py3: system-stringify list of attributes to be forwarded from commandserver.py
# skip-blame just some r'' prefixes
py3: byte-stringify most literals in test-commandserver.t
print() calls will be replaced by bprint().
# skip-blame just tons of b'' prefixes.
localrepo: ensure we properly %-format int in exception throw
I'm not thrilled with this, but it'll do.
Differential Revision: https://phab.mercurial-scm.org/D5107
py3: add a r'' prefix in mercurial/exchange.py
# skip-blame because just r'' prefix
This fixes test-narrow-acl.t on py3 which was broken by one of the
earlier patches.
Differential Revision: https://phab.mercurial-scm.org/D5149
py3: add 5 new passing tests to whitelist caught by buildbot
Thanks to everyone who is putting efforts in making hg py3 compatible.
Differential Revision: https://phab.mercurial-scm.org/D5148
py3: restore perfstartup() prior to
b456b2e0ad9f on Windows
Otherwise the test errors out with:
--- c:/Users/Matt/projects/hg_py3/tests/test-contrib-perf.t
+++ c:/Users/Matt/projects/hg_py3/tests/test-contrib-perf.t.err
@@ -184,6 +184,8 @@
$ hg perfrevrange
$ hg perfrevset 'all()'
$ hg perfstartup
+ 'b'c:' is not recognized as an internal or external command,
+ operable program or batch file.
$ hg perfstatus
$ hg perftags
$ hg perftemplating
help: document the server capabilities added by the LFS extension
I didn't bother marking these experimental because it references the extension
that is already marked experimental.
commands: adjust metavariables as appropriate
Apart from looking better in hg help command, these strings are also helpful
when generating shell completions programmatically.
match: fix up a repr to not crash on Python 3
Differential Revision: https://phab.mercurial-scm.org/D5120
narrow: when widening, don't include manifests the client already has
When widening, we already don't include the changelog (since
f1844a10ee19) and files that the client already has (since
c73c7653dfb9). However, we still include all manifests needed for the
new narrowspec. When using flat manifests, that means we resend all
the manifests even though the client necessarily has all of them. For
tree manifests, we unnecessarily resend the root manifests and any
subdirectory manifests that the client already has.
This patch makes it so we no longer resend manifests that the client
already has. It does so by passing an extra matcher to the changegroup
packer and it uses that for filtering out directories matching the old
matcher's visitdir(). For consistency between directories and files,
it also makes the filtering of files look at both old and new matcher
rather than passing in a diff matcher as we did before.
Differential Revision: https://phab.mercurial-scm.org/D4895