bug-mumi
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#70871: [PATCH] Show avatars from Libravatar.


From: Felix Lechner
Subject: bug#70871: [PATCH] Show avatars from Libravatar.
Date: Mon, 13 May 2024 12:30:02 -0700

Hi Arun,

On Mon, May 13 2024, Arun Isaac wrote:

> would you mind deduplicating these three avatars into a single
> function or similar?

Personally, I do not see the benefit of that abstraction for HTML.  For
one, the code is a lot easier to parse next to a Web Inspector that
shows the same tags.

There is also not much to abstract:  It's a single 'img' tag.

Moreover, the avatars are also in different positions.  They do
different things on the page.  Most significantly, they are subject to
different CSS.

You'll see why that matters just below.

> There seems to be one more place in mumi/web/view/html.scm where there
> is an avatar.

There is no third avatar.  If you haven't yet followed the CSS with a
Web Inspector, please look for the 'display: none' for that 'img' tag.
It comes from here:

    https://git.savannah.gnu.org/cgit/guix/mumi.git/tree/assets/mumi.scss#n394

Does the "third" avatar more or less duplicate with the big one in
front?  Do you plan to use the third avatar?  I'd drop it.

> I also just remembered that the alt attribute of the img tag is meant to
> be used in place of the image (visually challenged readers using a
> screen reader, image download failed on a slow connection, etc.). So, we
> shouldn't use the alt attribute as a fallback for users without
> avatars.

Would you please explain that rationale?  A blind person who disables
the loading of images would still "see" (or perhaps hear) the 'alt'
character.  How does my code break anything for a blind person, please?

> Also, we now need to check for guile-avatar in configure.ac.

It is not necessary for my deployment on patchwise.org, which already
shows the avatars.

For an interpreted language like Guile there is no need to provide
prerequisite checks at configure time unless they enable workarounds to
build without.  The Guile module detections currently found in
configure.ac are luxuries.  I would remove them.

If you prefer to keep them and add guile-avatar for the sake of
consistency, please feel free to amend the patch. I do not insist on Git
authorship for my contributions.  I am sorry, but I am super busy.

> what is patchwise.org? Did I miss some conversation about it?

That's the domain I hope to use in the near future---and in cooperation
with the FSF---to show off my plans for a Debbugs upgrade based on Mumi.

FSF declined to upgrade debbugs.gnu.org to the latest Debbugs version
(from Debian) because my packaging was based on Guix. FSF wants Trisquel
plus Ansible.

FSF encouraged me to complete my work elsewhere and demonstrate it when
done.  It was not my first choice, please believe me.  I protested
loudly in various places.  Either way, Patchwise is licensed under GNU
Affero.  There are no hidden goals.

> I was under the impression that we should be sending to
> bug-mumi@gnu.org.

I had hoped to open the bug by bouncing the message from patchwise.org
to gnu.org.  The Patchwise bouncer works in all other cases, such as
when amending or controlling bugs (please try it, or look at the headers
for this message!) or access to the mailing lists.  I'll have to talk to
FSF about using the envelope address for submissions, as well.

The Patchwise bouncer is essential for the automatic sequencer that will
allow folks to submit a patch series in one go.  I think your command
line tool may do something similar.

Also, the patchwise.org effort has not been made public.  You are one of
the first people to hear about it.

I'd be happy to cooperate, if you agree to be less picky about my
patches.

Kind regards
Felix





reply via email to

[Prev in Thread] Current Thread [Next in Thread]