match: making visitdir() deal with non-recursive entries
Primarily as an optimization to avoid recursing into directories that will
never have a match inside, this classifies each matcher pattern's root as
recursive or non-recursive (erring on the side of keeping it recursive,
which may lead to wasteful directory or manifest walks that yield no matches).
I measured the performance of "rootfilesin" in two repos:
- The Firefox repo with tree manifests, with
"hg files -r . -I rootfilesin:browser".
The browser directory contains about 3K files across 249 subdirectories.
- A specific Google-internal directory which contains 75K files across 19K
subdirectories, with "hg files -r . -I rootfilesin:REDACTED".
I tested with both cold and warm disk caches. Cold cache was produced by
running "sync; echo 3 > /proc/sys/vm/drop_caches". Warm cache was produced
by re-running the same command a few times.
These were the results:
Cold cache Warm cache
Before After Before After
firefox 0m5.1s 0m2.18s 0m0.22s 0m0.14s
google3 dir 2m3.9s 0m1.57s 0m8.12s 0m0.16s
Certain extensions, notably narrowhg, can depend on this for correctness
(not trying to recurse into directories for which it has no information).
Test update logic when there are renames or weird same-name cases between dirs
and files
Update with local changes across a file rename
$ hg init r1 && cd r1
$ echo a > a
$ hg add a
$ hg ci -m a
$ hg mv a b
$ hg ci -m rename
$ echo b > b
$ hg ci -m change
$ hg up -q 0
$ echo c > a
$ hg up
merging a and b to b
warning: conflicts while merging b! (edit, then use 'hg resolve --mark')
0 files updated, 0 files merged, 0 files removed, 1 files unresolved
use 'hg resolve' to retry unresolved file merges
[1]
Test update when local untracked directory exists with the same name as a
tracked file in a commit we are updating to
$ hg init r2 && cd r2
$ echo root > root && hg ci -Am root # rev 0
adding root
$ echo text > name && hg ci -Am "name is a file" # rev 1
adding name
$ hg up 0
0 files updated, 0 files merged, 1 files removed, 0 files unresolved
$ mkdir name
$ hg up 1
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
Test update when local untracked directory exists with some files in it and has
the same name a tracked file in a commit we are updating to. In future this
should be updated to give an friendlier error message, but now we should just
make sure that this does not erase untracked data
$ hg up 0
0 files updated, 0 files merged, 1 files removed, 0 files unresolved
$ mkdir name
$ echo text > name/file
$ hg st
? name/file
$ hg up 1
abort: *: '$TESTTMP/r1/r2/name' (glob)
[255]
$ cd ..
#if symlink
Test update when two commits have symlinks that point to different folders
$ hg init r3 && cd r3
$ echo root > root && hg ci -Am root
adding root
$ mkdir folder1 && mkdir folder2
$ ln -s folder1 folder
$ hg ci -Am "symlink to folder1"
adding folder
$ rm folder
$ ln -s folder2 folder
$ hg ci -Am "symlink to folder2"
$ hg up 1
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd ..
#endif
#if rmcwd
Test that warning is printed if cwd is deleted during update
$ hg init r4 && cd r4
$ mkdir dir
$ cd dir
$ echo a > a
$ echo b > b
$ hg add a b
$ hg ci -m "file and dir"
$ hg up -q null
current directory was removed
(consider changing to repo root: $TESTTMP/r1/r4)
#endif