[Development] Trouble getting reviews in months and years

[Tor Arne Vestbø]

Hi Ilya,

You’re completely right that simplifying the process of contributing to Qt goes beyond technical roadblocks such as cloning repositories, building, and Gerrit setup. Once those issues are solved, it comes down to making a convincing and thorough case for why a change is both correct, and aligns with the exiting architecture and goals of Qt.

A challenge there is that what may be an obvious change for one developer may not be obvious for another. We see this all the time when working on Qt, not just for external contributions. You may have been working on a change for a while, researching various options, trying out various things, and landed on what to you I the natural conclusion of that work. But without the context and insight into that research it’s hard for other reviewers to asses whether that conclusion is correct.

One thing we are actively trying to do to mitigate this is to document the research that went into a change, and be explicit about the use-cases that we’re trying to solve. This may take the form of a JIRA task for larger changes, or just a well written and detailed commit message for smaller changes.

In your case I notice that many of your patches relate to DPI handling on X11. To follow the process outlined above I think a JIRA task where you describe in detail the use-cases you are trying to solve would help a lot. With e.g. screenshots, and examples of the effects and pros and cons of the solution. Think of it as explaining the use-case and solution to someone not familiar with high-DPI and/or X11 :)

What do you think?

Cheers,
Tor Arne


> On Oct 6, 2024, at 22:41, Ilya Fedin <fedin-ilja2010 at ya.ru> wrote:
> 
> If anyone has any interest on how
> https://lists.qt-project.org/pipermail/development/2024-April/045167.html
> did end, here's the summary:
> 
>> https://codereview.qt-project.org/c/qt/qtbase/+/343628
> 
> Maybe there was any misundestanding during the review but it has felt
> like reviewers are gaslighting me. Like, I checked the patch on a wide
> range of DEs, I was trying to explain that the UX is better when
> reading X resources on KDE but reviewers were not hearing that and just
> repeating the same thing in a circle that "KDE uses fontconfig". While
> the reality is that it supports both.
> 
>> https://codereview.qt-project.org/c/qt/qtbase/+/444859
> 
> I just abandoned this one. No one reviewed it and I'm pretty much
> sure they would set -1 or -2 with a similar kind of
> bureaucratic/gaslighting explanation.
> 
>> https://codereview.qt-project.org/c/qt/qtbase/+/427313
>> https://codereview.qt-project.org/c/qt/qtbase/+/417817
> 
> There's gaslighting again with statements like
> 
>> There is neither a bug, nor the need to change anything.
> 
> (QTBUG-124303 exists)
> 
> It feels pretty much like that reviewers just don't want to take their
> time to see what's the real state of the things in the modern Linux.
> 
>> OpenSuSE, RHEL, Debian and Ubuntu maintain correct DPI values in
>> XSettings. I haven't checked, if they also support Xresources and if
>> so, which one is leading.
> 
>> https://codereview.qt-project.org/c/qt/qtbase/+/471045
> 
> Reviewers said they can't reproduce the bug and gave -1 because of
> that but the bug report is still open. IDK what to do about that and I
> have tired aruging with them so abandoned.
> 
>> https://codereview.qt-project.org/c/qt/qtbase/+/501396
> 
> I'm tired on previous ones so just abandoned.
> 
>> https://codereview.qt-project.org/c/qt/qtbase/+/519479
> 
> I'm tired on previous ones so just abandoned.
> 
>> https://codereview.qt-project.org/c/qt/qtbase/+/515131
> 
> The only one I haven't abandoned but it's still ignored. Going to
> abandon it in a week if it won't get any attention...
> 
>> https://codereview.qt-project.org/c/qt/qtbase/+/528953
> 
> The only one merged.
> 
> 
> There was a thread on the mailing list on how to get new contributors.
> It was discussing problems like simplifying cloning the repo. I'd say
> the problem is not that but just that Qt doesn't really need or want
> contributions. It seems the maintainers are fine with current behavior
> and bugs and don't want any changes (contributions). The icing on the
> cake is that one has to ping for years just to see that the
> contributions aren't really wanted. Although perhaps most contributors
> get that by the reviewers silence itself, I guess a little amount of
> people ping for years or go to mailing list.
> 
> Another thing is that various proposals such as
> https://lists.qt-project.org/pipermail/development/2022-June/042650.html
> get ignored on the mailing list while bikeshedding on C++ things gets
> LOTS of replies. Well, even this thread I had to ping multiple times to
> get any attention.
> 
> Personally I now have way less desire to upstream patches to Qt. I
> tried to push some changes fixing font/style update on gerrit a week
> ago but the reaction so far looks like that it's not wanted either. It
> seems working around Qt bugs is a way more productive approach than
> fixing them.
> -- 
> Development mailing list
> Development at qt-project.org
> https://lists.qt-project.org/listinfo/development