From fcc6bf93f7b336919220b77c5e7402deacabe60d Mon Sep 17 00:00:00 2001
From: Lukas Zachar <lzachar@redhat.com>
Date: Fri, 15 Jul 2022 18:51:27 +0200
Subject: [PATCH] Shallow git clone if no reference is used

If later `fetch_repo()` specifies ref,
git fetch --unshallow is called to obtain whole history.

Fixes: #169
---
 fmf/utils.py             | 24 +++++++++++++++++++++---
 tests/unit/test_utils.py |  9 +++++++++
 2 files changed, 30 insertions(+), 3 deletions(-)

diff --git a/fmf/utils.py b/fmf/utils.py
index 598c9d09..4318d828 100644
--- a/fmf/utils.py
+++ b/fmf/utils.py
@@ -720,9 +720,13 @@ def fetch_repo(url, ref=None, destination=None, env=None):
             # Write PID to lockfile so we know which process got it
             with open(lock.lock_file, 'w') as lock_file:
                 lock_file.write(str(os.getpid()))
-            # Clone the repository
+            # Clone the repository, use depth=1 when ref is not used
+            if ref is None:
+                depth = ['--depth=1']
+            else:
+                depth = []
             if not os.path.isdir(os.path.join(destination, '.git')):
-                run(['git', 'clone', url, destination], cwd=cache, env=env)
+                run(['git', 'clone'] + depth + [url, destination], cwd=cache, env=env)
             # Detect the default branch if 'ref' not provided
             if ref is None:
                 ref = default_branch(destination)
@@ -735,7 +739,21 @@ def fetch_repo(url, ref=None, destination=None, env=None):
             if age >= CACHE_EXPIRATION:
                 run(['git', 'fetch'], cwd=destination)
             # Checkout branch
-            run(['git', 'checkout', '-f', ref], cwd=destination, env=env)
+            try:
+                run(['git', 'checkout', '-f', ref], cwd=destination, env=env)
+            except subprocess.CalledProcessError as error:
+                # Another way to check for shallow clone is to call
+                # `git rev-parse --is-shallow-repository` but it prints true/false
+                if os.path.isfile(os.path.join(destination, '.git', 'shallow')):
+                    # Make fetch get all remote refs (branches...)
+                    run(["git", "config", "remote.origin.fetch",
+                        "+refs/heads/*:refs/remotes/origin/*"], cwd=destination)
+                    # Fetch the whole history
+                    run(['git', 'fetch', '--unshallow'], cwd=destination)
+                    run(['git', 'checkout', '-f', ref], cwd=destination, env=env)
+                else:
+                    # History is already fetched so ref is wrong (no need to retry)
+                    raise error
             # Reset to origin to get possible changes but no exit code check
             # ref could be tag or commit where it is expected to fail
             run(['git', 'reset', '--hard', "origin/{0}".format(ref)],
diff --git a/tests/unit/test_utils.py b/tests/unit/test_utils.py
index 05f1dcd6..c5c079c4 100644
--- a/tests/unit/test_utils.py
+++ b/tests/unit/test_utils.py
@@ -215,6 +215,15 @@ def test_switch_branches(self):
         output, _ = run(['git', 'rev-parse', '--abbrev-ref', 'HEAD'], repo)
         assert 'main' in output
 
+    def test_shallow_clone(self, tmpdir):
+        fetched_to = utils.fetch_repo(GIT_REPO, destination=str(tmpdir))
+        out, _ = run(["git", "rev-parse", "--is-shallow-repository"], fetched_to)
+        assert 'true' in out
+
+        fetched_to = utils.fetch_repo(GIT_REPO, destination=str(tmpdir), ref='fedora')
+        out, _ = run(["git", "rev-parse", "--is-shallow-repository"], fetched_to)
+        assert 'false' in out
+
     def test_fetch_valid_id(self):
         repo = utils.fetch_repo(GIT_REPO, '0.10')
         assert utils.os.path.isfile(utils.os.path.join(repo, 'fmf.spec'))