[email protected], le sam. 07 nov. 2026 04:53:57 +0000, a ecrit: > November 6, 2026 at 8:14 PM, "Samuel Thibault" <[email protected] > mailto:[email protected]?to=%22Samuel%20Thibault%22%20%3Csamuel.thibault%40gnu.org%3E > > wrote: > > jbranso--- via Bug reports for the GNU Hurd, le ven. 06 nov. 2026 20:00:31 > > +0000, a ecrit: > > > > > > > > November 6, 2026 at 1:12 PM, "Samuel Thibault" <[email protected] > > > mailto:[email protected]?to=%22Samuel%20Thibault%22%20%3Csamuel.thibault%40gnu.org%3E > > > > wrote: > > > Mikhail Karpov, le jeu. 05 nov. 2026 20:57:16 +0700, a ecrit: > > > > > > > > > > > a potentially better solution would be to adjust the signatures > > > > of these functions to match those of diskfs_create_protid and > > > > diskfs_make_peropen. > > > > > > > That would break all externally-maintained translators, I'd rather avoid > > > that. > > > > > > May I ask how many "externally-maintained translators" there are? > > > > > By definition we don't know. > > > > All we know is the incubator and hurdextras, so something like one or > > two dozen. And then there are private repositories we don't necessarily > > know about. > > I really hope that you don't mind me pushing back a bit... > > But how much work goes in the incubater / hurdextras ?
There is some, and there is interest in those funny projects. > But if someone is building code based on a Hurd library...then they shouldn't > be super upset if the library API slightly changes. Just update your code a > bit. Breaking them just for the sake of making the interface a bit different, is not really worth making people have to fix their code. > gtk2 vs. gtk3 vs gtk4 has some library changes no ? Yes, and it's a pain for applications, and only happens for very large API overhaul, and thus only every decade or so. Samuel
