keyword: specific regular expressions depending on read mode

keyword: specific regular expressions depending on read mode

More safeguarding against accidental (un)expansion:

Reading filelog: act only on \$(kw1|kw2|..)\$ as keywords are always
                 stored unexpanded.
Reading wdir:    act only on \$(kw1|kw2|..): [^$\n\r]*? \$ as we only
                 are interested in expanded keywords in this situation.
                 Note: we cannot use ..): [^$\n\r]+? \$ because e.g.
                 the {branch} template might be empty.

hg record is a special case as we read from the working directory and
need one regex each for modified and added files. Therefore test
recording an added file.

This way we finally also forbid sequences like $Id:  $ being treated
as keywords.

--- a/hgext/keyword.py	Fri Oct 08 18:39:46 2010 +0100
+++ b/hgext/keyword.py	Fri Oct 08 18:39:46 2010 +0100
@@ -163,9 +163,9 @@
                                   for k, v in kwmaps)
         else:
             self.templates = _defaultkwmaps(self.ui)
-        escaped = map(re.escape, self.templates.keys())
-        kwpat = r'\$(%s)(: [^$\n\r]*? )??\$' % '|'.join(escaped)
-        self.re_kw = re.compile(kwpat)
+        escaped = '|'.join(map(re.escape, self.templates.keys()))
+        self.re_kw = re.compile(r'\$(%s)\$' % escaped)
+        self.re_kwexp = re.compile(r'\$(%s): [^$\n\r]*? \$' % escaped)
 
         templatefilters.filters.update({'utcdate': utcdate,
                                         'svnisodate': svnisodate,
@@ -196,7 +196,7 @@
         expansion are not symbolic links.'''
         return [f for f in cand if self.match(f) and not 'l' in ctx.flags(f)]
 
-    def overwrite(self, ctx, candidates, lookup, expand):
+    def overwrite(self, ctx, candidates, lookup, expand, recsubn=None):
         '''Overwrites selected files expanding/shrinking keywords.'''
         if self.restrict or lookup: # exclude kw_copy
             candidates = self.iskwfile(candidates, ctx)
@@ -206,6 +206,8 @@
         if self.restrict or expand and lookup:
             mf = ctx.manifest()
         fctx = ctx
+        subn = (self.restrict and self.re_kw.subn or
+                recsubn or self.re_kwexp.subn)
         msg = (expand and _('overwriting %s expanding keywords\n')
                or _('overwriting %s shrinking keywords\n'))
         for f in candidates:
@@ -218,11 +220,11 @@
             if expand:
                 if lookup:
                     fctx = self.repo.filectx(f, fileid=mf[f]).changectx()
-                data, found = self.substitute(data, f, fctx, self.re_kw.subn)
+                data, found = self.substitute(data, f, fctx, subn)
             elif self.restrict:
                 found = self.re_kw.search(data)
             else:
-                data, found = _shrinktext(data, self.re_kw.subn)
+                data, found = _shrinktext(data, subn)
             if found:
                 self.ui.note(msg % f)
                 self.repo.wwrite(f, data, ctx.flags(f))
@@ -234,7 +236,7 @@
     def shrink(self, fname, text):
         '''Returns text with all keyword substitutions removed.'''
         if self.match(fname) and not util.binary(text):
-            return _shrinktext(text, self.re_kw.sub)
+            return _shrinktext(text, self.re_kwexp.sub)
         return text
 
     def shrinklines(self, fname, lines):
@@ -242,7 +244,7 @@
         if self.match(fname):
             text = ''.join(lines)
             if not util.binary(text):
-                return _shrinktext(text, self.re_kw.sub).splitlines(True)
+                return _shrinktext(text, self.re_kwexp.sub).splitlines(True)
         return lines
 
     def wread(self, fname, data):
@@ -569,12 +571,15 @@
             # record returns 0 even when nothing has changed
             # therefore compare nodes before and after
             ctx = repo['.']
+            modified, added = repo[None].status()[:2]
             ret = orig(ui, repo, commitfunc, *pats, **opts)
-            recordctx = repo['.']
-            if ctx != recordctx:
-                candidates = [f for f in recordctx.files() if f in recordctx]
+            recctx = repo['.']
+            if ctx != recctx:
+                modified = [f for f in modified if f in recctx]
+                added = [f for f in added if f in recctx]
                 kwt.restrict = False
-                kwt.overwrite(recordctx, candidates, False, True)
+                kwt.overwrite(recctx, modified, False, True, kwt.re_kwexp.subn)
+                kwt.overwrite(recctx, added, False, True, kwt.re_kw.subn)
                 kwt.restrict = True
             return ret
         finally:

--- a/tests/test-keyword.t	Fri Oct 08 18:39:46 2010 +0100
+++ b/tests/test-keyword.t	Fri Oct 08 18:39:46 2010 +0100
@@ -373,7 +373,6 @@
 
   $ hg status -A a
   C a
-  $ rm msg
 
 rollback and revert expansion
 
@@ -418,6 +417,25 @@
   $ hg update -C
   1 files updated, 0 files merged, 0 files removed, 0 files unresolved
 
+record added file
+
+  $ echo '$Id$' > r
+  $ hg add r
+  $ hg -v record -l msg -d '1 12' r<<EOF
+  > y
+  > EOF
+  diff --git a/r b/r
+  new file mode 100644
+  examine changes to 'r'? [Ynsfdaq?] 
+  r
+  committed changeset 3:899491280810
+  overwriting r expanding keywords
+  $ hg --verbose rollback
+  rolling back to revision 2 (undo commit)
+  overwriting r shrinking keywords
+  $ hg forget r
+  $ rm msg r
+
 Test patch queue repo
 
   $ hg init --mq