Bug#1072407: Fix is coming

[John Horigan]

I am fixing this FFmpeg API change in upstream. The new version builds
under unstable and experimental. I have uploaded an updated package to
mentors.debian.net. You should see it soon.

-- john

Bug#1067603: contextfree: FTBFS on armhf: test failure with missing codec

[John Horigan]

Could this bug be the cause?

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1012791

If libavcodec cannot access libx264.so due to some executable stack
security issue, then it would fail to load the libx264 encoder.

-- john

Bug#1067603: contextfree: FTBFS on armhf: test failure with missing codec

[John Horigan]

That error message indicates that libavcodec60 does not support the libx264
codec for encoding H.264 files. libavcodec60 lists libx264-164 as a
dependency, with no exception for armel or armhf systems, so I don't know
why the codec won't load.

Can you run the command

ffmpeg -codecs

on an armhf system and report back the output?

-- john

On Sun, 24 Mar 2024 21:04:49 +0900 Kentaro HAYASHI  wrote:
> Package: contextfree
> Version: 3.4+dfsg-1.1
> Severity: important
> Tags: ftbfs
> X-Debbugs-Cc: ken...@xdump.org
>
> Dear Maintainer,
>
>
>* What led up to the situation?
>
>contextfree can't build on armel,armhf.
>
>* What exactly did you do (or not do) that was effective (or
>  ineffective)?
>
> apt-get source contextfree
> cd contextfree-3.4+dfsg
> debuild -us -uc
>
>* What was the outcome of this action?
>
> See
>
https://buildd.debian.org/status/fetch.php?pkg=contextfree=armhf=3.4%2Bdfsg-1.1=1711207519=0
>
> input/ziggy_v2.cfdg   pass Reading rules file input/mtree.cfdg
> Restarting as a version 3 design
> 8 rules loaded
> Generating 8bit gray-scale Quicktime movie, variation FFGH...
> Failed to create movie file: codec not found
> make[1]: *** [Makefile:188: test] Error 8
> make[1]: Leaving directory '/home/kenhys/work/contextfree-3.4+dfsg'
> dh_auto_test: error: make -j8 test returned exit code 2
> make: *** [debian/rules:6: build] Error 25
> dpkg-buildpackage: error: debian/rules build subprocess returned exit
> status 2 debuild: fatal error at line 1184:
> dpkg-buildpackage -us -uc -ui failed
>
>* What outcome did you expect instead?
>
> contextfree can build on armel/armhf.
>
> -- System Information:
> Debian Release: trixie/sid
>   APT prefers unstable-debug
>   APT policy: (500, 'unstable-debug'), (500, 'unstable')
> Architecture: armhf (armv8l)
>
> Kernel: Linux 6.1.0-18-arm64 (SMP w/8 CPU threads)
> Locale: LANG=C, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set
> Shell: /bin/sh linked to /usr/bin/dash
> Init: unable to detect
>
> Versions of packages contextfree depends on:
> pn  libagg2
> ii  libavcodec60   7:6.1.1-3
> ii  libavformat60  7:6.1.1-3
> ii  libavutil587:6.1.1-3
> ii  libc6  2.37-15.1
> ii  libgcc-s1  14-20240315-1

Bug#1038018: agg: Depends on SDL 1.2

[John Horigan]

libagg/libagg2 come with several example programs that exercise the
library. After the library is built, these programs are built to ensure
that the headers are correct and the library exports the correct symbols.
The resulting binaries are discarded. If libsdl1.2-compat-dev works for
this test built step then I would prefer to use this instead of porting the
examples to SDL2.

I propose to make a new release of libagg/libagg2 using the latest upstream
rev, close this bug using libsdl1.2-compat-dev, and closing a minor bug.

-- john

On Sun, 25 Jun 2023 14:33:38 +0100 Simon McVittie  wrote:
> On Thu, 15 Jun 2023 at 12:31:50 +0100, Simon McVittie wrote:
> > If possible, please port this package to SDL 2 and close this bug.
>
> I couldn't find any sign that agg's reverse dependencies (contextfree,
> desmume, exactimage, svgpp) actually need the SDL 1.2 platform plugin,
> so another possible way to close this bug would be to disable the SDL
> 1.2 plugin. If you do this, please coordinate with the maintainers of
> the reverse dependencies to make sure this won't break them.
>
> > 4. Install libsdl1.2-compat-dev and recompile the package.
>
> I tried this on a porterbox and it seems to build fine. I didn't test
> the resulting binaries.
>
> smcv
>
>

Bug#1004803: contextfree: diff for NMU version 3.3+dfsg1-1.1

[John Horigan]

Sebastian,

I see no reason to delay your NMU patch. I apologize for not handling this
bug in a timely manner. My Debian development system died and I am still a
few weeks away from getting a replacement. When I have a working Debian
system again I will confirm functionality with FFmpeg5.0 and migrate your
patch to upstream.

