On Thu, May 08, 2025 at 04:50:28PM +0200, Ludovic Courtès wrote: > Hi, > > Nicolas Graves <ngra...@ngraves.fr> writes: > > > It's not the first time I run into something like this : > > > > $ ./pre-inst-env guix refresh emacs-haskell-mode > > > > [...]/gnu/packages/emacs-xyz.scm:2971:13: warning: 17.5 is greater than > > the latest known version of emacs-haskell-mode (16.1) > > > > This happens when the github interface has "releases" but stopped > > publishing them, even though some later tags exist. IMHO, the better > > developpement experience would be to use the latest tag when it's > > greater than the latest "release". > > (This should probably be a bug report.) > > More generally, I wonder what the ‘github’ updater buys us compared to > the ‘generic-git’ updater. Can the former do anything that the latter > cannot? Or does it perform better somehow? > > If not should we remove it?
Let's say it's faster than the generic-git updater. If the github updater says it's up-to-date then we'd want to run the generic-git updater too, and running both of them is definitely slower than running just one. I suppose if a project had releases but didn't tag the releases in the git repo then the github updater would be preferable, but that's the only example I can think of where we'd necessarily need the github updater. On the other hand, if we were to remove it we'd need to swap all the sources to git-fetch. I suppose we could instead change it so that after it parsed the URI it could send the repository over to the generic-git updater to check for updates. -- Efraim Flashner <efr...@flashner.co.il> אפרים פלשנר GPG key = A28B F40C 3E55 1372 662D 14F7 41AA E7DC CA3D 8351 Confidentiality cannot be guaranteed on emails sent or received unencrypted
signature.asc
Description: PGP signature
