feat,tests(svn, hg): Parsing updates and tests (#381)

Follow up to #374, fixes #382

Subversion, Mercurial parsing:
- Fixes
- doctests
- Add pytest modules

Author: tony

libvcs/conftest.py

@@ -12,6 +12,8 @@
 
 from libvcs._internal.run import run, which
 from libvcs.projects.git import GitProject, GitRemote
+from libvcs.projects.hg import MercurialProject
+from libvcs.projects.svn import SubversionProject
 
 skip_if_git_missing = pytest.mark.skipif(
     not which("git"), reason="git is not available"
@@ -333,6 +335,32 @@ def git_repo(projects_path: pathlib.Path, git_remote_repo: pathlib.Path):
     return git_repo
 
 
+@pytest.fixture
+def hg_repo(
+    projects_path: pathlib.Path, hg_remote_repo: pathlib.Path
+) -> MercurialProject:
+    """Pre-made hg clone of remote repo checked out to user's projects dir."""
+    hg_repo = MercurialProject(
+        url=f"file://{hg_remote_repo}",
+        dir=str(projects_path / "hg_repo"),
+    )
+    hg_repo.obtain()
+    return hg_repo
+
+
+@pytest.fixture
+def svn_repo(
+    projects_path: pathlib.Path, svn_remote_repo: pathlib.Path
+) -> SubversionProject:
+    """Pre-made svn clone of remote repo checked out to user's projects dir."""
+    svn_repo = SubversionProject(
+        url=f"file://{svn_remote_repo}",
+        dir=str(projects_path / "svn_repo"),
+    )
+    svn_repo.obtain()
+    return svn_repo
+
+
 @pytest.fixture(autouse=True)
 def add_doctest_fixtures(
     doctest_namespace: dict[str, Any],

libvcs/parse/hg.py

@@ -28,9 +28,10 @@
 RE_PATH = r"""
     ((?P<user>.*)@)?
     (?P<hostname>([^/:]+))
-    (?P<separator>[:,/])?
+    (:(?P<port>\d{1,4}))?
+    (?P<separator>/)?
     (?P<path>
-      (\w[^:.]*)  # cut the path at . to negate .hg
+      /?(\w[^:.]*)
     )?
 """
 
@@ -151,9 +152,9 @@ class HgURL(URLProtocol, SkipDefaultFieldsReprMixin):
     >>> HgURL(url='ssh://username@machinename/path/to/repo')
     HgURL(url=ssh://username@machinename/path/to/repo,
             scheme=ssh,
+            user=username,
             hostname=machinename,
             path=path/to/repo,
-            user=username,
             matcher=core-hg)
 
     - Compatibility checking: :meth:`HgURL.is_valid()`
@@ -162,9 +163,11 @@ class HgURL(URLProtocol, SkipDefaultFieldsReprMixin):
 
     url: str
     scheme: Optional[str] = None
-    hostname: Optional[str] = None
-    path: Optional[str] = None
     user: Optional[str] = None
+    hostname: str = dataclasses.field(default="")
+    port: Optional[int] = None
+    separator: str = dataclasses.field(default="/")
+    path: str = dataclasses.field(default="")
 
     #
     # commit-ish: tag, branch, ref, revision
@@ -233,23 +236,63 @@ def to_url(self) -> str:
         >>> hg_location.to_url()
         'https://hg.mozilla.org/mobile-browser'
 
-        Switch them to hglab:
+        Switch them to localhost:
 
         >>> hg_location.hostname = 'localhost'
         >>> hg_location.scheme = 'http'
 
         >>> hg_location.to_url()
         'http://localhost/mobile-browser'
 
