Hi Lukas, On Mon, May 06, 2019 at 06:51:57PM +0200, Lukas Tribus wrote: > There is room for improvement here. Can you confirm that attaching a > patch file per commit to the email would fix this usability issue?
I'd say yes, provided the attachments are prefixed with a sequence number, like git-format-patch does. However I think we should set reasonable limits to 10 patches or less to save all list members from being bombed if someone sends a huge (or even bogus) PR. I think that till a handful of patches (let's say two hands :-)) it's still possible to let various participants give their opinion on different patches. When you see patch 137/375 it's unlikely that anyone will have a look at it so then falling back to the current mode would work better. Similarly I think that if a patch is too large the series should not be forwarded. I'm well aware that "too large" is a bit vague. The idea is that few people if any will spend their time reviewing a 1000-line patch. But this could be applied to the whole series if easier. In the end I think that such thresholds will serve as a hint for reviewers about what to expect when seeing the announce. > I just finally fixed the script for authors with non-ascii names (it > crashed previously), I can add the file attachments to my todo list, > if useful. This could definitely be useful. But don't spend too much valuable time on this, as this is not the source of contribs I'd like to encourage the most considering the amount of fixes they usually require before being merged. By the way I noticed that something changed recently on Github, the curl command receives a redirect and now needs -L to follow it. Thanks! Willy