-- john

Bug#951939: contextfree: FTBFS: src-common/cfdg.cpp:75:71: error: no matching function for call to ‘find(std::array, 18>::const_iterator, std::array

[John Horigan]

This bug is due to several source files using STL algorithms without
having the required #include . This will be fixed in
upstream and a new feature release of Context Free will be uploaded to
Debian soon.

Bug#940662: contextfree FTCBFS: strips with the wrong strip

[John Horigan]

Helmut,

A new version of contextfree with your patch is awaiting the package
sponsor.
It can be viewed at https://mentors.debian.net/package/contextfree
This package also incorporates a new upstream version and updates
debhelper compatibility to v12.

-- john

Bug#921111: agg: symbols adjustments to support build with -O3

[John Horigan]

There are several more template symbols besides those two. Shouldn't they
all be marked as optional? This is my first time generating a package with
a symbols file.

-- john

On Fri, Feb 1, 2019 at 9:03 AM Steve Langasek 
wrote:

> Package: agg
> Version: 1:2.6.0-r132+dfsg1-2
> Severity: minor
> Tags: patch
> User: ubuntu-de...@lists.ubuntu.com
> Usertags: origin-ubuntu disco ubuntu-patch
>
> Hi John,
>
> The latest version of agg, 1:2.6.0-r132+dfsg1-2, fails to build from source
> on ppc64el in Ubuntu due to a difference in the symbols output for the
> shared library.  This is because Ubuntu defaults to -O3 on ppc64el, which
> means some additional template symbols are omitted from the output that are
> otherwise seen when building with -O2.
>
> Since these are template symbols and not part of the shared library ABI, a
> correct fix to make the library compatible with -O3 is to mark these
> symbols
> optional as in the attached patch.  Please consider applying in Debian.
>
> Thanks,
> --
> Steve Langasek   Give me a lever long enough and a Free OS
> Debian Developer   to set it on, and I can move the world.
> Ubuntu Developer   https://www.debian.org/
> slanga...@ubuntu.com vor...@debian.org
>

Bug#919322: CVE-2019-6245 CVE-2019-6247

[John Horigan]

I have prepared a new version of libagg-dev that fixes #377270 and I was
about to ask the package sponsor to upload it when this issue surfaced. I
will fold in fixes for the new bugs too.

-- john

On Mon, 14 Jan 2019 23:27:23 +0100 Moritz Muehlenhoff 
wrote:
> Source: agg
> Severity: grave
>
> Please see
> http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2019-6245
> http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2019-6247
>
> Given that the package isn't exactly fast-moving, how about
> revisiting #377270 for buster? Right now we need to coordinate
> rebuilds against the fixed agg...
>
> Cheers,
> Moritz
>
>

Bug#914740: agg FTCBFS: configures for the build architecture

[John Horigan]

I am preparing an update to libagg-dev with your patch and an upstream fix.
I am getting a lintian
warning: pkg-config-unavailable-for-cross-compilation because the pkgconfig
file is being written to /usr/lib/pkgconfig instead
of /usr/share/pkgconfig.  Do I need to fix this in order to support
cross-compilation?

-- john

Bug#853424: gnash: ftbfs with GCC-7

[John Horigan]

It is unfortunate that the author of revision 63 chose to remove rgba8_pre,
rather than create linear and sRGB variants.

If you replace agg::rgba8_pre(r,g,b,a) with
agg::rgba8(r,g,b,a).premultiply() it should work for both the old and new
versions of libagg

-- john

On Sat, Dec 16, 2017 at 12:11 AM Juhani Numminen 
wrote:

