Poor handling of Chromium-only bugs

[Evangelos Foutras]

If I am going to the trouble of triaging a bug and properly reporting it to an upstream bug tracker, I expect it to be handled efficiently by upstream developers. This has been my experience so far with most projects, but not with Chromium.

My most recent report is https://crbug.com/710701 and has to do with broken .jsx file syntax highlighting in DevTools. Here's how it has been dealt with:

1. First response does not consider that my opening message has enough information.
2. I respond with more information and a screencast showing the exact problem.
3. I get asked for a sample .html file and expected behavior. I doubt anyone so far has made an effort to understand the issue.
4. I self-host an example .jsx file and restate what the issue is and provide repro steps again.
5. A month later, a third person with @chromium.org address comes and closes the ticket as WontFix. He says that it's obsolete since "stable is now on M58", even though I was using Chromium 58 in my last comment.

Why should I continue forwarding some of the bug reports I receive for Chromium downstream in Arch Linux if the end result is similar to piping the bug report to /dev/null? Maybe my expectations are too high from a busy bug tracker such as Chromium's or perhaps I could do something differently?

I have complained about this "only Chrome matters" feeling in the past. While surely not a bad business decision, the situation is frustrating for actual users of Chromium (and doubly so for distro maintainers).

[Lei Zhang]

Thanks for taking the time to write this email. We all want our bugs
processed and fixed in a timely manner, but the reality is there are
55K open Chromium bugs, and 1800+ in Platform>DevTools alone. Given
these numbers, it is a busy bug tracker, so I hope you can understand
sometimes it takes time to triage bugs. Please also understand that
developers are humans, and humans make mistakes. However, some of
these mistakes are easy to correct, and we can reopen the bug of
course.

This isn't so much that only Chrome matters. It's that most Chromium
developers are not Linux Chromium packagers, and not everyone is an
expert with working with release branches. Many developers are only
used to working with builds from git checkouts, not source tarballs.
Many of them work with Chromium (not Chrome) on trunk, and rarely
touch release branches. So even though you are working with Chromium
source code, most developers do not understand how your copy of the
source code arrived in that state.

With regards to your bug:

(1) The first response happened within 48 hours, so at least
developers are trying to respond in a timely manner. I agree there
wasn't quite enough information. Since Chromium is open source, one
has no idea how your Chromium binary is built, so without more
context, no one has any clue has to why your Chromium binary is
behaves the way it does.

The information you provided in (2) was helpful. (3) was a fair
question. Thank you for responding in (4) with a sample file, and
insight into what you think the problem is.

I don't know what happened afterwards and why the bug got dropped. We
certainly could have done better there. I'm also not sure why the bug
got closed, but that's a mistake so I will reopen the bug and CC more
people to help take a look.

> --
> You received this message because you are subscribed to the Google Groups
> "chromium-packagers" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to chromium-packag...@chromium.org.
> To view this discussion on the web visit
> https://groups.google.com/a/chromium.org/d/msgid/chromium-packagers/309284bd-6dc1-c200-b50a-d7956fe5fd16%40foutrelis.com.

[Evangelos Foutras]

Thank you for the thoughtful reply. My takeaways are that Chromium has
indeed a very busy bug tracker and that Chromium developers don't
necessarily understand the way Linux distributions package Chromium.

[Lei Zhang]

I hope you continue to file bugs and not lose hope. [1] If there are
packaging-specific issues, and other developers are confused, feel
free to ping this group.

[1] Though sometimes bugs take a long time to fix. e.g. https://crbug.com/108144

[Evangelos Foutras]

On 19 May 2017 at 02:28, Lei Zhang <the...@google.com> wrote:

I don't know what happened afterwards and why the bug got dropped. We
certainly could have done better there. I'm also not sure why the bug
got closed, but that's a mistake so I will reopen the bug and CC more
people to help take a look.

It's worth noting that https://crbug.com/710701#c22 was closed for the second time as WontFix, this time without even an explanation as to why.

I work around this issue downstream in Arch by putting the parent commit into build/util/LASTCHANGE, but I don't like carrying such workarounds for ever. :-)