36104
|
1 Integration with the share extension needs improvement. Right now
|
|
2 we've seen some odd bugs, and the way we modify the contents of the
|
|
3 .hg/shared file is unfortunate. See wrappostshare() and unsharenarrowspec().
|
|
4
|
|
5 Resolve commentary on narrowrepo.wraprepo.narrowrepository.status
|
|
6 about the filtering of status being done at an awkward layer. This
|
|
7 came up the import to hgext, but nobody's got concrete improvement
|
|
8 ideas as of then.
|
|
9
|
|
10 Fold most (or preferably all) of narrowrevlog.py into core.
|
|
11
|
|
12 Address commentary in narrowrevlog.excludedmanifestrevlog.add -
|
|
13 specifically we should improve the collaboration with core so that
|
|
14 add() never gets called on an excluded directory and we can improve
|
|
15 the stand-in to raise a ProgrammingError.
|
|
16
|
|
17 Figure out how to correctly produce narrowmanifestrevlog and
|
|
18 narrowfilelog instances instead of monkeypatching regular revlogs at
|
|
19 runtime to our subclass. Even better, merge the narrowing logic
|
|
20 directly into core.
|
|
21
|
|
22 Reason more completely about rename-filtering logic in
|
|
23 narrowfilelog. There could be some surprises lurking there.
|
|
24
|
|
25 Formally document the narrowspec format. Unify with sparse, if at all
|
|
26 possible. For bonus points, unify with the server-specified narrowspec
|
|
27 format.
|
|
28
|
|
29 narrowrepo.setnarrowpats() or narrowspec.save() need to make sure
|
|
30 they're holding the wlock.
|
|
31
|
|
32 Implement a simple version of the expandnarrow wireproto command for
|
|
33 core. Having configurable shorthands for narrowspecs has been useful
|
|
34 at Google (and sparse has a similar feature from Facebook), so it
|
|
35 probably makes sense to implement the feature in core. (Google's
|
|
36 handler is entirely custom to Google, with a custom format related to
|
|
37 bazel's build language, so it's not in the narrowhg distribution.)
|