[PATCH] revset: added generatorset class with cached __contains__ method

[Pierre-Yves David]

On 02/18/2014 02:29 PM, Lucas Moscovicz wrote:
> # HG changeset patch
> # User Lucas Moscovicz <lmoscovicz at fb.com>
> # Date 1391642591 28800
> #      Wed Feb 05 15:23:11 2014 -0800
> # Node ID 7dfddd4a822ee7e5a3ce6f39a0bdf9c533cd76ff
> # Parent  c29948fed40a2d9755ecaa01ec05bfa542f65670
> revset: added generatorset class with cached __contains__ method
>
> diff --git a/mercurial/revset.py b/mercurial/revset.py
> --- a/mercurial/revset.py
> +++ b/mercurial/revset.py
> @@ -2131,6 +2131,37 @@
>       def set(self):
>           return set([r for r in self])
>
> +class generatorset(object):
> +    """Wrapper structure for generators that provides lazy membership."""

Could use more doc about all its behavior.

> +    def __init__(self, gen):
> +        self._gen = gen
> +        self._iter = iter(gen)
> +        self._cache = {}
> +
> +    def __contains__(self, x):
> +        if x in self._cache:
> +            return self._cache[x]
> +
> +        while True:
> +            try:
> +                l = self._iter.next()
> +                self._cache[l] = True
> +                if l == x:
> +                    return True
> +            except (StopIteration):
> +                break
> +
> +        self._cache[x] = False
> +        return False
> +
> +    def __iter__(self):
> +        for item in self._gen:
> +            self._cache[item] = True
> +            yield item
> +
> +    def set(self):
> +        return self

Sounds wrong to me, is there precendence of a set() method returning a 
non set object?