Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!

Bug 722448

Summary: media-libs/gd-2.3.0 - fails gdimagestring16/gdimagestring16 and gdimagestringup16/gdimagestringup16 tests
Product: Gentoo Linux Reporter: ernsteiswuerfel <erhard_f>
Component: Current packagesAssignee: Codec Project <codec>
Status: RESOLVED FIXED    
Severity: normal CC: sam
Priority: Normal Keywords: PullRequest, TESTFAILURE
Version: unspecified   
Hardware: All   
OS: Linux   
See Also: https://github.com/gentoo/gentoo/pull/15754
https://bugs.gentoo.org/show_bug.cgi?id=608730
Whiteboard:
Package list:
Runtime testing required: ---
Attachments: build.log (ppc)
emerge --info
test-suite.log (ppc)

Description ernsteiswuerfel archtester 2020-05-11 11:25:21 UTC
Created attachment 638156 [details]
build.log (ppc)

[...]
FAIL: gdimagestring16/gdimagestring16
FAIL: gdimagestringup16/gdimagestringup16
[...]
============================================================================
Testsuite summary for GD 2.3.0
============================================================================
# TOTAL: 202
# PASS:  200
# SKIP:  0
# XFAIL: 0
# FAIL:  2
# XPASS: 0
# ERROR: 0
Comment 1 ernsteiswuerfel archtester 2020-05-11 11:26:00 UTC
Created attachment 638158 [details]
emerge --info
Comment 2 ernsteiswuerfel archtester 2020-05-11 11:28:35 UTC
Created attachment 638160 [details]
test-suite.log (ppc)
Comment 3 Sam James archtester Gentoo Infrastructure gentoo-dev Security 2020-05-11 16:36:45 UTC
Thakns for checking this so quickly. Could you report this upstream?
Comment 4 ernsteiswuerfel archtester 2020-05-11 20:42:12 UTC
Hmm, interesting... Before reporting upstream I built 2.3.0 and git-master from upstream sources.


Upstream 2.3.0 fails 4 tests:

FAIL: gdimagecopyresampled/bug00201
===================================

gdimagecopyresampled/bug00201.c:68: Total pixels changed: 190 with a maximum channel difference of 255.
FAIL gdimagecopyresampled/bug00201 (exit status: 1)

FAIL: gdimagegrayscale/basic
============================

gdimagegrayscale/basic.c:21: Total pixels changed: 16800 with a maximum channel difference of 131.
FAIL gdimagegrayscale/basic (exit status: 1)

FAIL: gdimagestring16/gdimagestring16
=====================================

gdimagestring16/gdimagestring16.c:25: Total pixels changed: 24 with a maximum channel difference of 255.
FAIL gdimagestring16/gdimagestring16 (exit status: 1)

FAIL: gdimagestringup16/gdimagestringup16
=========================================

gdimagestringup16/gdimagestringup16.c:25: Total pixels changed: 24 with a maximum channel difference of 255.
FAIL gdimagestringup16/gdimagestringup16 (exit status: 1)


Upstream git-master fails only 2 tests:

FAIL: gdimagecopyresampled/bug00201
===================================

gdimagecopyresampled/bug00201.c:68: Total pixels changed: 190 with a maximum channel difference of 255.
FAIL gdimagecopyresampled/bug00201 (exit status: 1)

FAIL: gdimagegrayscale/basic
============================

gdimagegrayscale/basic.c:21: Total pixels changed: 16800 with a maximum channel difference of 131.
FAIL gdimagegrayscale/basic (exit status: 1)


So upstream git-master fails 2 tests Gentoo's 2.3.0 does not fail. I am a bit unsure about what to report upstream then?
Comment 5 Sam James archtester Gentoo Infrastructure gentoo-dev Security 2020-05-11 20:45:55 UTC
I've excluded the two "old tests"

FAIL: gdimagecopyresampled/bug00201
===================================

gdimagecopyresampled/bug00201.c:68: Total pixels changed: 190 with a maximum channel difference of 255.
FAIL gdimagecopyresampled/bug00201 (exit status: 1)

FAIL: gdimagegrayscale/basic
============================

which failed in the past, because they failed on my machine too, and were not regressions.

