Re: [PATCH v2 0/5] MR2901: winegstreamer: Test and fix H264 timestamps and concatenated streams.

[Rémi Bernon (＠rbernon)]

On Thu Jun  1 17:15:46 2023 +0000, Zebediah Figura wrote:
> > Well I don't know that, it's been mostly a week without any activity
>
> and it's not the first time winegstreamer MR have been left unnoticed. I
> have been upstreaming winegstreamer patches for a big part of last year
> and the slow iteration have been making it very difficult. We have
> patches piling up in Proton once again, and diverging implementations
> already. I would very much like to avoid getting in the same situation
> as we've been.
> It had been six days, which is nominally a bit slow, but the slowness
> was largely due to (a) multiple winegstreamer patches from you in that
> time [and while I appreciate smaller patch series, that doesn't mean
> that I'm going to review three patch series all in the same day], (b)
> the weekend, (c) a US holiday during which I wasn't doing any work, and
> (d) patch 5/5 which does multiple things at once and was therefore hard
> to read.
> > I don't think these changes present any risk, they have been
>
> thoroughly tested, by Paul and I, but also with a lot of Wine tests, and
> the changes have been in Proton for a long while already. I also
> maintain wg_transform and their client, so, although I'm happy to take
> constructive comments, I'm ultimately going to be responsible for the changes.
> As I've stated when the issue of maintainership was brought up last, I
> still would very much like to be consulted on winegstreamer changes.
> Whether they have been tested or not, I find that winegstreamer code has
> been getting increasingly complex and difficult to maintain, in ways
> that are fixable and in ways that aren't, and I see things that can be
> improved in many incoming winegstreamer patches.
> If you want to call for no-confidence in my maintainership, fine, but
> please let's be clear about that and not abuse gitlab.

It is not a matter of no-confidence, although I could argue that being told that my statements are non-sensical in several occasions isn't a great way to build confidence. Perhaps a matter of diverging opinions, but moreover it is a matter of trying hard to keep downstream from diverging.

I took maintainership of the wg_transform code, not just because I wrote most of it, but also in order to help with the bandwidth, and avoid bothering two persons on every single change. I believe I have taken every comment into account very carefully, and haven't bypassed your review or Nikolay's for MF related things, on any meaningful change.