[PATCH 5 of 8] manifest: add manifestlog.get to obtain subdirectory instances

[Durham Goode]

On 9/15/16 11:23 AM, Martin von Zweigbergk wrote:
> On Wed, Sep 14, 2016 at 4:04 PM, Durham Goode <durham at fb.com> wrote:
>> # HG changeset patch
>> # User Durham Goode <durham at fb.com>
>> # Date 1473893509 25200
>> #      Wed Sep 14 15:51:49 2016 -0700
>> # Node ID 33a7df42b989c555972280f6b84e2fac38babf7b
>> # Parent  d41da1522f8efb5bf5aa75a51f0093b1129b6b5a
>> manifest: add manifestlog.get to obtain subdirectory instances
>>
>> Previously manifestlog only allowed obtaining root level manifests. Future
>> patches will need direct access to subdirectory manifests as part of changegroup
>> creation, so let's add a get() function that knows how to deal with
>> subdirectories.
>>
>> diff --git a/mercurial/manifest.py b/mercurial/manifest.py
>> --- a/mercurial/manifest.py
>> +++ b/mercurial/manifest.py
>> @@ -1040,20 +1040,34 @@ class manifestlog(object):
>>           """Retrieves the manifest instance for the given node. Throws a KeyError
>>           if not found.
>>           """
>> -        if node in self._mancache:
>> -            cachemf = self._mancache[node]
>> -            # The old manifest may put non-ctx manifests in the cache, so skip
>> -            # those since they don't implement the full api.
>> -            if (isinstance(cachemf, manifestctx) or
>> -                isinstance(cachemf, treemanifestctx)):
>> -                return cachemf
>> +        return self.get('', node)
>>
>> -        if self._treeinmem:
>> -            m = treemanifestctx(self._revlog, '', node)
>> +    def get(self, dir, node):
>> +        """Retrieves the manifest instance for the given node. Throws a KeyError
>> +        if not found.
>> +        """
>> +        if dir:
>> +            if self._treeinmem:
>> +                m = treemanifestctx(self._revlog.dirlog(dir), dir, node)
> This does not use the _mancache and there doesn't seem to be a
> _mancache on the revlog either. I've forgotten whether or not you plan
> to have one manifestlog per repo or per directory, but I think you
> said per repo. So it seems like only the root directory manifests will
> be cached at this point. Am I reading that right? Perhaps you're
> fixing that later, but even then it seems like an unfortunate
> transitional step to lose it here.
>
You are right.  I'll send a new series with a patch (immediately after 
this patch) that changes the manifestlog's cache to have an entry per tree.