I meant to report them upstream but have not got to it yet. So, when you report, you could note that we've seen those two on ppc, amd64, all kinds of things, but those two new ones (gdimagestring16, gdimagestringup16) are new & observed only on ppc.

Sorry for the confusion.
Comment 6 ernsteiswuerfel archtester 2020-05-11 22:02:50 UTC
(In reply to Sam James (sec padawan) from comment #5)
> I meant to report them upstream but have not got to it yet. So, when you
> report, you could note that we've seen those two on ppc, amd64, all kinds of
> things, but those two new ones (gdimagestring16, gdimagestringup16) are new
> & observed only on ppc.

Had a closer look at upstream now. gdimagestring16, gdimagestringup16 already seem fixed by: https://github.com/libgd/libgd/commit/ec0b4397698bd358557d33641aac4beb1912f43b

gdimagecopyresampled/bug00201 and gdimagegrayscale/basic test failures have already been brought up in https://github.com/libgd/libgd/issues/480 and https://github.com/libgd/libgd/issues/594

I added my data to upstream issue 594 as it is the most recent. I was not able to reproduce gdimagecopyresampled/bug00201 and gdimagegrayscale/basic failures on git-master on my amd64 box so I suspect them to be 32bit specific.
Comment 7 Sam James archtester Gentoo Infrastructure gentoo-dev Security 2020-05-11 22:09:17 UTC
(In reply to ernsteiswuerfel from comment #6)
> (In reply to Sam James (sec padawan) from comment #5)
> > I meant to report them upstream but have not got to it yet. So, when you
> > report, you could note that we've seen those two on ppc, amd64, all kinds of
> > things, but those two new ones (gdimagestring16, gdimagestringup16) are new
> > & observed only on ppc.
> 
> Had a closer look at upstream now. gdimagestring16, gdimagestringup16
> already seem fixed by:
> https://github.com/libgd/libgd/commit/
> ec0b4397698bd358557d33641aac4beb1912f43b
> 
> gdimagecopyresampled/bug00201 and gdimagegrayscale/basic test failures have
> already been brought up in https://github.com/libgd/libgd/issues/480 and
> https://github.com/libgd/libgd/issues/594
> 
> I added my data to upstream issue 594 as it is the most recent. I was not
> able to reproduce gdimagecopyresampled/bug00201 and gdimagegrayscale/basic
> failures on git-master on my amd64 box so I suspect them to be 32bit
> specific.

Oh, perfect. I'd actually looked at the fixes in master since the release and I guess I saw that
but given we weren't hitting the failures at that time, ignored it...

I've subscribed to those threads but we've got the tests disabled for now.

I really appreciate your investigation. Interestingly, like I said, I actually hit
those two original failures on master still (if I recall correctly) on amd64.

In any case, it's not Gentoo specific and it seems to be a problem with the tests themselves, so I'm happy.

I will apply that patch and hopefully we're good!
Comment 8 Larry the Git Cow gentoo-dev 2020-05-12 22:30:22 UTC
The bug has been closed via the following commit(s):

https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=394bb39d8e64a4b5e16c32c5ce03aa28d6f713e2

commit 394bb39d8e64a4b5e16c32c5ce03aa28d6f713e2
Author:     Sam James (sam_c) <sam@cmpct.info>
AuthorDate: 2020-05-11 23:06:20 +0000
Commit:     Thomas Deutschmann <whissi@gentoo.org>
CommitDate: 2020-05-12 22:29:59 +0000

    media-libs/gd: Add upstream patch for 2 tests
    
    Fixes gdimagestring16 and gdimagestringup16 tests.
    
    Thanks-to: ernsteiswuerfel <erhard_f@mailbox.org>
    Closes: https://bugs.gentoo.org/722448
    Package-Manager: Portage-2.3.99, Repoman-2.3.22
    Signed-off-by: Sam James (sam_c) <sam@cmpct.info>
    Closes: https://github.com/gentoo/gentoo/pull/15754
    Signed-off-by: Thomas Deutschmann <whissi@gentoo.org>

 .../gd/files/gd-2.3.0-fix-tests-bug722448.patch    | 62 ++++++++++++++++++++++
 media-libs/gd/gd-2.3.0.ebuild                      |  1 +
 2 files changed, 63 insertions(+)