GNU bug report logs - #70093
Remove unnecessary `image-tests--map-equal'

Previous Next

Full log

Message #62 received at 70093 <at> debbugs.gnu.org (full text, mbox):

From: Joseph Turner <joseph <at> breatheoutbreathe.in>
To: David Ponce <da_vid <at> orange.fr>
Cc: 70093 <at> debbugs.gnu.org, Eli Zaretskii <eliz <at> gnu.org>,
 Stefan Kangas <stefankangas <at> gmail.com>
Subject: Re: bug#70093: Remove unnecessary `image-tests--map-equal'
Date: Mon, 03 Mar 2025 09:09:50 -0800

David Ponce <da_vid <at> orange.fr> writes:

> On 2025-03-02 20:37, Joseph Turner wrote:
>> David Ponce <da_vid <at> orange.fr> writes:
>>
>>> On 2025-03-02 05:52, Joseph Turner wrote:
>>>> Stefan Kangas <stefankangas <at> gmail.com> writes:
>>>>
>>>>> Eli Zaretskii <eliz <at> gnu.org> writes:
>>>>>
>>>>>>> Was this patch installed?
>>>>>>
>>>>>> No.
>>>>>>
>>>>>>> Should it be?
>>>>>>
>>>>>> I'm not sure, and I said so in the discussion.
>>>>>
>>>>> Now I see that there were some issues with the patch reported by David
>>>>> Ponce, but I see no followup to his last message from Joseph.
>>>>>
>>>>> Joseph, do you still consider this patch a good idea?
>>>> Yes.  The stricter tests which I propose in the patch all pass on my
>>>> machine.  David, are you able to reproduce this issue with emacs -Q?
>>>> Thank you!
>>>> Joseph
>>>
>>> Here here are the results of running image map tests with emacs -Q
>> [...]
>> Thank you!  In this case, I suppose we should not apply this patch
>> until
>> we figure out why the generated image map varies on some machines.
>> Joseph
>
> Hello,
>
> I am still trying to understand why the the rotated-map and
> flipped-map cases in `image--compute-map-and-original-map' test fails
> for me.
>
> I wrote the function `svg-draw-image-map' (see below) to draw in SVG
> an image map, in order to help visualize image maps (maybe a such
> function could be useful to "kind of debug" image maps?).
>
> I also adapted the code from `image--compute-map-and-original-map'
> to be evaluated in the *scratch* buffer, in order to show computed
> and reference image maps.
>
> And, for me this reveals that the reference maps `flipped-map' and
> `rotated-map' are not as expected, contrary to the corresponding
> result of `image--compute-map' which looks as expected.  This explains
> why these test case are failing for me.  But I still can't explain
> why the same test cases passes for you?
>
> Maybe could you try the code below to see if it help clarify this?

[...]

The original-map generation logic is only on Emacs 30, not Emacs 29, so
I'll test this after upgrading to Emacs 30 (on the main Guix channel).

Thank you!

Joseph

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.