[VOTE] fund work on Cambalache macOS port

Hello,

based on the previous discussion, I propose the following VOTE:

  1. TDF to fund work on the Cambalache macOS port
  2. Budget 10.000 € for this
  3. Authorize Florian, in case this vote gets approved, to sign the contracts on behalf of TDF

The vote runs 72h from now. It ends earlier if all board members have casted their vote.

Florian

+1

+1

@floeff , this proposal ignores the previous discussion - which we were just in the middle of, with some alternatives, possibly better ones, being thought up and considered.

Michael Weghorn already responded to your questions in the other thread.

To summarize: Your suggestion to use some other tool/process to convert the files back and forth to Qt Designer XML doesn’t work because that file format is not feature-compatible with the GTK UI designer XML format we currently use. In the long term we may consider migrating to a different UI file format, but any effort in this area will take a lot of time to complete, and it is important that we maintain a working toolset for our contributors while we are working on whatever else comes next.

The GLADE tool we were previously using has been discontinued, is not supported by developers, and is difficult to install and run on modern systems. Cambalache is a drop-in alternative, but currently has the drawback of not being usable by contributors on all platforms. This proposal will address that drawback.

First, there are objections and alternative suggestions, not questions.

Second, you wrote

but that is not true. People would still be without a relevant program on most computers - specifically, on Linux with X11, and on Windows (and on *BSD). So, the Cambaleche proposal also “doesn’t work”.

Third, about the format difference: Can it not be bridged through in and out conversion? And if not - why not, instead, extend the Qt UI designer to support another, or a richer, format? Instead of shoehorning Wayland onto Macs, we could just broaden the capabilities of an already-multiplatform tool.

This makes a conversion-filter-based solution even more attractive.

Cambalache runs fine in an X11 session on Linux. (The fact that it uses Wayland technologies internally to embed the preview into the main window doesn’t mean that the host system needs to be running a Wayland session as well.)
For Windows, running it in WSL might possibly be an option (but I haven’t tried myself). (On Windows, we already recommend using WSL for the LibreOffice build itself as well.)

I’m not opposing this but since the discussion is still in progress I am abstaining until I am confident that the matter is adequately discussed beyond the staff.

S.

Hi Florian, all

My vote is in favor.

Best,
Eliane Domingos

+1 for me on all points.

Ciao

Paolo

+1

–Osvaldo

how many ? :thinking:

…
4. publish the contract as signed, or at least the key details (deliverable(s), deadline, …)
5. properly track the tasks in a public bugzilla or redmine or else.

I would assume it’s one, but we need to get (and evaluate) it first.

@ilmari gaves and gives regular status updates on the Bugzilla contract, and I would assume the same will be done here.

Indeed, both of fpy’s points apply to the Bugzilla project (key details instead of full contract).