dirstate-v2: add test that shows a collision in ignore patterns hash

This hash is used for optimizing dirstate `status`. We demonstrate that the
hash is incorrectly ignoring the changes to the semantics of the ignore
files just because the contents (but not their source) haven't changed.

This is fixed in the next changeset.

tests/test-hgignore.t

@@ -433,5 +433,33 b' This is an optimization that is only rel'
   sha1=dea19cc7119213f24b6b582a4bae7b0cb063e34e
   $ hg debugstate --docket | grep ignore
   ignore pattern hash: dea19cc7119213f24b6b582a4bae7b0cb063e34e
+  $ cd ..
+
+Check that the hash depends on the source of the hgignore patterns
+(otherwise the context is lost and things like subinclude are cached improperly)
+
+  $ hg init ignore-collision
+  $ cd ignore-collision
+  $ echo > .hg/testhgignorerel
+
+  $ mkdir dir1/ dir1/subdir
+  $ touch dir1/subdir/f dir1/subdir/ignored1
+  $ echo 'ignored1' > dir1/.hgignore
+
+  $ mkdir dir2 dir2/subdir
+  $ touch dir2/subdir/f dir2/subdir/ignored2
+  $ echo 'ignored2' > dir2/.hgignore
+  $ echo 'subinclude:dir2/.hgignore' >> .hgignore
+  $ echo 'subinclude:dir1/.hgignore' >> .hgignore
+
+  $ hg commit -Aqm_
+
+  $ > dir1/.hgignore
+  $ echo 'ignored' > dir2/.hgignore
+  $ echo 'ignored1' >> dir2/.hgignore
+  $ hg status
+  M dir1/.hgignore
+  M dir2/.hgignore
+  ? dir1/subdir/ignored1 (missing-correct-output !)
 
 #endif