path_auditor: cache names of audited directories
We use a separate cache to avoid problems with
audit = path_auditor(repo.root)
audit("subrepo")
audit("subrepo/file")
whitelisting "subrepo" (which is fine) and then using the same whitelist
with "subrepo/file" (which is not fine).
Since we create a separate path_auditor for every path on the command line,
a "hg add dir/a dir/b dir/c" will still lstat dir 3 times just to audit
the paths.
#!/bin/sh
# http://www.selenic.com/mercurial/bts/issue322
echo % file replaced with directory
hg init a
cd a
echo a > a
hg commit -Ama
rm a
mkdir a
echo a > a/a
echo % should fail - would corrupt dirstate
hg add a/a
cd ..
echo % directory replaced with file
hg init c
cd c
mkdir a
echo a > a/a
hg commit -Ama
rm -r a
echo a > a
echo % should fail - would corrupt dirstate
hg add a
cd ..
echo % directory replaced with file
hg init d
cd d
mkdir b
mkdir b/c
echo a > b/c/d
hg commit -Ama
rm -r b
echo a > b
echo % should fail - would corrupt dirstate
hg add b
exit 0