> Control: retitle 853424 gnash: FTBFS with agg 1:2.4-r127+dfsg1-1
>
> Hi,
>
> This is the first warning in my build log (attached):
>
> make[4]: Entering directory
> '/build/gnash-0.8.11~git20160608/tmp.build/librender'
>   CXX  libgnashrender_la-Renderer_agg.lo
> In file included from ../../librender/agg/Renderer_agg.cpp:146:0:
> ../../librender/agg/Renderer_agg_style.h: In member function 'void
> gnash::AddStyles::operator()(const gnash::SolidFill&) const':
> ../../librender/agg/Renderer_agg_style.h:617:28: error: 'rgba8_pre' is not
> a member of 'agg'
>  _sh.add_color(agg::rgba8_pre(color.m_r, color.m_g, color.m_b,
> ^
> ../../librender/agg/Renderer_agg_style.h:617:28: note: suggested
> alternative: 'rgba_pre'
>  _sh.add_color(agg::rgba8_pre(color.m_r, color.m_g, color.m_b,
> ^
>
> It seems that rgba8_pre really was removed from the headers in the last
> upload of agg.
>
> https://sources.debian.org/src/agg/2.5+dfsg1-11/include/agg_color_rgba.h/#L436
>
> https://sources.debian.org/src/agg/1:2.4-r127+dfsg1-1/include/agg_color_rgba.h
>
> In the upstream repository, the removal was done in revision 63.
> https://sourceforge.net/p/agg/svn/63/?page=2#diff-17
>
> https://sourceforge.net/p/agg/svn/63/tree/agg-2.4/include/agg_color_rgba.h?diff=518286af2718467b8b34c637:62
>
>
> Best regards,
> Juhani
>

Bug#872738: contextfree FTBFS on ppc64el: error: 'CTL_HW' was not declared in this scope

[John Horigan]

I have applied Frederic'c patch to the upstream source as well as fixes for
the FTBFS issues with other platforms (ppc, alpha, hurd-i386, x32). I also
eliminated the use of the deprecated readdir_r() function. Context Free
3.0.11.5 package is being reviewed by the package sponsor.

-- john

On Sun, Aug 20, 2017 at 10:09 AM Adrian Bunk  wrote:

> Source: contextfree
> Version: 3.0.11.4+dfsg1-1
> Severity: serious
>
>
> https://buildd.debian.org/status/fetch.php?pkg=contextfree=ppc64el=3.0.11.4%2Bdfsg1-1=1503230603=0
>
> ...
> g++ -g -O2 -fdebug-prefix-map=/<>/contextfree-3.0.11.4+dfsg1=.
> -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time
> -D_FORTIFY_SOURCE=2 -Isrc-common -Isrc-unix -Iobjs -Isrc-common/agg-extras
> -Isrc-ffmpeg/include -I/usr/local/include -O2 -Wall -Wextra
> -Wno-parentheses -std=c++14  -c -o objs/posixSystem.o
> src-unix/posixSystem.cpp
> src-unix/posixSystem.cpp: In member function 'virtual
> std::vector
> PosixSystem::findTempFiles()':
> src-unix/posixSystem.cpp:220:53: warning: 'int readdir_r(DIR*, dirent*,
> dirent**)' is deprecated [-Wdeprecated-declarations]
>  while (readdir_r(dirp.get(), , ) == 0 && der) {
>  ^
> In file included from src-unix/posixSystem.cpp:36:0:
> /usr/include/dirent.h:183:12: note: declared here
>  extern int readdir_r (DIR *__restrict __dirp,
> ^
> src-unix/posixSystem.cpp: In member function 'virtual size_t
> PosixSystem::getPhysicalMemory()':
> src-unix/posixSystem.cpp:246:14: error: 'CTL_HW' was not declared in this
> scope
>  mib[0] = CTL_HW;
>   ^~
> src-unix/posixSystem.cpp:246:14: note: suggested alternative: 'CTL_PM'
>  mib[0] = CTL_HW;
>   ^~
>   CTL_PM
> Makefile:206: recipe for target 'objs/posixSystem.o' failed
> make[1]: *** [objs/posixSystem.o] Error 1
>

Bug#522360: econvert: Incorrect output from SVG to PNG

[John Horigan]

reassign 522360 agg 2.5+dfsg1-12
thanks

This bug is not in the libagg-dev binary package. It is in the agg source
package that libagg-dev is built from. Specifically, the source files in
agg-2.5+dfsg1/examples/svg_test that are incorporated in exact-image. This
bug will be marked as fixed when agg 1:2.4-r127+dfsg1 enters the debian
archive.

The exact-image maintainer will have to copy source files from
agg-2.4-r127+dfsg1/examples/svg_test into the exact-image codecs directory,
making API changes as needed. Upstream changes made to the agg_svg files
have been migrated to upstream agg so this should be easy.

-- john

Bug#827993: I Intend to Adopt agg

[John Horigan]

retitle 827993 ITA: agg -- AntiGrain Geometry graphical toolkit
owner 827993 !
thanks

I am an upstream maintainer of AGG. I am working with the sponsor of the
previous maintainer to bring package agg up to the latest upstream revision
and fix all outstanding bugs.

-- john