Ezio Melotti <ezio.melo...@gmail.com> added the comment: > I think about it in the reverse: You want a featurefull library, and > disable some things (zlib, ssl, threads) only if you specifically > don’t want them (if you’re a Twisted fan for example <async wink>).
I have a few arguments in favor of my position: 1) Python and its test suite should always run without errors, even when some dependencies are missing (I've found and fixed several issues related to zlib because it was missing here); 2) The default Python installed with the OS will use whatever dependencies are already provided, so if you add all the missing ones manually, you'll be testing in a scenario that is different from the "default" one (you might argue that if no one installs all the missing dependencies, we can't know if they all work together everywhere (IIRC at some point we had a discussion about having a buildbot without the optional dependencies too)); 3) Hard work pays off later. Laziness pays off now! > the addition of this message and further edits to its wording were done > exactly to avoid people worrying that their python binary is not usable. But it's still missing an "optional" before modules imho. The second message that says to "look in setup.py in detect_modules()" doesn't seem too useful too (but maybe I should take a look ;). There are a couple of threads on Core-Mentorship by people that got confused by these messages (and some include discussions about finding the dependencies). > Okay; do we need to collect info for all major OSes > (I’m taking the list of stable 3.2 buildbots to determine “major”)? You can also do it incrementally, adding e.g. "For Debian and derivative systems you can use ``aptitude build-dep pythonX.Y``." and then add other items to the list as soon as we figure them out (you can also try to figure them out in advanced, as long as finding the command for some obscure platform doesn't keep on hold the list of commands for other well known platforms). ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue13472> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
