Only use pygit2's blame via opt-in

This can be reconsidered if pygit2's blame algorithm improves.
The option can still be useful if the 'git' command is not available
on the system.

5 files changed, 107 insertions, 4 deletions

diff --git a/git_deps/blame.py b/git_deps/blame.py
new file mode 100644
index 0000000..41d79fe
--- /dev/null
+++ b/git_deps/blame.py
@@ -0,0 +1,61 @@
+import subprocess
+import re
+from dataclasses import dataclass
+
+# The following classes are introduced to imitate their counterparts in pygit2,
+# so that the output of 'blame_via_subprocess' can be swapped with pygit2's own
+# blame output.
+
+@dataclass
+class GitRef:
+    """
+    A reference to a commit
+    """
+    hex: str
+
+@dataclass
+class BlameHunk:
+    """
+    A chunk of a blame output which has the same commit information
+    for a consecutive set of lines
+    """
+    orig_commit_id: GitRef
+    orig_start_line_number: int
+    final_start_line_number: int
+    lines_in_hunk: int = 1
+
+
+def blame_via_subprocess(path, commit, start_line, num_lines):
+    """
+    Generate a list of blame hunks by calling 'git blame' as a separate process.
+    This is a workaround for the slowness of pygit2's own blame algorithm.
+    See https://github.com/aspiers/git-deps/issues/1
+    """
+    cmd = [
+        'git', 'blame',
+        '--porcelain',
+        '-L', "%d,+%d" % (start_line, num_lines),
+        commit, '--', path
+    ]
+    output = subprocess.check_output(cmd, universal_newlines=True)
+
+    current_hunk = None
+    for line in output.split('\n'):
+        m = re.match(r'^([0-9a-f]{40}) (\d+) (\d+) (\d+)$', line)
+
+        if m: # starting a new hunk
+            if current_hunk:
+                yield current_hunk
+            dependency_sha1, orig_line_num, line_num, length = m.group(1, 2, 3, 4)
+            orig_line_num = int(orig_line_num)
+            line_num = int(line_num)
+            length = int(length)
+            current_hunk = BlameHunk(
+                orig_commit_id=GitRef(dependency_sha1),
+                orig_start_line_number = orig_line_num,
+                final_start_line_number = line_num,
+                lines_in_hunk = length
+            )
+
+    if current_hunk:
+        yield current_hunk
diff --git a/git_deps/cli.py b/git_deps/cli.py
index 34a45e1..701b004 100755
--- a/git_deps/cli.py
+++ b/git_deps/cli.py
@@ -75,6 +75,8 @@ def parse_args():
                         '[%(default)s]')
     parser.add_argument('-d', '--debug', dest='debug', action='store_true',
                         help='Show debugging')
+    parser.add_argument('--pygit2-blame', dest='pygit2_blame', action='store_true',
+                        help="Use pygit2's blame algorithm (slower than git's)")
 
     options, args = parser.parse_known_args()
 
diff --git a/git_deps/detector.py b/git_deps/detector.py
index a7036c9..82996a6 100644
--- a/git_deps/detector.py
+++ b/git_deps/detector.py
@@ -7,6 +7,7 @@ from git_deps.utils import abort, standard_logger
 from git_deps.gitutils import GitUtils
 from git_deps.listener.base import DependencyListener
 from git_deps.errors import InvalidCommitish
+from git_deps.blame import blame_via_subprocess
 
 
 class DependencyDetector(object):
@@ -236,10 +237,16 @@ class DependencyDetector(object):
             self.notify_listeners("new_dependent", dependent)
 
     def run_blame(self, hunk, parent, path):
-        return self.repo.blame(path,
-                    newest_commit=parent.hex,
-                    min_line=hunk.old_start,
-                    max_line=hunk.old_start + hunk.old_lines - 1)
+        if self.options.pygit2_blame:
+            return self.repo.blame(path,
+                        newest_commit=parent.hex,
+                        min_line=hunk.old_start,
+                        max_line=hunk.old_start + hunk.old_lines - 1)
+        else:
+            return blame_via_subprocess(path,
+                        parent.hex,
+                        hunk.old_start,
+                        hunk.old_lines)
 
     def is_excluded(self, commit):
         if self.options.exclude_commits is not None:
diff --git a/tests/self_test.sh b/tests/self_test.sh
index a0b2272..52e1a28 100755
--- a/tests/self_test.sh
+++ b/tests/self_test.sh
@@ -13,12 +13,21 @@ echo "Running test suite"
 echo "* Dependencies of 4f27a1e, a regular commit"
 git-deps 4f27a1e^! | sort | diff tests/expected_outputs/deps_4f27a1e -
 
+echo "* Same, but via pygit2's blame algorithm"
+git-deps --pygit2-blame 4f27a1e^! | sort | diff tests/expected_outputs/deps_4f27a1e -
+
 echo "* Dependencies of 1ba7ad5, a merge commit"
 git-deps 1ba7ad5^! | sort | diff tests/expected_outputs/deps_1ba7ad5 -
 
+echo "* Same, but via pygit2's blame algorithm"
+git-deps --pygit2-blame 1ba7ad5^! | sort | diff tests/expected_outputs/deps_1ba7ad5 -
+
 echo "* Dependencies of the root commit"
 git-deps b196757^! | sort | diff tests/expected_outputs/deps_b196757 -
 
+echo "* Same, but via pygit2's blame algorithm"
+git-deps --pygit2-blame b196757^! | sort | diff tests/expected_outputs/deps_b196757 -
+
 echo "* Recursive dependencies of a4f27a1e, a regular commit"
 git-deps -r 4f27a1e^! | sort | diff tests/expected_outputs/recursive_deps_4f27a1e -
 
diff --git a/tests/test_blame.py b/tests/test_blame.py
new file mode 100644
index 0000000..410e7de
--- /dev/null
+++ b/tests/test_blame.py
@@ -0,0 +1,24 @@
+
+from git_deps.blame import blame_via_subprocess, BlameHunk, GitRef
+
+def test_blame_via_subprocess():
+    hunks = list(blame_via_subprocess(
+        'INSTALL.md',
+        '04f5c095d4eccf5808db6dbf90c31a535f7f371c',
+        12, 4))
+
+    expected_hunks = [
+        BlameHunk(
+            GitRef('6e23a48f888a355ad7e101c797ce1b66c4b7b86a'),
+            orig_start_line_number=12,
+            final_start_line_number=12,
+            lines_in_hunk=2),
+        BlameHunk(
+            GitRef('2c9d23b0291157eb1096384ff76e0122747b9bdf'),
+            orig_start_line_number=10,
+            final_start_line_number=14,
+            lines_in_hunk=2)
+    ]
+
+    assert hunks == expected_hunks
+