-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Ralf S. Engelschall wrote: > On Fri, Apr 14, 2006, Max Bowsher wrote: > >> [...] >> OK, I'll backport it. Just out of curiosity, is there any problem caused >> by the weird UUIDs? > > I'm not aware of any functional problems. > > I'm just still trying hard to fix the existing UUID implementations > to make sure that everything which looks like a UUID actually is also > a standard-conforming UUID. You would be surprised how many software > exists on the net which claim to produce "UUIDs" but actually generates > nothing more than 128 bit of pseudo-randomness, printed as hex values > with a few dashes between them ;-) > >> Should I be considering backporting it to 0.9.x as well? (Anyone using >> Apache 2.0.x must continue to use APR 0.9.x, and 0.9.x will remain the >> default bundled with Subversion for the forseeable future.) >> [...] > > If APR 0.9.x will be still productive even in the forseeable future I > suggest to backport the fix to APR 0.9.x, too. Because even if there is > most of the time no functional problem with the bogus UUIDs (as long as > nobody tries to _decode_ them as I do with OSSP uuid ;-) it looks rather > unprofessional when modern applications like Subversion generate such > bogus UUIDs into every repository. > > Thanks for your investigation.
OK, backports committed to 1.2.x and 0.9.x branches. Max. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.1 (Cygwin) iD8DBQFEQBRofFNSmcDyxYARAhPHAKDWpDjYW6WyyN9TB23mec8pdiYAZACeJrJ/ KisLMv6dk/bzSFvKnER0jB4= =Emsa -----END PGP SIGNATURE-----
