match: handle excludes using new differencematcher

As I've said on earlier patches, I'm hoping to use more composition of
simpler matchers instead of the single complex matcher we currently
have. This extracts a first new matcher that composes two other
matchers. It matches if the first matcher matches but the second does
not. As such, we can use it for excludes, which this patch also
does. We'll remove the now-unncessary code for excludes in the next
patch.

mercurial/match.py

@@ -142,10 +142,15 b' def match(root, cwd, patterns, include=N'
                 kindpats.append((kind, pats, source))
             return kindpats
 
-    return matcher(root, cwd, normalize, patterns, include=include,
-                   exclude=exclude, default=default, exact=exact,
-                   auditor=auditor, ctx=ctx, listsubrepos=listsubrepos,
-                   warn=warn, badfn=badfn)
+    m = matcher(root, cwd, normalize, patterns, include=include, exclude=None,
+                default=default, exact=exact, auditor=auditor, ctx=ctx,
+                listsubrepos=listsubrepos, warn=warn, badfn=badfn)
+    if exclude:
+        em = matcher(root, cwd, normalize, [], include=exclude, exclude=None,
+                     default=default, exact=False, auditor=auditor, ctx=ctx,
+                     listsubrepos=listsubrepos, warn=warn, badfn=None)
+        m = differencematcher(m, em)
+    return m
 
 def exact(root, cwd, files, badfn=None):
     return match(root, cwd, files, exact=True, badfn=badfn)
@@ -418,6 +423,62 b' class matcher(basematcher):'
                 (self._files, self.patternspat, self.includepat,
                  self.excludepat))
 
+class differencematcher(basematcher):
+    '''Composes two matchers by matching if the first matches and the second
+    does not. Well, almost... If the user provides a pattern like "-X foo foo",
+    Mercurial actually does match "foo" against that. That's because exact
+    matches are treated specially. So, since this differencematcher is used for
+    excludes, it needs to special-case exact matching.
+
+    The second matcher's non-matching-attributes (root, cwd, bad, explicitdir,
+    traversedir) are ignored.
+
+    TODO: If we want to keep the behavior described above for exact matches, we
+    should consider instead treating the above case something like this:
+    union(exact(foo), difference(pattern(foo), include(foo)))
+    '''
+    def __init__(self, m1, m2):
+        super(differencematcher, self).__init__(m1._root, m1._cwd)
+        self._m1 = m1
+        self._m2 = m2
+        self.bad = m1.bad
+        self.explicitdir = m1.explicitdir
+        self.traversedir = m1.traversedir
+
+    def matchfn(self, f):
+        return self._m1(f) and (not self._m2(f) or self._m1.exact(f))
+
+    @propertycache
+    def _files(self):
+        if self.isexact():
+            return [f for f in self._m1.files() if self(f)]
+        # If m1 is not an exact matcher, we can't easily figure out the set of
+        # files, because its files() are not always files. For example, if
+        # m1 is "path:dir" and m2 is "rootfileins:.", we don't
+        # want to remove "dir" from the set even though it would match m2,
+        # because the "dir" in m1 may not be a file.
+        return self._m1.files()
+
+    def visitdir(self, dir):
+        if self._m2.visitdir(dir) == 'all':
+            # There's a bug here: If m1 matches file 'dir/file' and m2 excludes
+            # 'dir' (recursively), we should still visit 'dir' due to the
+            # exception we have for exact matches.
+            return False
+        return bool(self._m1.visitdir(dir))
+
+    def isexact(self):
+        return self._m1.isexact()
+
+    def anypats(self):
+        return self._m1.anypats() or self._m2.anypats()
+
+    def prefix(self):
+        return not self.always() and not self.isexact() and not self.anypats()
+
+    def __repr__(self):
+        return ('<differencematcher m1=%r, m2=%r>' % (self._m1, self._m2))
+
 class subdirmatcher(basematcher):
     """Adapt a matcher to work on a subdirectory only.
 

