[Python-Dev] Support for the Haiku OS

"Martin v. Löwis" martin at v.loewis.de
Thu Jan 15 20:36:54 CET 2009


> I don't see such a commitment in this case, but if a
> believable one comes up I'm sure Martin would happily revert his
> position.

Indeed. I have myself added support for AtheOS, even though I had
never used the system. The AtheOS maintainer ran away, the code
rotted, and eventually get ripped out. The same happened for a lot
of other systems whose code was recently removed. Unless a core
committer actively works on the port, I see little change that it
remains in a usable shape over time.

In case you wonder why it might break: configure gets rewritten to
have new features (like --enable-shared), and then people contributing
such a change can only contribute it for the systems they can test it
on. New variables get introduced, which don't get set for Haiku, and
the Makefile breaks - perhaps in a trivial way, but nonetheless useless
for the end user who wants to build Haiku.

In addition, I expect that a *true* Haiku port would have many
additional modules that provide access to API specific to the system
(such as support for GUI applications). The true Haiku port will
have to provide all these things, but they won't be in the core.
So people using Python on Haiku will have to get it from elsewhere,
anyway.

So I think adding the patch to Python has little advantage to Haiku
users, and (if the patch is small) maintaining it outside of
python.org should be little effort to the authors of the patch.
(e.g. compared to writing all these other modules)

Regards,
Martin


More information about the Python-Dev mailing list