bug#71213: [PATCH] Use the current face foreground for read-passwd icons

[Michael Albinus via "Bug reports for GNU Emacs, the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>]

Kévin Le Gouguec <kevin.legouguec@gmail.com> writes:

Hi Kévin,

>>> We've done something similar for other icons, e.g. in 2021-11-02 "Use
>>> current face foreground for SVG icons in customize" (11702a6dd7c).  I
>>> think it makes sense for these too, since "#2e3436" (and any shade of
>>> not-strictly-gray-but-close-enough) is liable to have low contrast with
>>> the mode line background depending on the theme (e.g. modus-vivendi).
>>>
>>> Let me know what you think.
>>
>> I have absolutely no knowledge about image formats. I've stolen the two
>> files from the Gnome repository, that's it. So I won't object any
>> change.
>
> Right, AFAIU we followed the same approach for the icons I mention above
> (grabbed the icons somewhere, realized their foreground fared poorly
> depending on the background, stripped their 'fill' attribute).

So I would say it is OK to install your patch. Other people will comment
only if they see an unexpected change, so this is likely the best way to
test your proposal :-)

>> While we're at it: do the two correspinding *.pbm files also need some
>> care?
>
> It's not a format I am familiar with, but empirically,
>
>   (insert-image (create-image "conceal.pbm"))
>
> inserts an icon that automagically changes color as I switch between
> light and dark themes.  pbm(5) says the format is monochrome, and (info
> "(elisp) Other Image Types") confirms Emacs does the obvious thing by
> default (frame foreground for one bits; frame background for zero bits).
>
>>       And do we need to create also *.xbm files?
>
> Not sure if we have a policy or guidelines there; etc/images/README
> explains how to convert from xpm > xbm > pbm, but does not provide
> judgments on which are needed.  (info "(elisp) Image Formats") says
>
>        Supported image formats (and the required support libraries) include
>     PBM and XBM (which do not depend on support libraries and are always
>     available),
>
> which makes me tempted to say "if we have lowest-common-denominator
> formats like PBM or XBM for our icons, then we're good, the rest is
> bonus"?

OK.

Best regards, Michael.