Fix EVR regex to not match email addresses with - in them (#445)

Fix EVR regex to not match email addresses with - in them

TODO:

 Write new tests or update the old ones to cover new functionality.
 Update doc-strings where appropriate.
 Update or write new documentation in packit/packit.dev.

Reviewed-by: Nikola Forró
Reviewed-by: Dan Čermák <[email protected]>

Author: softwarefactory-project-zuul[bot]

specfile/changelog.py

@@ -46,6 +46,23 @@ class ChangelogEntry:
         content: List of lines forming the content of the entry.
     """
 
+    _EVR_RE = re.compile(
+        r"""
+            ^.*
+            \d{4}                    # year
+            \s+                      # whitespace
+            .+                       # author
+            \s+                      # preceding whitespace
+            ((?P<sb>\[)|(?P<rb>\())? # optional opening bracket
+            (?P<evr>\S+?)            # EVR
+            (?(sb)\]|(?(rb)\)))      # matching closing bracket
+            :?                       # optional colon
+            \s*                      # optional following whitespace
+            $
+            """,
+        re.VERBOSE,
+    )
+
     def __init__(
         self,
         header: str,
@@ -83,25 +100,24 @@ def __repr__(self) -> str:
         return f"ChangelogEntry({self.header!r}, {self.content!r}, {self._following_lines!r})"
 
     @property
-    def evr(self) -> Optional[str]:
+    def evr(self) -> Optional[EVR]:
         """EVR (epoch, version, release) of the entry."""
-        m = re.match(
-            r"""
-            ^.*
-            \s+                       # preceding whitespace
-            ((?P<sb>\[)|(?P<rb>\())?  # optional opening bracket
-            (?P<evr>(\d+:)?\S+-\S+?)  # EVR
-            (?(sb)\]|(?(rb)\)))       # matching closing bracket
-            :?                        # optional colon
-            \s*                       # optional following whitespace
-            $
-            """,
-            self.header,
-            re.VERBOSE,
-        )
+        m = self._EVR_RE.match(self.header)
+
         if not m:
             return None
-        return m.group("evr")
+
+        evr_s = m.group("evr")
+        try:
+            evr = EVR.from_string(evr_s)
+        except SpecfileException:
+            return None
+
+        # looks like an email
+        if evr_s.startswith("<") and evr_s.endswith(">"):
+            return None
+
+        return evr
 
     @property
     def extended_timestamp(self) -> bool:
@@ -263,7 +279,9 @@ def copy(self) -> "Changelog":
         return copy.copy(self)
 
     def filter(
-        self, since: Optional[str] = None, until: Optional[str] = None
+        self,
+        since: Optional[Union[str, EVR]] = None,
+        until: Optional[Union[str, EVR]] = None,
     ) -> "Changelog":
         """
         Filters changelog entries with EVR between since and until.
@@ -278,9 +296,11 @@ def filter(
             Filtered changelog.
         """
 
-        def parse_evr(s):
+        def parse_evr(str_or_evr: Union[str, EVR]) -> EVR:
+            if isinstance(str_or_evr, EVR):
+                return str_or_evr
             try:
-                return EVR.from_string(s)
+                return EVR.from_string(str_or_evr)
             except SpecfileException:
                 return EVR(version="0")
 
@@ -290,8 +310,8 @@ def parse_evr(s):
             start_index = next(
                 (
                     i
-                    for i, e in enumerate(self.data)
-                    if parse_evr(e.evr) >= parse_evr(since)
+                    for i, entry in enumerate(self.data)
+                    if entry.evr >= parse_evr(since)
                 ),
                 len(self.data) + 1,
             )
@@ -301,8 +321,8 @@ def parse_evr(s):
             end_index = next(
                 (
                     i + 1
-                    for i, e in reversed(list(enumerate(self.data)))
-                    if parse_evr(e.evr) <= parse_evr(until)
+                    for i, entry in reversed(list(enumerate(self.data)))
+                    if entry.evr <= parse_evr(until)
                 ),
                 0,
             )

tests/unit/test_changelog.py

@@ -3,44 +3,57 @@
 
 import copy
 import datetime
+from typing import Optional
 
 import pytest
 
 from specfile.changelog import Changelog, ChangelogEntry
 from specfile.sections import Section
+from specfile.utils import EVR
 
 
 @pytest.mark.parametrize(
     "header, evr",
     [
         ("* Thu Jan 04 2007 Michael Schwendt <[email protected]>", None),
+        ("* Thu Jan 04 2007 Michael Schwendt <[email protected]>", None),
+        (
+            "* Fri Jul 26 2024 Miroslav Suchý <[email protected]> - ss981107-67",
+            EVR(version="ss981107", release="67"),
+        ),
         (
             "* Mon Jul 13 2020 Tom Stellard <[email protected]> 4.0-0.4.pre2",
-            "4.0-0.4.pre2",
+            EVR(version="4.0", release="0.4.pre2"),
+        ),
+        (
+            "* Fri Jul 20 2018 Gwyn Ciesla <[email protected]> - 0.52-6",
+            EVR(version="0.52", release="6"),
         ),
-        ("* Fri Jul 20 2018 Gwyn Ciesla <[email protected]> - 0.52-6", "0.52-6"),
         (
             "* Mon Feb 23 2009 Fedora Release Engineering <[email protected]> "
             "- 1.23-3.20081106gitbe42b4",
-            "1.23-3.20081106gitbe42b4",
+            EVR(version="1.23", release="3.20081106gitbe42b4"),
         ),
         (
             "* Thu Feb 04 2016 Marcin Zajaczkowski <mszpak ATT wp DOTT pl> - 1:0.9.10-6",
-            "1:0.9.10-6",
+            EVR(epoch=1, version="0.9.10", release="6"),
         ),
         (
             "* Mon Jan 03 2022 Fedora Kernel Team <[email protected]> [5.16-0.rc8.55]",
-            "5.16-0.rc8.55",
+            EVR(version="5.16", release="0.rc8.55"),
+        ),
+        (
+            "* Wed Jan 23 2002 Karsten Hopp <[email protected]> (4.6-1)",
+            EVR(version="4.6", release="1"),
         ),
-        ("* Wed Jan 23 2002 Karsten Hopp <[email protected]> (4.6-1)", "4.6-1"),
         (
             "* Thu Apr  9 2015 Jeffrey C. Ollie <[email protected]> - 13.3.2-1:",
-            "13.3.2-1",
+            EVR(version="13.3.2", release="1"),
         ),
     ],
 )
-def test_entry_evr(header, evr):
-    assert ChangelogEntry(header, [""]).evr == evr
+def test_entry_evr(header, evr: Optional[EVR]):
+    assert evr == ChangelogEntry(header, [""]).evr
 
 
 @pytest.mark.parametrize(
@@ -133,7 +146,9 @@ def test_filter(since, until, evrs):
             ),
         ]
     )
-    assert [e.evr for e in changelog.filter(since=since, until=until)] == evrs
+    assert [e.evr for e in changelog.filter(since=since, until=until)] == [
+        EVR.from_string(evr) for evr in evrs
+    ]
 
 
 def test_parse():