gitdir: Add methods to inspect contents

This adds methods to list and read files. The difference between doing
this to a commit and the currrent working tree is abstracted over by
whether the passed ref is None or omitted.

2 files changed, 142 insertions, 0 deletions

diff --git a/morphlib/gitdir.py b/morphlib/gitdir.py
index f40190ff..cb247303 100644
--- a/morphlib/gitdir.py
+++ b/morphlib/gitdir.py
@@ -17,10 +17,27 @@
 
 
 import cliapp
+import glob
+import os
 
 import morphlib
 
 
+class NoWorkingTreeError(cliapp.AppException):
+
+    def __init__(self, repo):
+        cliapp.AppException.__init__(
+            self, 'Git directory %s has no working tree '
+                  '(is bare).' % repo.dirname)
+
+
+class InvalidRefError(cliapp.AppException):
+    def __init__(self, repo, ref):
+        cliapp.AppException.__init__(
+            self, 'Git directory %s has no commit '
+                  'at ref %s.' %(repo.dirname, ref))
+
+
 class GitDirectory(object):
 
     '''Represent a git working tree + .git directory.
@@ -126,6 +143,58 @@ class GitDirectory(object):
         '''Run "git remote update --prune".'''
         self._runcmd(['git', 'remote', 'update', '--prune'])
 
+    def is_bare(self):
+        '''Determine whether the repository has no work tree (is bare)'''
+        return self.get_config('core.bare') == 'true'
+
+    def list_files(self, ref=None):
+        '''Return an iterable of the files in the repository.
+
+        If `ref` is specified, list files at that ref, otherwise
+        use the working tree.
+
+        If this is a bare repository and no ref is specified, raises
+        an exception.
+
+        '''
+        if ref is None and self.is_bare():
+            raise NoWorkingTreeError(self)
+        if ref is None:
+            return self._list_files_in_work_tree()
+        else:
+            return self._list_files_in_ref(ref)
+
+    def _rev_parse_tree(self, ref):
+        try:
+            return self._runcmd(['git', 'rev-parse', '--verify',
+                                 '%s^{tree}' % ref]).strip()
+        except cliapp.AppException as e:
+            raise InvalidRefError(self, ref)
+
+    def _list_files_in_work_tree(self):
+        for dirpath, subdirs, filenames in os.walk(self.dirname):
+            if dirpath == self.dirname and '.git' in subdirs:
+                subdirs.remove('.git')
+            for filename in filenames:
+                yield os.path.join(dirpath, filename)[len(self.dirname)+1:]
+
+    def _list_files_in_ref(self, ref):
+        tree = self._rev_parse_tree(ref)
+        output = self._runcmd(['git', 'ls-tree', '--name-only', '-rz', tree])
+        # ls-tree appends \0 instead of interspersing, so we need to
+        # strip the trailing \0 before splitting
+        paths = output.strip('\0').split('\0')
+        return paths
+
+    def read_file(self, filename, ref=None):
+        if ref is None and self.is_bare():
+            raise NoWorkingTreeError(self)
+        if ref is None:
+            with open(os.path.join(self.dirname, filename)) as f:
+                return f.read()
+        tree = self._rev_parse_tree(ref)
+        return self.cat_file('blob', tree, filename)
+
 
 def init(dirname):
     '''Initialise a new git repository.'''
diff --git a/morphlib/gitdir_tests.py b/morphlib/gitdir_tests.py
index 2494981a..175b8ee7 100644
--- a/morphlib/gitdir_tests.py
+++ b/morphlib/gitdir_tests.py
@@ -64,3 +64,76 @@ class GitDirectoryTests(unittest.TestCase):
         gitdir.set_remote_fetch_url('origin', url)
         self.assertEqual(gitdir.get_remote_fetch_url('origin'), url)
 
+class GitDirectoryContentsTests(unittest.TestCase):
+
+    def setUp(self):
+        self.tempdir = tempfile.mkdtemp()
+        self.dirname = os.path.join(self.tempdir, 'foo')
+        os.mkdir(self.dirname)
+        gd = morphlib.gitdir.init(self.dirname)
+        for fn in ('foo', 'bar.morph', 'baz.morph', 'quux'):
+            with open(os.path.join(self.dirname, fn), "w") as f:
+                f.write('dummy morphology text')
+        gd._runcmd(['git', 'add', '.'])
+        gd._runcmd(['git', 'commit', '-m', 'Initial commit'])
+        os.rename(os.path.join(self.dirname, 'foo'),
+                  os.path.join(self.dirname, 'foo.morph'))
+        self.mirror = os.path.join(self.tempdir, 'mirror')
+        gd._runcmd(['git', 'clone', '--mirror', self.dirname, self.mirror])
+
+    def tearDown(self):
+        shutil.rmtree(self.tempdir)
+
+    def test_lists_files_in_work_tree(self):
+        gd = morphlib.gitdir.GitDirectory(self.dirname)
+        self.assertEqual(sorted(gd.list_files()),
+                         ['bar.morph', 'baz.morph', 'foo.morph', 'quux'])
+
+    def test_read_file_in_work_tree(self):
+        gd = morphlib.gitdir.GitDirectory(self.dirname)
+        self.assertEqual(gd.read_file('bar.morph'),
+                         'dummy morphology text')
+
+    def test_list_raises_no_ref_no_work_tree(self):
+        gd = morphlib.gitdir.GitDirectory(self.mirror)
+        self.assertRaises(morphlib.gitdir.NoWorkingTreeError,
+                          gd.list_files)
+
+    def test_read_raises_no_ref_no_work_tree(self):
+        gd = morphlib.gitdir.GitDirectory(self.mirror)
+        self.assertRaises(morphlib.gitdir.NoWorkingTreeError,
+                          gd.read_file, 'bar.morph')
+
+    def test_lists_files_in_HEAD(self):
+        for gitdir in (self.dirname, self.mirror):
+            gd = morphlib.gitdir.GitDirectory(gitdir)
+            self.assertEqual(sorted(gd.list_files('HEAD')),
+                             ['bar.morph', 'baz.morph', 'foo', 'quux'])
+
+    def test_read_files_in_HEAD(self):
+        for gitdir in (self.dirname, self.mirror):
+            gd = morphlib.gitdir.GitDirectory(gitdir)
+            self.assertEqual(gd.read_file('bar.morph', 'HEAD'),
+                             'dummy morphology text')
+
+    def test_lists_files_in_named_ref(self):
+        for gitdir in (self.dirname, self.mirror):
+            gd = morphlib.gitdir.GitDirectory(gitdir)
+            self.assertEqual(sorted(gd.list_files('master')),
+                             ['bar.morph', 'baz.morph', 'foo', 'quux'])
+
+    def test_read_file_in_named_ref(self):
+        for gitdir in (self.dirname, self.mirror):
+            gd = morphlib.gitdir.GitDirectory(gitdir)
+            self.assertEqual(gd.read_file('bar.morph', 'master'),
+                             'dummy morphology text')
+
+    def test_list_raises_invalid_ref(self):
+        gd = morphlib.gitdir.GitDirectory(self.dirname)
+        self.assertRaises(morphlib.gitdir.InvalidRefError,
+                          gd.list_files, 'no-such-ref')
+
+    def test_read_raises_invalid_ref(self):
+        gd = morphlib.gitdir.GitDirectory(self.dirname)
+        self.assertRaises(morphlib.gitdir.InvalidRefError,
+                          gd.read_file, 'bar', 'no-such-ref')