Bug#908323: libgtk3-perl: FTBFS: t/overrides.t failure

2018-09-13 Thread gregor herrmann
On Thu, 13 Sep 2018 13:16:46 +0100, Iain Lane wrote:

> On Wed, Sep 12, 2018 at 11:59:49PM +0200, gregor herrmann wrote:
> > That's great news, thanks.
> > Please give us a short ping once the fixed package is available.
> I've just uploaded gdk-pixbuf 2.38.0+dfsg-5. There will be the normal
> delay, but hopefully this fixes the issue for you.

Thanks!

Indeed libgtk3-perl now seems to work with my previous patch and
gir1.2-gdkpixbuf-2.0 2.38.0+dfsg-5.
I'll upload in a second.


Cheers,
gregor

-- 
 .''`.  https://info.comodo.priv.at -- Debian Developer https://www.debian.org
 : :' : OpenPGP fingerprint D1E1 316E 93A7 60A8 104D  85FA BB3A 6801 8649 AA06
 `. `'  Member VIBE!AT & SPI Inc. -- Supporter Free Software Foundation Europe
   `-   NP: Gypsy Sun And Rainbows: Izabella


signature.asc
Description: Digital Signature


Bug#908323: libgtk3-perl: FTBFS: t/overrides.t failure

2018-09-13 Thread Iain Lane
On Wed, Sep 12, 2018 at 11:59:49PM +0200, gregor herrmann wrote:
> That's great news, thanks.
> Please give us a short ping once the fixed package is available.

I've just uploaded gdk-pixbuf 2.38.0+dfsg-5. There will be the normal
delay, but hopefully this fixes the issue for you.

Cheers,

-- 
Iain Lane  [ i...@orangesquash.org.uk ]
Debian Developer   [ la...@debian.org ]
Ubuntu Developer   [ la...@ubuntu.com ]


signature.asc
Description: PGP signature


Bug#908323: libgtk3-perl: FTBFS: t/overrides.t failure

2018-09-12 Thread gregor herrmann
On Wed, 12 Sep 2018 13:42:16 +0100, Iain Lane wrote:

> > I tried with
> > 
> > #v+
> > --- a/lib/Gtk3.pm
> > +++ b/lib/Gtk3.pm
> > [...]
> > #v-
> 
> This diff looks correct to me - I think it could be forwarded upstream
> and/or uploaded.

Thanks!
Uploading right now is a bit difficult as the tests still fail due to
the next issue :) But I'll prepare it as a proper patch.
 
> > but then the tests failed differently, with:
> > 
> > Gtk3::Gdk::Pixbuf::new_from_inline: passed too few parameters (expected 4, 
> > got 3) at t/overrides.t line 762.
> > # Looks like your test exited with 255 just after 162.
> 
> ...and this one is a real bug in gdk-pixbuf. I forwarded a potential fix
> to upstream - will wait a little while to see what they say. I've
> confirmed that after applying the patch already in this bug and that
> proposed patch libgtk3-perl's autopkgtests pass again. Once it gets an
> ack or if there is no reply then I'll upload to Debian.

That's great news, thanks.
Please give us a short ping once the fixed package is available.


Cheers,
gregor

-- 
 .''`.  https://info.comodo.priv.at -- Debian Developer https://www.debian.org
 : :' : OpenPGP fingerprint D1E1 316E 93A7 60A8 104D  85FA BB3A 6801 8649 AA06
 `. `'  Member VIBE!AT & SPI Inc. -- Supporter Free Software Foundation Europe
   `-   NP: Donovan: Celia Of The Seals


signature.asc
Description: Digital Signature


Bug#908323: libgtk3-perl: FTBFS: t/overrides.t failure

2018-09-12 Thread Iain Lane
Control: clone -1 -2
Control: reassign -2 gir1.2-gdkpixbuf-2.0
Control: found -2 2.38.0+dfsg-4
Control: retitle -2 gdk_pixbuf_new_from_inline introspection metadata is broken
Control: forwarded -2 https://gitlab.gnome.org/GNOME/gdk-pixbuf/issues/92
Control: tags -2 + upstream confirmed patch
Control: affects -1 =
Control: affects -2 src:libgtk3-perl

Hi there,

On Sat, Sep 08, 2018 at 03:43:21PM +0200, gregor herrmann wrote:
> Control: tag -1 + confirmed upstream
> 
> On Sat, 08 Sep 2018 15:03:13 +0300, Niko Tyni wrote:
> 
> > I'm guessing this regressed with gdk-pixbuf 2.38.0+dfsg-1 or so, but I
> > haven't verified that (and the ci.debian.net machinery doesn't seem to
> > have pinpointed it yet either.)
> 
> Sounds plausible; the changelog mentions
> "Generate separate introspection data for GdkPixdata API"
> and there's now a
> /usr/lib/x86_64-linux-gnu/girepository-1.0/GdkPixdata-2.0.typelib
> besides the old
> /usr/lib/x86_64-linux-gnu/girepository-1.0/GdkPixbuf-2.0.typelib

Right - this part is "expected". I should have tried build testing
libgtk3-perl with the new version, sorry about that.

> 
> I tried with
> 
> #v+
> --- a/lib/Gtk3.pm
> +++ b/lib/Gtk3.pm
> [...]
> #v-

This diff looks correct to me - I think it could be forwarded upstream
and/or uploaded.

