GNU bug report logs - #36403
27.0.50; Trivial image.c bugs

Previous Next

Full log

Message #63 received at 36403 <at> debbugs.gnu.org (full text, mbox):

From: "Basil L. Contovounesios" <contovob <at> tcd.ie>
To: Lars Ingebrigtsen <larsi <at> gnus.org>
Cc: Alan Third <alan <at> idiocy.org>, 36403 <at> debbugs.gnu.org,
 Pip Cet <pipcet <at> gmail.com>
Subject: Re: bug#36403: 27.0.50; Trivial image.c bugs
Date: Tue, 04 Oct 2022 16:52:59 +0300

[Message part 1 (text/plain, inline)]

Lars Ingebrigtsen [2020-08-21 13:26 +0200] wrote:

> Pip Cet <pipcet <at> gmail.com> writes:
>
>> Paul's suggestion was to use equal () instead of !NILP (Fequal (...)).
>> I'm against that, because the F in Fequal kind of hints at the
>> difficulties of using equal, of which there are many: in the current
>> implementation, it can signal, quit, be asymmetric (signalling for
>> (equal a b) whereas (equal b a) works), and is susceptible to equality
>> bombs that take forever to compare.
>
> Yeah, your equal_lists is better in all ways, I think.  It should be
> much faster, too -- Fequal on a list checks whether the string members
> are equal, too, which is slow.  So I think this will speed things up if
> you have a buffer that displays images where the data comes from a
> string (which can be huge) instead of a file.

We now have Fequal again:

  Restore Emacs 27 image cache semantics
  ac341cd629 2020-12-09 00:42:11 +0100
  https://git.sv.gnu.org/cgit/emacs.git/commit/?id=ac341cd629

Which means image-test-circular-specs signals:

  (circular-list (:dummy . #0))

So how important is it to support image specs with circular property
values?  Should the test be marked :expected-result :failed?

Either way, WDYT of the attached minor cleanup?

Thanks,

-- 
Basil

[0001-Touch-up-image-circular-tests.el.patch (text/x-diff, attachment)]

Previous Next

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