sapling/tests/test-fb-hgext-treemanifest-noflat.t
Jun Wu 9dc21f8d0b codemod: import from the edenscm package
Summary:
D13853115 adds `edenscm/` to `sys.path` and code still uses `import mercurial`.
That has nasty problems if both `import mercurial` and
`import edenscm.mercurial` are used, because Python would think `mercurial.foo`
and `edenscm.mercurial.foo` are different modules so code like
`try: ... except mercurial.error.Foo: ...`, or `isinstance(x, mercurial.foo.Bar)`
would fail to handle the `edenscm.mercurial` version. There are also some
module-level states (ex. `extensions._extensions`) that would cause trouble if
they have multiple versions in a single process.

Change imports to use the `edenscm` so ideally the `mercurial` is no longer
imported at all. Add checks in extensions.py to catch unexpected extensions
importing modules from the old (wrong) locations when running tests.

Reviewed By: phillco

Differential Revision: D13868981

fbshipit-source-id: f4e2513766957fd81d85407994f7521a08e4de48
2019-01-29 17:25:32 -08:00

84 lines
2.4 KiB
Perl

TODO: Make this test compatibile with obsstore enabled.
$ setconfig experimental.evolution=
$ . "$TESTDIR/library.sh"
This file tests that normal mercurial operations almost never read the flat manifests
$ cat >> $TESTTMP/flatcheck.py <<EOF
> import sys, traceback
> from edenscm.mercurial import extensions, manifest
> def uisetup(ui):
> extensions.wrapfunction(manifest.manifestrevlog, 'revision', readmf)
> def readmf(orig, self, nodeorrev, **kwargs):
> if nodeorrev != -1:
> print >> sys.stderr, 'read flat manifest'
> stack = traceback.extract_stack()
> print >> sys.stderr, ''.join(traceback.format_list(stack[-3:-2]))
> return orig(self, nodeorrev, **kwargs)
> EOF
$ hg init master
$ hg clone -q ssh://user@dummy/master client
- Add a bunch of files so the manifest is large enough to use deltas
$ cd master
$ echo a >> a
$ echo a >> b
$ echo a >> c
$ echo a >> d
$ echo a >> e
$ echo a >> f
$ echo a >> g
$ echo a >> h
$ hg commit -Aqm 'add a-f'
$ echo a >> a
$ hg commit -Aqm 'modify a'
$ cd ../client
$ cat >> .hg/hgrc <<EOF
> [extensions]
> fastmanifest=
> flatcheck=$TESTTMP/flatcheck.py
> treemanifest=
>
> [remotefilelog]
> usefastdatapack=True
> reponame=master
>
> [fastmanifest]
> usetree=True
> usecache=False
>
> [treemanifest]
> autocreatetrees=True
> EOF
# Test there is one flat read is expected on the first pull, since
# manifest.revdiff cannot hit the fast path since the first manifest is not a delta.
$ hg pull -q -r 0
read flat manifest
File "*/mercurial/revlog.py", line *, in revdiff (glob)
self.revision(rev* (glob)
# Test that no flat manifests are read during pull and update
$ hg pull -q -r 1
$ hg up 0
8 files updated, 0 files merged, 0 files removed, 0 files unresolved
# Test only one flat read is expected on each commit, to get the p1 fulltext to
# produce the delta.
$ echo a >> b && hg commit -Aqm 'modify b'
read flat manifest
File "*/fastmanifest/implementation.py", line *, in add (glob)
p1text = origself.revision(p1)
# Test that rebase access the flat text only once, for the final commit
$ hg rebase -d 1 -r 2
read flat manifest
File "*/fastmanifest/implementation.py", line *, in add (glob)
p1text = origself.revision(p1)
rebasing 2:667a26a14261 "modify b" (tip)
saved backup bundle to $TESTTMP/client/.hg/strip-backup/667a26a14261-d769c687-rebase.hg (glob)