On Fri, 1 Jul 2016 21:05:09 -0700 Leo Gaspard <l...@gaspard.io> wrote:
Hey Leo, > Actually, I'd think if you give people push access to their patch branch > it may be easier for them than having to export a patch and update the > wiki: they already rebase the patches for themselves, they would just > have to git push and that would be done. but this never actually happens. I know of people who have even private collections of patches updated, but would never get the idea to push them to the wiki, as the entire structure is inconsistent across projects. To remove strain for admins, the best choice really is to automate the process, which I've already managed to do for st. On each new release-tag, a patch for this specific version is added to the wiki. Every other solution will fail, as it is both too much work for the wiki admins (as they already have enough to do) and won't encourage people to do it (one more "wall" to prevent contributions). > This idea of setup does not take into account the cost for maintenance > of a setup where selected people are allowed to push to selected > branches, as I have not (yet) inquired more into that. > > This idea does not take into account the keeping alive of old patches > either; which may be implemented by auto-generating a tag when a branch > is force-pushed, but requires even more setup from the suckless server > admins. A simpler solution would be to disable force-pushes, but this > would mean mergeing all the time and an unclean history for patches. > > This idea only takes into account the price for the patch-submitting > end-user. Yes, and that is the big issue here. In my opinion, only those people should have a say in this process who actually spend time on fixing the patches (like Matthias Schoth, Jochen Sprickerhof, Eric Pruitt, Ayrton, myself and others). They actually do real work instead of phantasizing here on this ML. Cheers FRIGN -- FRIGN <d...@frign.de>
