future of HgKit

Andrey Somov py4fun at gmail.com
Mon Oct 5 03:10:24 CDT 2009


>> - it is unclear whether Jython will ever be able to run Mercurial
> 
> well, it not too far from running.
> i did play with it for some time, and found a
> couple of bugs in jython.
> fixing that and changing mercurial to not use
> buffer() [literally a ~10 lines patch] got me
> to commit, push, pull, checkout & co.
> 
> jython still misses a bz2 module to work
> with bundles

Why everybody says it is easy and nobody could make it work ?

> 
>> - Jython runs at least a few times slower the CPython
>> - currently Mercurial is heavily using C modules to improve performance,
>> since Jython has to stick to pure Python modules the performance
>> degradation will be even worse
> 
> hm.. jython in jitted and has diferent libraries.
> moreover, i think that most of the operiations
> are disk limited, not cpu.

Do not think but make an experiment. You will be impressed.

> 
>> - talking to a socket will allow to overcome the license restriction but
>>  anyway developers have to install Jython with Mercurial (Eclipse
>> cannot include it as CVS or EGit)
>>
>> I think using CPython through a socket may already improve the
>> situation. Anyway this is the first (and the only) step we can do.
>>
> 
> i dont understand what the socket interface change
> from the direct hg invocation.
> it does get away with the startup, but hg is
> really pretty fast already.
> it does offer a more parse friendly output?
> that maybe can be got with more templates?
> 

well, the suggestion was to launch Jython and keep it running (either as 
daemon or a process in Eclipse)



More information about the Mercurial mailing list