D4686: changegroup: reintroduce some comments that have gotten lost over the years
durin42 (Augie Fackler)
phabricator at mercurial-scm.org
Thu Sep 20 19:39:38 UTC 2018
durin42 created this revision.
Herald added a subscriber: mercurial-devel.
Herald added a reviewer: hg-reviewers.
REVISION SUMMARY
I got concerned about the correctness of the pruning logic, but I was
misreading it. I didn't figure that out until I walked all the way
back to https://phab.mercurial-scm.org/rHG0252abaafb8a54baab71e40cdee9761ecc25a670 from 20111, where I was finally able to see (in
the deleted side of the change!) a complete explanation from
https://phab.mercurial-scm.org/rHGb6d9ea0bc1070f75648ac9b729682a7066968fca in 2005.
REPOSITORY
rHG Mercurial
REVISION DETAIL
https://phab.mercurial-scm.org/D4686
AFFECTED FILES
mercurial/changegroup.py
CHANGE DETAILS
diff --git a/mercurial/changegroup.py b/mercurial/changegroup.py
--- a/mercurial/changegroup.py
+++ b/mercurial/changegroup.py
@@ -1071,8 +1071,14 @@
store = mfl.getstorage(tree)
if not self._filematcher.visitdir(store.tree[:-1] or '.'):
+ # No nodes to send because this directory is out of
+ # the client's view of the repository (probably
+ # because of narrow clones).
prunednodes = []
else:
+ # Avoid sending any manifest nodes we can prove the
+ # client already has by checking linkrevs. See the
+ # related comment in generatefiles().
prunednodes = self._prunemanifests(store, nodes, commonrevs)
if tree and not prunednodes:
continue
@@ -1164,6 +1170,10 @@
return linkrevnodes[x]
frev, flr = filerevlog.rev, filerevlog.linkrev
+ # Skip sending any filenode we know the client already
+ # has. This avoids over-sending files relatively
+ # inexpensively, so it's not a problem if we under-filter
+ # here.
filenodes = [n for n in linkrevnodes
if flr(frev(n)) not in commonrevs]
To: durin42, #hg-reviewers
Cc: mercurial-devel
More information about the Mercurial-devel
mailing list