From 297ea519dd5a50d08323ae72c8031c3d28856480 Mon Sep 17 00:00:00 2001
From: David Roman <droman@ifae.es>
Date: Fri, 22 Aug 2025 13:13:51 +0200
Subject: [PATCH] fix parsing of atoms

Previously atoms like category/a-long-name were parsed in such a way
that "name" was taken as the version
---
 ebuildtester/atom.py | 79 +++++++++++++++++++++++++++++++++++++++++---
 test/test_atom.py    | 19 ++++++++---
 2 files changed, 89 insertions(+), 9 deletions(-)

diff --git a/ebuildtester/atom.py b/ebuildtester/atom.py
index 7a7d04c..9015876 100644
--- a/ebuildtester/atom.py
+++ b/ebuildtester/atom.py
@@ -1,11 +1,82 @@
 """An Atom."""
 
+import re
+
 
 class AtomException(Exception):
     """An exception in this class."""
     pass
 
 
+# from lib/portage/version.py
+# PMS 3.1.3: A slot name may contain any of the characters [A-Za-z0-9+_.-].
+# It must not begin with a hyphen or a dot.
+_slot = r"([\w+][\w+.-]*)"
+
+# 2.1.1 A category name may contain any of the characters [A-Za-z0-9+_.-].
+# It must not begin with a hyphen or a dot.
+_cat = r"[\w+][\w+.-]*"
+
+# 2.1.2 A package name may contain any of the characters [A-Za-z0-9+_-].
+# It must not begin with a hyphen,
+# and must not end in a hyphen followed by one or more digits.
+_pkg = r"[\w+][\w+-]*?"
+
+_v = r"(\d+)((\.\d+)*)([a-z]?)((_(pre|p|beta|alpha|rc)\d*)*)"
+_rev = r"\d+"
+_vr = _v + "(-r(" + _rev + "))?"
+
+_cp = "(" + _cat + "/" + _pkg + "(-" + _vr + ")?)"
+_cpv = "(" + _cp + "-" + _vr + ")"
+_pv = (
+    "(?P<pn>"
+    + _pkg
+    + "(?P<pn_inval>-"
+    + _vr
+    + ")?)"
+    + "-(?P<ver>"
+    + _v
+    + ")(-r(?P<rev>"
+    + _rev
+    + "))?"
+)
+
+_pv_re = None
+
+
+def _get_pv_re():
+    global _pv_re
+    if _pv_re is not None:
+        return _pv_re
+
+    _pv_re = re.compile(r"^" + _pv + r"$", re.VERBOSE | re.UNICODE)
+
+    return _pv_re
+
+
+def _pkgsplit(mypkg: str):
+    """
+    @param mypkg: pv
+    @return:
+    1. None if input is invalid.
+    2. (pn, ver, rev) if input is pv
+    """
+    m = _get_pv_re().match(mypkg)
+    if m is None:
+        return None
+
+    if m.group("pn_inval") is not None:
+        # package name appears to have a version-like suffix
+        return None
+
+    rev = m.group("rev")
+    if rev is None:
+        rev = "0"
+    rev = "r" + rev
+
+    return (m.group("pn"), m.group("ver"), rev)
+
+
 class Atom(object):
 
     def __init__(self, atom):
@@ -26,10 +97,10 @@ def __init__(self, atom):
 
         # Split off version.
         try:
-            temp = self.package.index("-")
-            if temp > -1:
-                self.version = self.package[temp + 1:]
-                self.package = self.package[:temp]
+            # If atom don't start with "=" we assume there is no version
+            if '=' in atom:
+                self.package, self.version, rev = _pkgsplit(self.package)
+                self.version = f"{self.version}-{rev}"
         except ValueError:
             pass
 
diff --git a/test/test_atom.py b/test/test_atom.py
index 95f8343..b92d4b1 100644
--- a/test/test_atom.py
+++ b/test/test_atom.py
@@ -7,7 +7,7 @@ class TestAtom(unittest.TestCase):
     def test_equals(self):
         atom = Atom("=CATEGORY/PACKAGE-1.0.0")
         self.assertEqual(atom.__str__()[0], "=")
-        self.assertEqual(atom.__str__(), "=CATEGORY/PACKAGE-1.0.0")
+        self.assertEqual(atom.__str__(), "=CATEGORY/PACKAGE-1.0.0-r0")
 
     def test_slash_1(self):
         with self.assertRaises(AtomException) as e:
@@ -19,12 +19,21 @@ def test_slash_2(self):
         self.assertEqual(atom.package, "PACKAGE")
 
     def test_version(self):
-        self.assertEqual(Atom("CATEGORY/PACKAGE-1").version, "1")
-        self.assertEqual(Atom("CATEGORY/PACKAGE-1.0").version, "1.0")
-        self.assertEqual(
-            Atom("CATEGORY/PACKAGE-1.0-r1").version, "1.0-r1")
+        self.assertEqual(Atom("CATEGORY/PACKAGE-1").version, None)
+        self.assertEqual(Atom("CATEGORY/PACKAGE-NAME-1").version, None)
+        self.assertEqual(Atom("CATEGORY/PACKAGE-1.0-r1").version, None)
+        self.assertEqual(Atom("CATEGORY/P-NAME-1.0-r1").version, None)
+
+        self.assertEqual(Atom("=CATEGORY/PACKAGE-1").version, "1-r0")
+        self.assertEqual(Atom("=CATEGORY/PACKAGE-NAME-1").version, "1-r0")
+        self.assertEqual(Atom("=CATEGORY/PACKAGE-NAME-1.0").version, "1.0-r0")
+        self.assertEqual(Atom("=CATEGORY/PACKAGE-1.0-r1").version, "1.0-r1")
 
     def test_str(self):
         atom_1 = Atom("=CATEGORY/PACKAGE-1.0.0-r1")
         atom_2 = Atom(str(atom_1))
         self.assertEqual(atom_1, atom_2)
+        atom_3 = Atom("=CATEGORY/PACKAGE-NAME-1.0.0-r1")
+        atom_4 = Atom(str(atom_3))
+        self.assertEqual(atom_3, atom_4)
+