tests/test-walk.t

@@ -76,7 +76,7 b''
   f  mammals/Procyonidae/raccoon     Procyonidae/raccoon
   f  mammals/skunk                   skunk
   $ hg debugwalk -X ../beans
-  matcher: <matcher files=[], patterns=None, includes=None, excludes='(?:beans(?:/|$))'>
+  matcher: <differencematcher m1=<matcher files=[], patterns=None, includes=None, excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:beans(?:/|$))', excludes=None>>
   f  fennel                          ../fennel
   f  fenugreek                       ../fenugreek
   f  fiddlehead                      ../fiddlehead
@@ -146,7 +146,7 b''
   f  fenugreek   ../fenugreek
   f  fiddlehead  ../fiddlehead
   $ hg debugwalk -X 'rootfilesin:'
-  matcher: <matcher files=[], patterns=None, includes=None, excludes='(?:^[^/]+$)'>
+  matcher: <differencematcher m1=<matcher files=[], patterns=None, includes=None, excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:^[^/]+$)', excludes=None>>
   f  beans/black                     ../beans/black
   f  beans/borlotti                  ../beans/borlotti
   f  beans/kidney                    ../beans/kidney
@@ -194,7 +194,7 b''
   matcher: <matcher files=[], patterns=None, includes='(?:^mammals/[^/]+$)', excludes=None>
   f  mammals/skunk  skunk
   $ hg debugwalk -X 'rootfilesin:mammals'
-  matcher: <matcher files=[], patterns=None, includes=None, excludes='(?:^mammals/[^/]+$)'>
+  matcher: <differencematcher m1=<matcher files=[], patterns=None, includes=None, excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:^mammals/[^/]+$)', excludes=None>>
   f  beans/black                     ../beans/black
   f  beans/borlotti                  ../beans/borlotti
   f  beans/kidney                    ../beans/kidney
@@ -289,35 +289,35 b''
   matcher: <matcher files=['beans'], patterns='(?:beans(?:/|$))', includes='(?:beans\\/black(?:/|$))', excludes=None>
   f  beans/black  beans/black
   $ hg debugwalk -Xbeans/black beans
-  matcher: <matcher files=['beans'], patterns='(?:beans(?:/|$))', includes=None, excludes='(?:beans\\/black(?:/|$))'>
+  matcher: <differencematcher m1=<matcher files=['beans'], patterns='(?:beans(?:/|$))', includes=None, excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:beans\\/black(?:/|$))', excludes=None>>
   f  beans/borlotti  beans/borlotti
   f  beans/kidney    beans/kidney
   f  beans/navy      beans/navy
   f  beans/pinto     beans/pinto
   f  beans/turtle    beans/turtle
   $ hg debugwalk -Xbeans/black -Ibeans
-  matcher: <matcher files=[], patterns=None, includes='(?:beans(?:/|$))', excludes='(?:beans\\/black(?:/|$))'>
+  matcher: <differencematcher m1=<matcher files=[], patterns=None, includes='(?:beans(?:/|$))', excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:beans\\/black(?:/|$))', excludes=None>>
   f  beans/borlotti  beans/borlotti
   f  beans/kidney    beans/kidney
   f  beans/navy      beans/navy
   f  beans/pinto     beans/pinto
   f  beans/turtle    beans/turtle
   $ hg debugwalk -Xbeans/black beans/black
-  matcher: <matcher files=['beans/black'], patterns='(?:beans\\/black(?:/|$))', includes=None, excludes='(?:beans\\/black(?:/|$))'>
+  matcher: <differencematcher m1=<matcher files=['beans/black'], patterns='(?:beans\\/black(?:/|$))', includes=None, excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:beans\\/black(?:/|$))', excludes=None>>
   f  beans/black  beans/black  exact
   $ hg debugwalk -Xbeans/black -Ibeans/black
-  matcher: <matcher files=[], patterns=None, includes='(?:beans\\/black(?:/|$))', excludes='(?:beans\\/black(?:/|$))'>
+  matcher: <differencematcher m1=<matcher files=[], patterns=None, includes='(?:beans\\/black(?:/|$))', excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:beans\\/black(?:/|$))', excludes=None>>
   $ hg debugwalk -Xbeans beans/black
-  matcher: <matcher files=['beans/black'], patterns='(?:beans\\/black(?:/|$))', includes=None, excludes='(?:beans(?:/|$))'>
+  matcher: <differencematcher m1=<matcher files=['beans/black'], patterns='(?:beans\\/black(?:/|$))', includes=None, excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:beans(?:/|$))', excludes=None>>
   f  beans/black  beans/black  exact
   $ hg debugwalk -Xbeans -Ibeans/black
-  matcher: <matcher files=[], patterns=None, includes='(?:beans\\/black(?:/|$))', excludes='(?:beans(?:/|$))'>
+  matcher: <differencematcher m1=<matcher files=[], patterns=None, includes='(?:beans\\/black(?:/|$))', excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:beans(?:/|$))', excludes=None>>
   $ hg debugwalk 'glob:mammals/../beans/b*'
   matcher: <matcher files=['beans'], patterns='(?:beans\\/b[^/]*$)', includes=None, excludes=None>
   f  beans/black     beans/black
   f  beans/borlotti  beans/borlotti
   $ hg debugwalk '-X*/Procyonidae' mammals
-  matcher: <matcher files=['mammals'], patterns='(?:mammals(?:/|$))', includes=None, excludes='(?:[^/]*\\/Procyonidae(?:/|$))'>
+  matcher: <differencematcher m1=<matcher files=['mammals'], patterns='(?:mammals(?:/|$))', includes=None, excludes=None>, m2=<matcher files=[], patterns=None, includes='(?:[^/]*\\/Procyonidae(?:/|$))', excludes=None>>
   f  mammals/skunk  mammals/skunk
   $ hg debugwalk path:mammals
   matcher: <matcher files=['mammals'], patterns='(?:^mammals(?:/|$))', includes=None, excludes=None>