Re: Maintainers, please check if you actually depend on intltool

[Milan Crha <mcrha redhat com>]

On Mon, 2018-12-03 at 18:15 +0000, Javier Jardón wrote:
> As you probably know, for some years we have been trying to move to
> upstream gettext [1]

        Hi,
I know intltool has some issues, the projects I work on faced some of
them, but it also provides very useful tools and makes life easier to
the developers, thus I rather do not understand the need of the move to
plain gettext. Consider also code (or rather script) duplication
(scripts to extract translatable strings from .desktop and various .xml
files). I do not have those scripts, neither I'm sure where to get them
or how to create them on my own.

Changes like [4] are kind of funny, especially those where
   # TRANSLATORS: Don't translate this text (this is icon name)
had been added. It was not needed before. Thus intltool can help to the
translators as well (and the .po files did have less bogus strings).

The [1] also doesn't mention a replacement to 'intltool-update -m',
which I use quite often.

> open MR's to reflect the real dependencies of your module, that would
> be great

I'm sorry, I never understood such requests, but it can be just me.
Consider simple "dependencies changed: dropped dependency on aaaa,
added dependency on bbbb" comment, which someone knowledgeable of the
internal things would use to make things right on the first shot, with
something like: "clone some repo, learn where things are stored, what
format is used, ideally also what implications some changes have, learn
how to test whether the change won't break anything obvious (like
compile it somehow locally at least), then open a pull request and add
it somehow in a fork of the project probably...". That's much more
complicated and discouraging for someone whom has no single idea of the
"target" project.

I know, you said "would be great", I understand it, but I also see an
increasing demand of creating pull requests while one suggests a one-
liner change easily achievable by the maintainer in a fraction of time
with compare of all the tasks a person not working with the project at
all would do. I'm not talking about real "code" changes here, there's a
difference when someone wants to suggest a change into the project on
his/her own, which he/she also wants to test in action.

Just my opinion.
        Bye,
        Milan

[4] https://gitlab.gnome.org/GNOME/gnome-menus/merge_requests/1.patch