On 3/1/2012 9:23 AM, Thorsten Glaser wrote: > Jakub Lach <jakub_lach <at> mailplus.pl> writes: > >> and how it worked before? mksh had letters >> added in version from start. > > Actually, mksh always had an uppercase R plus a number,
The issue isn't the R, it's the letter that comes after the number. The way that we test version strings R40 is greater than R40c. That's to accommodate the usual convention of having <number><letter> indicate a beta version. We had to pick one way or the other, which is not to say that either one is wrong, only that we had to choose one. We have the ability to handle this appropriately in the port's Makefile, but for whatever (possibly entirely valid) reason miwi chose not to do that. What I suspect the OP saw was a situation something like R30 evaluated to greater than R20b, then R40 evaluated to greater than R30*, but when R40c came out he was stuck in the default version test logic which caused his installed R40 to seem newer than R40c. > because I > originally did not want to make minor versions or other versioning > games. Well, I'm not sure what kind of games you're worried about, but there is nothing inherently wrong with minor version numbers. :) The meme of major versions introduce new features and minor versions are bug fixes is well understood, and very common. > The website has a recommendation for packagers that need to map > these versions to numbers: R22 maps to 22.1, R22x maps to 22.y > for x=[bcdefgh] → y=[2345678], and there will never be an R22i > or up, so 22.9.yyyymmdd can be used for -current. Well hopefully miwi is paying attention to your excellent advice this time around. :) Doug _______________________________________________ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"