OoO Vers la fin de l'après-midi du dimanche 28 février 2010, vers 16:46, Josselin Mouette <j...@debian.org> disait :
>> It would be far easier to let Python 2.6 be the default, then file (or >> upgrade) serious bugs and solve them in a week or two. > Yeah sure, let’s knowingly break dozens of packages by switching instead > of fixing them before and make it painless for users. Sorry, I really don't see any relation between FTBFS and breaking a package. I don't have any handy stat to support my claims (and I don't maintain enough Python packages to turn this affirmation into an universal one), but there are packages that fail to build from sources because they are not able to run unittests but work fine with Python 2.6 as default. Therefore, we get RC bugs for packages that build fine (with current default Python) and run fine (with all supported Python). I also tend to believe that there are a lot of packages that will just fail to run with Python 2.6 but will have no problem to build, because for most packages, building just means to copy files in the right location. The later we switch to Python 2.6, the more difficult it will be to catch those bugs. -- NON-FLAMMABLE, IS NOT A CHALLENGE NON-FLAMMABLE, IS NOT A CHALLENGE NON-FLAMMABLE, IS NOT A CHALLENGE -+- Bart Simpson on chalkboard in episode BABF13
pgpJUpisf8xbD.pgp
Description: PGP signature
