Current state of revset performance.

[Pierre-Yves David]

On 10/31/2014 06:37 PM, Martin von Zweigbergk wrote:
> Speaking of revset performance, I just found another case of a matcher
> that gets re-created for each revision. This time it was in the status
> functions ('modifies(filename)' etc). I'll send a patch shortly,
> although it's probably not a very commonly used feature.

Thanks for catching it!

I would like to remember all that the revset code is in terrible 
terrible terrible shape. Anyone staring at it for more than 2 minutes 
will probably have a trolley of patch to submit. It is probably a good 
entry point if one wants its 3 years old to get an open-source contribution.

-- 
Pierre-Yves David