[Anarchid]

There's too much spam in #zkdev and makes it less nice than it could be

AI disclosure: Claude Opus 5.5 in Claude Code

[Anarchid]

Unsure about the fewer hooks, synchronize could maybe stay

[sprunk]

"Push" is imo ok as-is, despite some commits being a bit too verbose. This can only be triggered by somebody with write access, once per commit, it has to have successfully gone through review, and is the only one about something that actually affects the codebase itself.

[GoogleFrog]

I prefer the solution where people who write spammy PR descriptions are shamed into learning how to write shorter ones. Moss orbit seems to have learnt. For reference, inserting images of text is a terrible way to do it.

[Anarchid]

I prefer the solution where people who write spammy PR descriptions are shamed into learning how to write shorter ones. Moss orbit seems to have learnt. For reference, inserting images of text is a terrible way to do it.

I think this is not entirely the same. The immediate motivation was this,yes, but a nice, human-written, non-spammy PR nevertheless also does a huge wall of text in #zkdev, and i think this suboptimal

Putting long text into the first image twitter-style is degenerate, yes, but so would be putting substantive texts into comments, and i think same would be true for "any multi-paragraph text is forever banned".

Social solutions have their places where there is no technofix possible (yet), but i think this isn't one of those cases.

[sprunk]

a nice, human-written, non-spammy PR nevertheless also does a huge wall of text in #zkdev

Dropping a one-time wall of nice human-written text sounds fine though.