Mercurial on IronPython?

[Chuck.Kirschman at bentley.com]

-----Original Message-----
From: mercurial-bounces at selenic.com [mailto:mercurial-bounces at selenic.com] On Behalf Of Tim Groeneveld
Sent: Saturday, March 22, 2008 5:57 PM
To: mercurial at selenic.com
Subject: Re: Mercurial on IronPython?

On Sunday 23 March 2008 5:48:07 am Chuck.Kirschman at bentley.com wrote:
>> I've done a couple ports of smaller internal projects from C python to
>> IronPython.  Any win32 calls need to be ported to .NET, some libraries
>> don't exist, and process doesn't work the same way so that may be an issue.
>>  There are also some bugs in IPy that we encountered, though most of those
>> were fixed once we reported them.  But the big problem is startup speed.
>>  Hg is started often, and you want the results quickly.  You'll be adding
>> 10-30 seconds to every invocation to get the whole VM up and run through
>> byte codes.  Once it's all running performance seems to be roughly
>> equivalent, so it wasn't a problem for our long-running processes.

> Well, I have done absolutely no work with IronPython, I have not even touched 
> it to be honest with you. What I will say, however, is that applications 
> written in C# and compiled with mcs - the C# compiler in Mono start up within 
> 7-10 seconds on my machine, and this is even after rebooting the machine - so 
> I can't say that it is cache making it that fast.

I think you reinforce my point.  7-10 seconds to start the VM is excessive when most of the Hg commands I run take about 1 second to complete.  That is a pretty high price.  I didn't want someone to do all the work to port Hg to Ipy only to find out that nobody wants it because it's too slow.

chuck