cargo: When loading wraps multiple packages can have the same URL

This avoids cloning the same repo multiple times, instead a single wrap
can provide multiple cargo dependencies.

3 files changed, 51 insertions, 24 deletions

diff --git a/mesonbuild/cargo/interpreter.py b/mesonbuild/cargo/interpreter.py
index c48efd6c5..82d2b5836 100644
--- a/mesonbuild/cargo/interpreter.py
+++ b/mesonbuild/cargo/interpreter.py
@@ -448,21 +448,41 @@ class Interpreter:
         ]
 
 
+def _parse_git_url(url: str, branch: T.Optional[str] = None) -> T.Tuple[str, str, str]:
+    if url.startswith('git+'):
+        url = url[4:]
+    parts = urllib.parse.urlparse(url)
+    query = urllib.parse.parse_qs(parts.query)
+    query_branch = query['branch'][0] if 'branch' in query else ''
+    branch = branch or query_branch
+    revision = parts.fragment or branch
+    directory = os.path.basename(parts.path)
+    if directory.endswith('.git'):
+        directory = directory[:-4]
+    if branch:
+        directory += f'-{branch}'
+    url = urllib.parse.urlunparse(parts._replace(params='', query='', fragment=''))
+    return url, revision, directory
+
+
 def load_wraps(source_dir: str, subproject_dir: str) -> T.List[PackageDefinition]:
     """ Convert Cargo.lock into a list of wraps """
 
-    wraps: T.List[PackageDefinition] = []
+    # Map directory -> PackageDefinition, to avoid duplicates. Multiple packages
+    # can have the same source URL, in that case we have a single wrap that
+    # provides multiple dependency names.
+    wraps: T.Dict[str, PackageDefinition] = {}
     filename = os.path.join(source_dir, 'Cargo.lock')
     if os.path.exists(filename):
         try:
             toml = load_toml(filename)
         except TomlImplementationMissing as e:
             mlog.warning('Failed to load Cargo.lock:', str(e), fatal=False)
-            return wraps
+            return []
         raw_cargolock = T.cast('raw.CargoLock', toml)
         cargolock = CargoLock.from_raw(raw_cargolock)
         for package in cargolock.package:
-            subp_name = _dependency_name(package.name, version.api(package.version))
+            meson_depname = _dependency_name(package.name, version.api(package.version))
             if package.source is None:
                 # This is project's package, or one of its workspace members.
                 pass
@@ -472,25 +492,24 @@ def load_wraps(source_dir: str, subproject_dir: str) -> T.List[PackageDefinition
                     checksum = cargolock.metadata[f'checksum {package.name} {package.version} ({package.source})']
                 url = f'https://crates.io/api/v1/crates/{package.name}/{package.version}/download'
                 directory = f'{package.name}-{package.version}'
-                wraps.append(PackageDefinition.from_values(subp_name, subproject_dir, 'file', {
-                    'directory': directory,
-                    'source_url': url,
-                    'source_filename': f'{directory}.tar.gz',
-                    'source_hash': checksum,
-                    'method': 'cargo',
-                }))
+                if directory not in wraps:
+                    wraps[directory] = PackageDefinition.from_values(meson_depname, subproject_dir, 'file', {
+                        'directory': directory,
+                        'source_url': url,
+                        'source_filename': f'{directory}.tar.gz',
+                        'source_hash': checksum,
+                        'method': 'cargo',
+                    })
+                wraps[directory].add_provided_dep(meson_depname)
             elif package.source.startswith('git+'):
-                parts = urllib.parse.urlparse(package.source[4:])
-                query = urllib.parse.parse_qs(parts.query)
-                branch = query['branch'][0] if 'branch' in query else ''
-                revision = parts.fragment or branch
-                url = urllib.parse.urlunparse(parts._replace(params='', query='', fragment=''))
-                wraps.append(PackageDefinition.from_values(subp_name, subproject_dir, 'git', {
-                    'directory': package.name,
-                    'url': url,
-                    'revision': revision,
-                    'method': 'cargo',
-                }))
+                url, revision, directory = _parse_git_url(package.source)
+                if directory not in wraps:
+                    wraps[directory] = PackageDefinition.from_values(directory, subproject_dir, 'git', {
+                        'url': url,
+                        'revision': revision,
+                        'method': 'cargo',
+                    })
+                wraps[directory].add_provided_dep(meson_depname)
             else:
                 mlog.warning(f'Unsupported source URL in {filename}: {package.source}')
-    return wraps
+    return list(wraps.values())
diff --git a/mesonbuild/wrap/wrap.py b/mesonbuild/wrap/wrap.py
index f8ed5e0fe..7273e77f4 100644
--- a/mesonbuild/wrap/wrap.py
+++ b/mesonbuild/wrap/wrap.py
@@ -318,6 +318,9 @@ class PackageDefinition:
             with open(self.get_hashfile(subproject_directory), 'w', encoding='utf-8') as file:
                 file.write(self.wrapfile_hash + '\n')
 
+    def add_provided_dep(self, name: str) -> None:
+        self.provided_deps[name] = None
+
 def get_directory(subdir_root: str, packagename: str) -> str:
     fname = os.path.join(subdir_root, packagename + '.wrap')
     if os.path.isfile(fname):
diff --git a/unittests/cargotests.py b/unittests/cargotests.py
index 7c09ab938..96a7b81ee 100644
--- a/unittests/cargotests.py
+++ b/unittests/cargotests.py
@@ -193,6 +193,10 @@ class CargoLockTest(unittest.TestCase):
                     name = "bar"
                     version = "0.1"
                     source = "git+https://github.com/gtk-rs/gtk-rs-core?branch=0.19#23c5599424cc75ec66618891c915d9f490f6e4c2"
+                    [[package]]
+                    name = "member"
+                    version = "0.1"
+                    source = "git+https://github.com/gtk-rs/gtk-rs-core?branch=0.19#23c5599424cc75ec66618891c915d9f490f6e4c2"
                     '''))
             wraps = load_wraps(tmpdir, 'subprojects')
             self.assertEqual(len(wraps), 2)
@@ -202,12 +206,13 @@ class CargoLockTest(unittest.TestCase):
             self.assertEqual(wraps[0].get('method'), 'cargo')
             self.assertEqual(wraps[0].get('source_url'), 'https://crates.io/api/v1/crates/foo/0.1/download')
             self.assertEqual(wraps[0].get('source_hash'), '8a30b2e23b9e17a9f90641c7ab1549cd9b44f296d3ccbf309d2863cfe398a0cb')
-            self.assertEqual(wraps[1].name, 'bar-0.1-rs')
-            self.assertEqual(wraps[1].directory, 'bar')
+            self.assertEqual(wraps[1].name, 'gtk-rs-core-0.19')
+            self.assertEqual(wraps[1].directory, 'gtk-rs-core-0.19')
             self.assertEqual(wraps[1].type, 'git')
             self.assertEqual(wraps[1].get('method'), 'cargo')
             self.assertEqual(wraps[1].get('url'), 'https://github.com/gtk-rs/gtk-rs-core')
             self.assertEqual(wraps[1].get('revision'), '23c5599424cc75ec66618891c915d9f490f6e4c2')
+            self.assertEqual(list(wraps[1].provided_deps), ['gtk-rs-core-0.19', 'bar-0.1-rs', 'member-0.1-rs'])
 
 class CargoTomlTest(unittest.TestCase):
     CARGO_TOML_1 = textwrap.dedent('''\