859-1?q?... at 08/10/2019 - 06:12

Rafal Luzynski wrote:
That's the wrong question to ask. The right question is: What reason is
there to choose an insecure protocol when HTTPS is available?

Anyway, the answer is yes:

220 GNU FTP server ready.
USER anonymous
230-NOTICE (Updated October 13 2017):
230-
230-Because of security concerns with plaintext protocols, we still
230-intend to disable the FTP protocol for downloads on this server
230-(downloads would still be available over HTTP and HTTPS), but we
230-will not be doing it on November 1, 2017, as previously announced
230-here. We will be sharing our reasons and offering a chance to
230-comment on this issue soon; watch this space for details.
230-
230-If you maintain scripts used to access ftp.gnu.org over FTP,
230-we strongly encourage you to change them to use HTTPS instead.

Björn Persson

Re: Does anybody care about gettext?

By Kevin Fenzi at 08/10/2019 - 11:43

On 8/10/19 4:12 AM, Björn Persson wrote:
I'm confused by this... https is already being used, it is just the
hostname that is 'ftp'. So, I would see no reason to change this, but I
suppose you could ask upstream to rename the host to avoid confusion...

kevin

Re: Does anybody care about gettext?

By =?iso-8859-1?q?... at 08/10/2019 - 14:48

Kevin Fenzi wrote:
When I posted, gettext.spec in the master branch still said:

Source: ftp://ftp.gnu.org/gnu/gettext/%{name}-%{tarversion}.tar.xz

The spec that Jerry James posted changed the URI scheme to https, which
is the right thing to do. That change is now also in Git.

The URL field should also be changed by the way. <a href="http://www.gnu.org" title="www.gnu.org">www.gnu.org</a> redirects
HTTP requests to HTTPS, but there is no reason to give an attacker the
chance to intercept the HTTP request and redirect you to a malicious
server instead.

Björn Persson

Re: Does anybody care about gettext?

By Kevin Fenzi at 08/10/2019 - 15:41

On 8/10/19 12:48 PM, Björn Persson wrote:
Alright, I was going from the quoted text.

Sure, 100% agree.

kevin

Re: Does anybody care about gettext?

By Nico Kadel-Garcia at 08/10/2019 - 11:51

On Sat, Aug 10, 2019 at 12:48 PM Kevin Fenzi < ... at scrye dot com> wrote:
The site still supports FTP, which I suspect is why they use the same
name for both services. Check out <a href="ftp://ftp.gnu.org/pub/gnu/" title="ftp://ftp.gnu.org/pub/gnu/">ftp://ftp.gnu.org/pub/gnu/</a>

Re: Does anybody care about gettext?

By Rafal Luzynski at 08/10/2019 - 06:31

10.08.2019 13:12 Björn Persson <Bjorn@rombobjörn.se> wrote:
This answers the question, thanks.

Rafal

Re: Does anybody care about gettext?

By Sundeep Anand at 08/09/2019 - 17:33

Not sure, quick (however, may be, improper) work-around could be fixing
formatting as
fprintf (a_fp, "%s", str) in libtextstyle/lib/libcroco/cr-statement.c

By this time *BuildRequires* looks like:
...
BuildRequires: teckit
BuildRequires: texlive-dvips
BuildRequires: texlive-dvipdfmx
BuildRequires: texinfo-tex
BuildRequires: texlive-xetex
...
Is teckit required? (perhaps teckit could not build from source:
<a href="https://src.fedoraproject.org/rpms/teckit/tree/master" title="https://src.fedoraproject.org/rpms/teckit/tree/master">https://src.fedoraproject.org/rpms/teckit/tree/master</a>) Or am I missing
something?

thanks,
sundeep

Re: Does anybody care about gettext?

By Alexander Ploumistos at 08/09/2019 - 17:29

On Fri, Aug 9, 2019 at 11:13 PM Jerry James < ... at gmail dot com> wrote:
Wow, thank you so much Jerry!
This should be really helpful for the gettext maintainers as well.

Best regards,
Alex

Re: Does anybody care about gettext?

By =?UTF-8?B?TWlyb... at 08/09/2019 - 04:54

On 09. 08. 19 10:09, Alexander Ploumistos wrote:
I've requested the package to be unretired and when that happened, I've retagged
the latest build to f31-pending. That eventually moved it back to f31.

Re: Does anybody care about gettext?

By Sundeep Anand at 08/09/2019 - 03:32

On Fri, Aug 9, 2019 at 1:40 PM Alexander Ploumistos <

I noticed that, thanks :)

Re: Does anybody care about gettext?

By =?UTF-8?B?TWlyb... at 08/09/2019 - 03:02

BTW I've just orphaned it, so hopefully somebody will take it and actually make it build.