On 13 February 2013 19:35, stephane ducasse <stephane.duca...@free.fr> wrote: > > On Feb 13, 2013, at 6:28 PM, Frank Shearar <frank.shea...@gmail.com> wrote: > >> On 13 February 2013 16:03, stephane ducasse <stephane.duca...@free.fr> wrote: >>> >>> On Feb 13, 2013, at 4:06 PM, Sean P. DeNigris <s...@clipperadams.com> wrote: >>> >>>> stephane ducasse wrote >>>>> @sean what was the problem because I tested it several time on 1.4 and 2.0 >>>> >>>> The tests were failing due to FS differences between 1.4 and 2.0. I >>>> uploaded >>>> fixes, so we could do a 4.5.1 with the latest packages (4.5 is tagged >>>> release so we shouldn't change it). Also, I moved the dependency on the >>>> CommandShell version from the baseline (where I originally put it in >>>> error), >>>> to the versions. >>> >>> Nobody use this 4.5 so we could just modify it. Because 4.5.1 looks boring >>> to me. >> >> As a general rule this is a very bad idea. The version number should >> allow you to say with certainty what's in an artifact. Just modifying >> it means that "4.5" tells you just about nothing about someone's setup >> when they say "4.5 is broken". > > frank I do not really need a lesson. ***I*** created the wrong 4.5 version > so probably nobody could reliably use it. So what is the point to get a major > version > broken. > I do not want to always have to remember that I should use 4.5.1.1 > better 4.5
Look, it's your software, so it's your call. If it was _my_ software - and it's not - I'd just figure out how I managed to release something broken, fix that, and release a 4.5.1. It's not like you're going to run out of version numbers. frank > Stef > >> >> frank >> >>>> For CommandShell, I moved CommandShellPharo into a #'pharo' block, and MVC >>>> into a #'squeak' one. >>>> >>>> >>>> >>>> -- >>>> View this message in context: >>>> http://forum.world.st/The-monkey-is-back-in-town-tp4658091p4669672.html >>>> Sent from the Pharo Smalltalk mailing list archive at Nabble.com. >>>> >>> >>> >> > >