yuja added inline comments.
INLINE COMMENTS
> quark wrote in localrepo.py:575
> I guess `hg log -r HASH --hidden` would go through this code path so `pass`
> makes sense.
Maybe. My concern was `getvisibilityexceptions()` "seemed" not
working correctly in that case. I know unfiltered repo has
yuja added a comment.
>> How long should the visibilityexception be preserved?
>
> The visibility exceptions must be preserved until we calculate filteredrevs
for that filter. After adding the expections, we clear filteredrevcache to make
sure we calculate that again.
Sounds like
quark added inline comments.
INLINE COMMENTS
> yuja wrote in localrepo.py:575
> raise ProgrammingError?
I guess `hg log -r HASH --hidden` would go through this code path so `pass`
makes sense.
REPOSITORY
rHG Mercurial
REVISION DETAIL
https://phab.mercurial-scm.org/D1285
To: pulkit,
pulkit added a comment.
In https://phab.mercurial-scm.org/D1285#27808, @yuja wrote:
> > I have spend a lot of time on this and I am unable to make this work
without making it class attribute because to calculate filterrevs,
unfilteredrepo is passed and I lose the reference to the
yuja added a comment.
> I have spend a lot of time on this and I am unable to make this work
without making it class attribute because to calculate filterrevs,
unfilteredrepo is passed and I lose the reference to the object on which I
initially stored the visibilityexceptions. If I call
pulkit added a comment.
In https://phab.mercurial-scm.org/D1285#27283, @yuja wrote:
> > As _visibilityexceptions is defined as a class attribute, mutating it is
likely to mutate it for all instances of repoview,
>
> Indeed.
I have spend a lot of time on this and I am unable
yuja requested changes to this revision.
yuja added a comment.
This revision now requires changes to proceed.
> As _visibilityexceptions is defined as a class attribute, mutating it is
likely to mutate it for all instances of repoview,
Indeed.
INLINE COMMENTS
> localrepo.py:575
> +
lothiraldan added inline comments.
INLINE COMMENTS
> repoview.py:189
> +# hidden revs which should be visible
> +_visibilityexceptions = set()
> +
As `_visibilityexceptions` is defined as a class attribute, mutating it is
likely to mutate it for all instances of repoview, I'm not sure
pulkit updated this revision to Diff 4092.
REPOSITORY
rHG Mercurial
CHANGES SINCE LAST UPDATE
https://phab.mercurial-scm.org/D1285?vs=4051=4092
REVISION DETAIL
https://phab.mercurial-scm.org/D1285
AFFECTED FILES
mercurial/localrepo.py
mercurial/repoview.py
CHANGE DETAILS
diff --git
pulkit updated this revision to Diff 4051.
REPOSITORY
rHG Mercurial
CHANGES SINCE LAST UPDATE
https://phab.mercurial-scm.org/D1285?vs=3956=4051
REVISION DETAIL
https://phab.mercurial-scm.org/D1285
AFFECTED FILES
mercurial/localrepo.py
mercurial/repoview.py
CHANGE DETAILS
diff --git
pulkit updated this revision to Diff 3956.
REPOSITORY
rHG Mercurial
CHANGES SINCE LAST UPDATE
https://phab.mercurial-scm.org/D1285?vs=3784=3956
REVISION DETAIL
https://phab.mercurial-scm.org/D1285
AFFECTED FILES
mercurial/localrepo.py
mercurial/repoview.py
CHANGE DETAILS
diff --git
quark accepted this revision.
quark added inline comments.
INLINE COMMENTS
> repoview.py:236-238
> +def addvisibilityexceptions(self, exceptions):
> +"""adds hidden revs which should be visible to set of exceptions"""
> +self._visibilityexceptions.update(exceptions)
Maybe
pulkit added a comment.
In https://phab.mercurial-scm.org/D1285#24461, @quark wrote:
> I think the state of "visibilityexception" should really be bounded into
whoever controls "filteredrevs", which is the `repoview` object. This allows
different exceptions to be set for different
pulkit updated this revision to Diff 3784.
pulkit edited the summary of this revision.
pulkit retitled this revision from "localrepo: add a new attribute
_visibilityexceptions and related API" to "repoview: add a new attribute
_visibilityexceptions and related API".
REPOSITORY
rHG Mercurial
14 matches
Mail list logo
