yeah, its tricky. i think it depends on how closely the project tracks cpython.
to minimize the stackless/cpython diff and to use mercurial to help with merging in both directions it would probably be easier to base a new stackless mercurial repository on the current cpython repository. preserving the complete history of the stackless repository in a clone of the cpython repository would take some effort but it may be possible. i'm willing to spend some time on either this approach or a clean slate if everyone is interested in moving to mercurial. thanks, gene On Tue, Mar 29, 2011 at 9:28 AM, Jeff Senn <[email protected]> wrote: > Thanks for the explanation! > > On Mar 29, 2011, at 10:16 AM, Kristján Valur Jónsson wrote: > > Well, we have the permission to host stackless at hg.python.org. The > question is just how. > It will be a new project and the question is whether we should try to > recreate the svn history in hg, (complicated, apparently) or just start > afresh with a clear slate and a few branches. > > > Though the historian in me cringes horribly ... *practically* speaking the > latter is probably sufficient (given that there probably aren't too many > projects around stuck in particular branches of stackless which will later > need carefully tended paths forward). > > Perhaps more significantly how will it be represented? as a branch from the > particular python version? or completely separate code that shares > a lot of similar pieces? I'm not sure I understand enough about merging in > this new world to know how to think about it... I just hope it's not > any *worse* than svnmerge :-) > > As for Nagare: > At pycon, we were sprinting on adding picking for (vanilla) python’s built > in iterators. This would reduce the diff between Stackless and CPython. > We have a clone of hg.python.org/cpython sitting at > http://nagare.org:8000/cpython and this is the repository that we used to > synchronize our work. > It’s just a temporary one, for this particular code sprint. > > K > > > *From:* [email protected] [mailto: > [email protected]] *On Behalf Of *Jeff Senn > *Sent:* 29. mars 2011 14:01 > *To:* The Stackless Python Mailing List > *Subject:* Re: [Stackless] Me Bad Re: Stackless sprint at PyCon > > So what is the plan for a source repository for stackless? (now that python > proper is moving/has moved to hg.python.org) > > K, you mentioned Nagare.... is there one in the Nagare repos somewhere? Is > that going to be "official"? Should this: > http://zope.stackless.com/svn/sdocument_view get updated? > > > On Mar 29, 2011, at 9:44 AM, Kristján Valur Jónsson wrote: > > > I just uploaded my most recent changes. I think we have all of the > iterators covered in cpython now, apart from “generators” > Please take a look. > The next step is to get this stuff into cpython proper, I´ll start the > process for that. > > K > > *From:* [email protected] [mailto: > [email protected]] *On Behalf Of *Kristján Valur Jónsson > *Sent:* 23. mars 2011 09:16 > *To:* The Stackless Python Mailing List > *Cc:* Andrew Francis > *Subject:* Re: [Stackless] Me Bad Re: Stackless sprint at PyCon > > I’m doing some review and cleanup of your contributions and uploading it > back to the nagare HG host. > I think all of you did a great job getting your hands dirty with the C > python api, but there are still some reference counting errors (mostly > leaks) in there and also potential crash cases with some of the __setstate__ > things. Mostly we have to be vigilant in error checking so that unpicling > bad data cannot crash the interpreter.+ > > Cheers, > > K > ... > > _______________________________________________ > Stackless mailing list > [email protected] > http://www.stackless.com/mailman/listinfo/stackless > > > > _______________________________________________ > Stackless mailing list > [email protected] > http://www.stackless.com/mailman/listinfo/stackless >
_______________________________________________ Stackless mailing list [email protected] http://www.stackless.com/mailman/listinfo/stackless
