treemanifests: store whether a lazydirs entry needs copied after materializing
Due to the way that things like manifestlog.get caches its values, without
making a copy (if necessary) after calling readsubtree(), we might end up
adjusting the state of the same object on different contexts, breaking things
like dirty state tracking (and probably other things).
Differential Revision: https://phab.mercurial-scm.org/D4874
# Blacklist for a full testsuite run with fsmonitor enabled.
# Used by fsmonitor-run-tests.
# The following tests all fail because they either use extensions that conflict
# with fsmonitor, use subrepositories, or don't anticipate the extra file in
# the .hg directory that fsmonitor adds.
#### mainly testing eol extension
test-eol-add.t
test-eol-clone.t
test-eol-hook.t
test-eol-patch.t
test-eol-tag.t
test-eol-update.t
test-eol.t
test-eolfilename.t
#### mainly testing largefiles extension
test-issue3084.t
test-largefiles-cache.t
test-largefiles-misc.t
test-largefiles-small-disk.t
test-largefiles-update.t
test-largefiles-wireproto.t
test-largefiles.t
test-lfconvert.t
#### mainly testing nested repositories
test-nested-repo.t
test-push-warn.t
test-subrepo-deep-nested-change.t
test-subrepo-recursion.t
test-subrepo.t
#### fixing these seems redundant, because these don't focus on
#### operations in the working directory or .hg
test-debugextensions.t
test-extension.t
test-help.t