On Thu, 16 Oct 2003 09:52:57 -0700, Stas Bekman <stas@stason.org> wrote: >> I think "yes", in that it should be reporting itself as 5.8.1 with 1 >> registered patch (MAINT #####) >> >> 5.8.1-to-be reported itself as 5.8.0 with a maint patch for most of the >> period between 5.8.0 and 5.8.1. IIRC Jarkko only advanced it to 5.8.1 >> (with a maint patch) shortly before the release candidates. >> >> Then again, if we are going to release 5.8.2 soon (for the hashing issues) >> then we'll be onto release candidates in the next few days anyway. :-) > >Is there any significant reason for not incrementing the release number >immediately after doing a release. This is just confusing for those who >maintain multiple builds of perl. And then somewhere down the road you have to >reinstall helluva lot things because suddenly the directories including 5.8.1 >should be 5.8.2. And there are lots of them. I'd rather switch to the 5.8.2 >notation right away and go with it all the way till it gets released, no >matter when this happens. Does it make any sense? Not to me. I consider 5.8.2 a "milestone", so only releases at or after that milestone should be labeled 5.8.2. Everything before that is 5.8.1 with local patches. I think only RCs should have the new version number, all dev snapshots/builds should use the previous one. There are also some subtle implications for use 5.8.2; when you go through old dev snapshots to search for a breaking change. But that is probably only a minor consideration. If you really want to use different version numbers for interim snapshots, then I think there should be dev and release version numbers: 5.8.2 public release 5.8.3 development 5.8.4 next public release after 5.8.2 But I find that excessive. Cheers, -JanThread Previous | Thread Next