last_change: Use git log --name-status to avoid repeated fork+exec

Signed-off-by: Anders Kaseorg <andersk@mit.edu>

tracext/git/PyGIT.py

@@ -14,6 +14,7 @@
 from threading import Lock
 from subprocess import Popen, PIPE
 from operator import itemgetter
+from contextlib import contextmanager
 import cStringIO
 import codecs
 
@@ -70,8 +71,11 @@ def __execute(self, git_cmd, *cmd_args):
     def cat_file_batch(self):
         return self.__pipe('cat-file', '--batch', stdin=PIPE, stdout=PIPE)
 
+    def log_pipe(self, *cmd_args):
+        return self.__pipe('log', *cmd_args, stdout=PIPE)
+
     def __getattr__(self, name):
-        if name[0] == '_' or name in ['cat_file_batch']:
+        if name[0] == '_' or name in ['cat_file_batch', 'log_pipe']:
             raise AttributeError, name
         return partial(self.__execute, name.replace('_','-'))
 
@@ -713,7 +717,52 @@ def sync(self):
         rev = self.repo.rev_list("--max-count=1", "--topo-order", "--all").strip()
         return self.__rev_cache_sync(rev)
 
-    def last_change(self, sha, path):
+    @contextmanager
+    def get_historian(self, sha, base_path):
+        p = []
+        change = {}
+        next_path = []
+
+        def name_status_gen():
+            p[:] = [self.repo.log_pipe('--pretty=format:%n%H', '--name-status',
+                                       sha, '--', base_path)]
+            f = p[0].stdout
+            for l in f:
+                if l == '\n': continue
+                old_sha = l.rstrip('\n')
+                for l in f:
+                    if l == '\n': break
+                    _, path = l.rstrip('\n').split('\t', 1)
+                    while path not in change:
+                        change[path] = old_sha
+                        if next_path == [path]: yield old_sha
+                        try:
+                            path, _ = path.rsplit('/', 1)
+                        except ValueError:
+                            break
+            f.close()
+            p[0].terminate()
+            p[0].wait()
+            p[:] = []
+            while True: yield None
+        gen = name_status_gen()
+
+        def historian(path):
+            try:
+                return change[path]
+            except KeyError:
+                next_path[:] = [path]
+                return gen.next()
+        yield historian
+
+        if p:
+            p[0].stdout.close()
+            p[0].terminate()
+            p[0].wait()
+
+    def last_change(self, sha, path, historian=None):
+        if historian is not None:
+            return historian(path)
         return self.repo.rev_list("--max-count=1",
                                   sha, "--",
                                   self._fs_from_unicode(path)).strip() or None

tracext/git/git_fs.py

@@ -387,8 +387,8 @@ def display_rev(self, rev):
     def short_rev(self, rev):
         return self.git.shortrev(self.normalize_rev(rev), min_len=self._shortrev_len)
 
-    def get_node(self, path, rev=None):
-        return GitNode(self, path, rev, self.log)
+    def get_node(self, path, rev=None, historian=None):
+        return GitNode(self, path, rev, self.log, None, historian)
 
     def get_quickjump_entries(self, rev):
         for bname, bsha in self.git.get_branches():
@@ -412,24 +412,26 @@ def get_changes(self, old_path, old_rev, new_path, new_rev, ignore_ancestry=0):
         if old_path != new_path:
             raise TracError("not supported in git_fs")
 
-        for chg in self.git.diff_tree(old_rev, new_rev, self.normalize_path(new_path)):
-            mode1, mode2, obj1, obj2, action, path, path2 = chg
+        with self.git.get_historian(old_rev, old_path.strip('/')) as old_historian:
+            with self.git.get_historian(new_rev, new_path.strip('/')) as new_historian:
+                for chg in self.git.diff_tree(old_rev, new_rev, self.normalize_path(new_path)):
+                    mode1, mode2, obj1, obj2, action, path, path2 = chg
 
-            kind = Node.FILE
-            if mode2.startswith('04') or mode1.startswith('04'):
-                kind = Node.DIRECTORY
+                    kind = Node.FILE
+                    if mode2.startswith('04') or mode1.startswith('04'):
+                        kind = Node.DIRECTORY
 
-            change = GitChangeset.action_map[action]
+                    change = GitChangeset.action_map[action]
 
-            old_node = None
-            new_node = None
+                    old_node = None
+                    new_node = None
 
-            if change != Changeset.ADD:
-                old_node = self.get_node(path, old_rev)
-            if change != Changeset.DELETE:
-                new_node = self.get_node(path, new_rev)
+                    if change != Changeset.ADD:
+                        old_node = self.get_node(path, old_rev, old_historian)
+                    if change != Changeset.DELETE:
+                        new_node = self.get_node(path, new_rev, new_historian)
 
-            yield old_node, new_node, kind, change
+                    yield old_node, new_node, kind, change
 
     def next_rev(self, rev, path=''):
         return self.git.hist_next_revision(rev)
@@ -469,7 +471,7 @@ def sync(self, rev_callback=None, clean=None):
                 rev_callback(rev)
 
 class GitNode(Node):
-    def __init__(self, repos, path, rev, log, ls_tree_info=None):
+    def __init__(self, repos, path, rev, log, ls_tree_info=None, historian=None):
         self.log = log
         self.repos = repos
         self.fs_sha = None # points to either tree or blobs
@@ -491,7 +493,7 @@ def __init__(self, repos, path, rev, log, ls_tree_info=None):
             self.fs_perm, k, self.fs_sha, self.fs_size, _ = ls_tree_info
 
             # fix-up to the last commit-rev that touched this node
-            rev = repos.git.last_change(rev, p)
+            rev = repos.git.last_change(rev, p, historian)
 
             if k == 'tree':
                 pass
@@ -537,8 +539,9 @@ def get_entries(self):
         if not self.isdir:
             return
 
-        for ent in self.repos.git.ls_tree(self.rev, self.__git_path()):
-            yield GitNode(self.repos, ent[-1], self.rev, self.log, ent)
+        with self.repos.git.get_historian(self.rev, self.path.strip('/')) as historian:
+            for ent in self.repos.git.ls_tree(self.rev, self.__git_path()):
+                yield GitNode(self.repos, ent[-1], self.rev, self.log, ent, historian)
 
     def get_content_type(self):
         if self.isdir: