<br><br><div class="gmail_quote">On Fri, Mar 27, 2009 at 11:50, Paul Moore <span dir="ltr"><<a href="mailto:p.f.moore@gmail.com">p.f.moore@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
2009/3/27 Collin Winter <<a href="mailto:collinw@gmail.com">collinw@gmail.com</a>>:<br>
<div class="im">> In particular, Windows support is one of those things we'll need to<br>
> address on our end. LLVM's Windows support may be spotty, or there may<br>
> be other Windows issues we inadvertently introduce. None of the three<br>
> of us have Windows machines, nor do we particularly want to acquire<br>
> them :), and Windows support isn't going to be a big priority. If we<br>
> find that some of our patches have Windows issues, we will certainly<br>
> fix those before proposing their inclusion in CPython.<br>
<br>
</div>On the assumption (sorry, I've done little more than read the press<br>
releases so far) that you're starting from the CPython base and<br>
incrementally patching things, you currently have strong Windows<br>
support. It would be a shame if that got gradually chipped away<br>
through neglect, until it became a big job to reinstate it.<br>
<br>
If the Unladen Swallow team doesn't include any Windows developers,<br>
you're a bit stuck, I guess, but could you not at least have a Windows<br>
buildbot which keeps tabs on the current status? Then you might<br>
encourage interested Windows bystanders to check in occasionally and<br>
maybe offer fixes.<br>
<br>
As things stand, the press releases give me the impression (as a<br>
Windows user without a lot of time to invest in contributing) that<br>
this project is irrelevant to me, and I should ignore it until you<br>
announce "proper" Windows support. By which time, it may have fallen<br>
completely off my radar. (On a smaller scale, this happened with<br>
virtualenv - I found to my surprise that it now supported Windows, and<br>
had for some time without me realising, because when it started it was<br>
Unix-only and I had not bothered to keep track of it). Maybe again<br>
it's something that could be clarified in the announcements.</blockquote><div><br>It's not a matter of chipping away support. It's a matter of wishing to not write our own JIT, but rather leverage other people's work. That currently means LLVM, but LLVM has a weak Windows story at the moment. Of course, LLVM has little Windows support because it doesn't have any Windows users :-) The changes done so far are (mostly) orthogonal to Windows (the actual performance benefits may depend a little on the platform), but the future work will not be. On the other hand, getting the Windows story straightened out is mostly a matter of getting Windows support in LLVM, and not work specific to Unladen Swallow; hopefully, we can use some of the Windows knowledge on python-dev or elsewhere in the world for that.<br>
</div></div><br>Notice how I said 'currently' and 'at the moment' and 'future work'. LLVM is also a work in progress, See, for instance, <a href="http://llvm.org/docs/GettingStartedVS.html">http://llvm.org/docs/GettingStartedVS.html</a> . (And if we were to write our own JIT, we would have the same problem but worse: none of us would be able to write an effective one for Windows, if at all, and we would have a much smaller developer pool to work with. And it would take much longer in the first place.)<br>
<br>-- <br>Thomas Wouters <<a href="mailto:thomas@python.org">thomas@python.org</a>><br><br>Hi! I'm a .signature virus! copy me into your .signature file to help me spread!<br>