On Apr 18, 2012, at 03:09 PM, Paul Elliott wrote: >Nobody thinks python3 is important enough to have a debhelper infrastructure? I wouldn't say that. I'd say that the higher level tools are simply lagging behind demand. The low-level tools are available and won't significantly change. As Scott says, all that will change is that the higher level tools will come along and hide some of the lower level stuff from you. If you use the lower level tools today, nothing will break when the higher level tools are ready. When they are, and you decide to adopt them, you'll make one change, which will satisfyingly be to remove code (and I don't say that sarcastically :). But you could also decide to leave things just the way they are, since nothing will be broken. >My upstream tells me he does not know of any python3 programs using his >module. It's a chicken-and-egg problem. In order to increase the momentum toward Python 3, we need to port things from the bottom up. Higher level libraries can't port until the ones they depend on are ported. Applications can't port until all their dependencies are ported. Most users and developers get most of their packages from their distro, so the upstream's judge of popularity isn't always sufficient to judge the pent up demand. >Maybe I should delay python3 support then. IMHO, no you shouldn't delay. The path is clear, and easy, and you don't risk doing busy work that you'll just have to undo later. There's no reason not support Python 3 in your package, and every reason to do it. Cheers, -Barry
Attachment:
signature.asc
Description: PGP signature