> but then the tests failed differently, with:
> 
> Gtk3::Gdk::Pixbuf::new_from_inline: passed too few parameters (expected 4, 
> got 3) at t/overrides.t line 762.
> # Looks like your test exited with 255 just after 162.

...and this one is a real bug in gdk-pixbuf. I forwarded a potential fix
to upstream - will wait a little while to see what they say. I've
confirmed that after applying the patch already in this bug and that
proposed patch libgtk3-perl's autopkgtests pass again. Once it gets an
ack or if there is no reply then I'll upload to Debian.

Cheers,

-- 
Iain Lane  [ i...@orangesquash.org.uk ]
Debian Developer   [ la...@debian.org ]
Ubuntu Developer   [ la...@ubuntu.com ]


signature.asc
Description: PGP signature


Bug#908323: libgtk3-perl: FTBFS: t/overrides.t failure

2018-09-08 Thread gregor herrmann
Control: tag -1 + confirmed upstream

On Sat, 08 Sep 2018 15:03:13 +0300, Niko Tyni wrote:

> I'm guessing this regressed with gdk-pixbuf 2.38.0+dfsg-1 or so, but I
> haven't verified that (and the ci.debian.net machinery doesn't seem to
> have pinpointed it yet either.)

Sounds plausible; the changelog mentions
"Generate separate introspection data for GdkPixdata API"
and there's now a
/usr/lib/x86_64-linux-gnu/girepository-1.0/GdkPixdata-2.0.typelib
besides the old
/usr/lib/x86_64-linux-gnu/girepository-1.0/GdkPixbuf-2.0.typelib

I tried with

#v+
--- a/lib/Gtk3.pm
+++ b/lib/Gtk3.pm
@@ -60,6 +60,7 @@
   Gtk-3.0   | Gtk3
   Gdk-3.0   | Gtk3::Gdk
   GdkPixbuf-2.0 | Gtk3::Gdk
+  GdkPixdata-2.0| Gtk3::Gdk
   Pango-1.0 | Pango
 
 =cut
@@ -84,6 +85,10 @@
 my $_GDK_PIXBUF_VERSION = '2.0';
 my $_GDK_PIXBUF_PACKAGE = 'Gtk3::Gdk';
 
+my $_GDK_PIXDATA_BASENAME = 'GdkPixdata';
+my $_GDK_PIXDATA_VERSION = '2.0';
+my $_GDK_PIXDATA_PACKAGE = 'Gtk3::Gdk';
+
 my $_PANGO_BASENAME = 'Pango';
 my $_PANGO_VERSION = '1.0';
 my $_PANGO_PACKAGE = 'Pango';
@@ -472,6 +477,11 @@
 flatten_array_ref_return_for => 
\@_GDK_PIXBUF_FLATTEN_ARRAY_REF_RETURN_FOR);
 
   Glib::Object::Introspection->setup (
+basename => $_GDK_PIXDATA_BASENAME,
+version => $_GDK_PIXDATA_VERSION,
+package => $_GDK_PIXDATA_PACKAGE);
+
+  Glib::Object::Introspection->setup (
 basename => $_PANGO_BASENAME,
 version => $_PANGO_VERSION,
 package => $_PANGO_PACKAGE);
#v-

but then the tests failed differently, with:

Gtk3::Gdk::Pixbuf::new_from_inline: passed too few parameters (expected 4, got 
3) at t/overrides.t line 762.
# Looks like your test exited with 255 just after 162.


so this needs a bit more (and more background knowledge).


Cheers,
gregor


-- 
 .''`.  https://info.comodo.priv.at -- Debian Developer https://www.debian.org
 : :' : OpenPGP fingerprint D1E1 316E 93A7 60A8 104D  85FA BB3A 6801 8649 AA06
 `. `'  Member VIBE!AT & SPI Inc. -- Supporter Free Software Foundation Europe
   `-   NP: Michael Penn: Whole Truth


signature.asc
Description: Digital Signature


Bug#908323: libgtk3-perl: FTBFS: t/overrides.t failure

2018-09-08 Thread Niko Tyni
Source: libgtk3-perl
Version: 0.034-1
Severity: serious
User: debian-p...@lists.debian.org
Usertags: autopkgtest
Control: block 902557 with -1

As noticed by the autopkgtest checks on ci.debian.net, this package fails
its test suite on current sid, making it also fail to build from source.

  
https://ci.debian.net/data/autopkgtest/unstable/amd64/libg/libgtk3-perl/953602/log.gz

  Undefined subroutine &Gtk3::Gdk::PIXDATA_HEADER_LENGTH called at 
/<>/blib/lib/Gtk3.pm line 2119.
  # Looks like your test exited with 255 just after 162.

  [...]

  Test Summary Report
  ---
  t/overrides.t(Wstat: 65280 Tests: 162 Failed: 0)
Non-zero exit status: 255
Parse errors: Bad plan.  You planned 233 tests but ran 162.
  Files=22, Tests=843,  5 wallclock secs ( 0.18 usr  0.03 sys +  4.58 cusr  
0.46 csys =  5.25 CPU)
  Result: FAIL
 
I'm guessing this regressed with gdk-pixbuf 2.38.0+dfsg-1 or so, but I
haven't verified that (and the ci.debian.net machinery doesn't seem to
have pinpointed it yet either.)
-- 
Niko Tyni   nt...@debian.org