This patch adds preliminary support for Haiku.  We still have to hand
tweak a few other files to get things to build.  We have further patches
to get python-2.5.4 to build, but it's failing some of the regression
tests so we're leaving those parts out of this submitted patch.
For more details on the progress of the Haiku port check here:
http://ports.haiku-files.org/wiki/dev-lang/python/2.5.4

Python 2.5 branch is now frozen: only security fixes are accepted. You 
should work on 2.6 or better on Python trunk for faster integration of 
your patches ;-) Your patch looks good.

What is the purpose of this patch? I.e. why are you submitting it to
this bug tracker?

On Tue, Jan 13, 2009 at 10:27 AM, STINNER Victor <report@bugs.python.org> wrote:
>
> STINNER Victor <victor.stinner@haypocalc.com> added the comment:
>
> Python 2.5 branch is now frozen: only security fixes are accepted. You
> should work on 2.6 or better on Python trunk for faster integration of
> your patches ;-) Your patch looks good.
>
> ----------
> nosy: +haypo
>
> _______________

2.6.x is next on my list, then we'll focus on 3.x.x.  Probably have a
patch cleaned and ready to post within a few days here.

Martin v. Löwis <martin@v.loewis.de> added the comment:

What is the purpose of this patch? I.e. why are you submitting it to
this bug tracker?

"This patch adds preliminary support for Haiku." In case you were
wondering what Haiku was, it is the open source replacement for BeOS,
see haiku-os.org for more background.
As for part two of your question, is there somewhere else I should be
posting patches?  If so direct me and I'll post them there.

-scottmc

> What is the purpose of this patch? I.e. why are you submitting it to
> this bug tracker?
> 
> "This patch adds preliminary support for Haiku." In case you were
> wondering what Haiku was, it is the open source replacement for BeOS,
> see haiku-os.org for more background.
> As for part two of your question, is there somewhere else I should be
> posting patches?  If so direct me and I'll post them there.

No, my question is whether you really want to see this patch considered
for inclusion into Python. If so, I guess I need to reject it, as we
decided some time ago to not accept patches for such minority platforms
anymore.

> we decided some time ago to not accept patches for such minority 
> platforms anymore.

I don't have any particular interest in BeOS variants, but this comes as 
a surprise given that http://python.org/about/ proclaims that "Python 
runs everywhere."

Maybe Haiku could become a supported platform instead of more or less 
defunct BeOS?  As far as I can tell there is no current BeOS port 
maintainer, but there is some crud left over from the better times.  If 
Haiku community can step in and clean up BeOS crud by either removing it 
or making sure it works on modern BeOS variants, this can only be a good 
thing.

> I don't have any particular interest in BeOS variants, but this comes as 
> a surprise given that http://python.org/about/ proclaims that "Python 
> runs everywhere."
> 
> Maybe Haiku could become a supported platform instead of more or less 
> defunct BeOS?

Maybe this can be discussed on python-dev, but there is a standing BDFL
pronouncement on this matter. It used to be the policy to accept patches
for any operating system, but this policy is reverted now.

> As far as I can tell there is no current BeOS port 
> maintainer, but there is some crud left over from the better times.  If 
> Haiku community can step in and clean up BeOS crud by either removing it 
> or making sure it works on modern BeOS variants, this can only be a good 
> thing.

No. The cost is too high, and there are more important issues to
resolve. Of course, the Haiku community can provide there fork of Python
if they want to (and they also get official blessing for forking if they
want to).

On Tue, Jan 13, 2009 at 10:38 PM, Martin v. Löwis
<report@bugs.python.org> wrote:
>
> Martin v. Löwis <martin@v.loewis.de> added the comment:
>
>> I don't have any particular interest in BeOS variants, but this comes as
>> a surprise given that http://python.org/about/ proclaims that "Python
>> runs everywhere."
>>
>> Maybe Haiku could become a supported platform instead of more or less
>> defunct BeOS?
>
> Maybe this can be discussed on python-dev, but there is a standing BDFL
> pronouncement on this matter. It used to be the policy to accept patches
> for any operating system, but this policy is reverted now.
>

I just joined that list, and will move further discussion of this
there, that is once I have a clean patch that can be applied to the
2.6.x branch.  One of Haiku's goals was to improve on POSIX
compliance, so in many ports that we've done (over 100 now) we've
actually had to un-workaround former BeOS workarounds.  So we'd be
fine with removing any BeOS cruft that we come across during a
potential Haiku port/fork.

>> As far as I can tell there is no current BeOS port
>> maintainer, but there is some crud left over from the better times.  If
>> Haiku community can step in and clean up BeOS crud by either removing it
>> or making sure it works on modern BeOS variants, this can only be a good
>> thing.
>
> No. The cost is too high, and there are more important issues to
> resolve. Of course, the Haiku community can provide there fork of Python
> if they want to (and they also get official blessing for forking if they
> want to).
>

I disagree about the cost being too high, as long as our patches are
clean I don't see an issue here, if anything we might find things that
are broken in python that weren't uncovered previously, I think it's a
win-win.  I don't think there's going to be a whole lot that needs to
be patched to get it working on Haiku, we've got a mostly working
2.5.2 binary already, but it's failing a few of the regression tests,
so we are looking into why on those and have left those patches out of
what was submitted, as we wouldn't want to have to undo one of our
previous patches.

-scottmc

Ok, I'm closing this patch as rejected. If the discussion on python-dev
turns out that Haiku support should be added, then please create a new
issue (with a patch for 2.7 - clearly, for 2.6, no new features can be
added).

FWIW, I've launched a discussion on python-dev.  If the OP is willing to
maintain the deltas over a long period of time (and perhaps provide a
buildbot for Haiku), then I fully support his effort.

[Copy of a post I just made to python-dev]

I'm with Martin. In these days of distributed version control systems, I
would think that the effort for the Haiku folks to maintain a branch of
Python in their own version control would be minimal. It is likely that
for each new Python version that comes out, initially it is broken on
Haiku, and then they have to go in and fix it. Doing that in their own
version control has the advantage that they don't have to worry about
not breaking support for any other minority operating systems, so I
expect that all in all the cost will be less for them than if they have
to submit these patches to core Python; and since that will definitely
be less work for core Python, I would call that a win-win solution.