-        todo
-        ----
+        Another example, `hugin <http://hugin.hg.sourceforge.net>`_:
+
+        >>> hugin = HgURL(url="http://hugin.hg.sourceforge.net:8000/hgroot/hugin/hugin")
+
+        >>> hugin
+        HgURL(url=http://hugin.hg.sourceforge.net:8000/hgroot/hugin/hugin,
+                scheme=http,
+                hostname=hugin.hg.sourceforge.net,
+                port=8000,
+                path=hgroot/hugin/hugin,
+                matcher=core-hg)
+
+        >>> hugin.to_url()
+        'http://hugin.hg.sourceforge.net:8000/hgroot/hugin/hugin'
+
+        SSH URL with a username, `graphicsmagic <http://graphicsmagick.org/Hg.html>`_:
+
+        >>> graphicsmagick = HgURL(
+        ...     url="ssh://yourid@hg.GraphicsMagick.org//hg/GraphicsMagick"
+        ... )
+
+        >>> graphicsmagick
+        HgURL(url=ssh://yourid@hg.GraphicsMagick.org//hg/GraphicsMagick,
+                scheme=ssh,
+                user=yourid,
+                hostname=hg.GraphicsMagick.org,
+                path=/hg/GraphicsMagick,
+                matcher=core-hg)
+
+        >>> graphicsmagick.to_url()
+        'ssh://yourid@hg.GraphicsMagick.org//hg/GraphicsMagick'
+
+        Switch the username:
+
+        >>> graphicsmagick.user = 'lucas'
+
+        >>> graphicsmagick.to_url()
+        'ssh://lucas@hg.GraphicsMagick.org//hg/GraphicsMagick'
 
-        - Formats: Show an example converting a hghub url from ssh -> https format,
-          and the other way around.
         """
-        if self.scheme is not None:
-            parts = [self.scheme, "://", self.hostname, "/", self.path]
-        else:
-            parts = [self.user or "hg", "@", self.hostname, ":", self.path]
+        parts = [self.scheme or "ssh", "://"]
+        if self.user:
+            parts.extend([self.user, "@"])
+
+        parts.append(self.hostname)
+
+        if self.port is not None:
+            parts.extend([":", f"{self.port}"])
+
+        parts.extend([self.separator, self.path])
 
         return "".join(part for part in parts if isinstance(part, str))

libvcs/parse/svn.py

@@ -29,17 +29,20 @@
 RE_PATH = r"""
     ((?P<user>.*)@)?
     (?P<hostname>([^/:]+))
-    (?P<separator>[:,/])?
+    (:(?P<port>\d{1,4}))?
+    (?P<separator>/)?
     (?P<path>
       (\w[^:.]*)
     )?
 """
 
+# Valid schemes for svn(1).
+# See Table 1.1 Repository access URLs in SVN Book
+# https://svnbook.red-bean.com/nightly/en/svn.basic.in-action.html#svn.basic.in-action.wc.tbl-1
 RE_SCHEME = r"""
     (?P<scheme>
       (
-        http|https|
-        svn\+ssh
+        file|http|https|svn|svn\+ssh
       )
     )
 """
@@ -154,10 +157,12 @@ class SvnURL(URLProtocol, SkipDefaultFieldsReprMixin):
     """
 
     url: str
-    scheme: str = dataclasses.field(init=False)
-    hostname: str = dataclasses.field(init=False)
-    path: str = dataclasses.field(init=False)
+    scheme: Optional[str] = None
     user: Optional[str] = None
+    hostname: str = dataclasses.field(default="")
+    port: Optional[int] = None
+    separator: str = dataclasses.field(default="/")
+    path: str = dataclasses.field(default="")
 
     #
     # commit-ish: ref
@@ -213,9 +218,9 @@ def to_url(self) -> str:
         >>> svn_location
         SvnURL(url=svn+ssh://my-username@my-server/vcs-python/libvcs,
                 scheme=svn+ssh,
+                user=my-username,
                 hostname=my-server,
                 path=vcs-python/libvcs,
-                user=my-username,
                 matcher=core-svn)
 
         Switch repo libvcs -> vcspull:
@@ -232,12 +237,15 @@ def to_url(self) -> str:
         >>> svn_location.to_url()
         'svn+ssh://tom@my-server/vcs-python/vcspull'
         """
-        if self.scheme is not None:
-            parts = [self.scheme, "://"]
-            if self.user:
-                parts.extend([self.user, "@"])
-            parts += [self.hostname, "/", self.path]
-        else:
-            parts = [self.user or "svn", "@", self.hostname, ":", self.path]
+        parts = [self.scheme or "ssh", "://"]
+        if self.user:
+            parts.extend([self.user, "@"])
+
+        parts.append(self.hostname)
+
+        if self.port is not None:
+            parts.extend([":", f"{self.port}"])
+
+        parts.extend([self.separator, self.path])
 
         return "".join(part for part in parts if isinstance(part, str))