Support Pytest 7.4+

Co-authored-by: Henry Schreiner <henryschreineriii@gmail.com>

.github/workflows/test.yml

@@ -54,6 +54,12 @@ jobs:
         PYTEST_MAJOR_VERSION: 7
         PYTEST_PLUGINS: pytest_github_actions_annotate_failures
 
+    - name: Run tests with PyTest 8
+      run: tox
+      env:
+        PYTEST_MAJOR_VERSION: 8
+        PYTEST_PLUGINS: pytest_github_actions_annotate_failures
+
   post-test:
     name: All tests passed
     if: always()

plugin_test.py

@@ -223,7 +223,7 @@ def test_fail():
     result = testdir.runpytest_subprocess("--rootdir=foo")
     result.stderr.fnmatch_lines(
         [
-            "::error file=test_annotation_fail_cwd.py,line=5::test_fail*assert 0*",
+            "::error file=test_annotation_fail_cwd0/test_annotation_fail_cwd.py,line=5::test_fail*assert 0*",
         ]
     )
 

pytest_github_actions_annotate_failures/plugin.py

@@ -6,7 +6,7 @@
 from typing import TYPE_CHECKING
 
 import pytest
-from _pytest._code.code import ExceptionRepr
+from _pytest._code.code import ExceptionRepr, ReprEntry
 from packaging import version
 
 if TYPE_CHECKING:
@@ -38,59 +38,65 @@ def pytest_runtest_makereport(item: Item, call):  # noqa: ARG001
         return
 
     if report.when == "call" and report.failed:
-        # collect information to be annotated
         filesystempath, lineno, _ = report.location
 
-        runpath = os.environ.get("PYTEST_RUN_PATH")
-        if runpath:
-            filesystempath = os.path.join(runpath, filesystempath)
-
-        # try to convert to absolute path in GitHub Actions
-        workspace = os.environ.get("GITHUB_WORKSPACE")
-        if workspace:
-            full_path = os.path.abspath(filesystempath)
-            try:
-                rel_path = os.path.relpath(full_path, workspace)
-            except ValueError:
-                # os.path.relpath() will raise ValueError on Windows
-                # when full_path and workspace have different mount points.
-                # https://github.com/utgwkk/pytest-github-actions-annotate-failures/issues/20
-                rel_path = filesystempath
-            if not rel_path.startswith(".."):
-                filesystempath = rel_path
-
         if lineno is not None:
             # 0-index to 1-index
             lineno += 1
 
-        # get the name of the current failed test, with parametrize info
         longrepr = report.head_line or item.name
 
         # get the error message and line number from the actual error
         if isinstance(report.longrepr, ExceptionRepr):
             if report.longrepr.reprcrash is not None:
                 longrepr += "\n\n" + report.longrepr.reprcrash.message
             tb_entries = report.longrepr.reprtraceback.reprentries
-            if len(tb_entries) > 1 and tb_entries[0].reprfileloc is not None:
+            if tb_entries:
+                entry = tb_entries[0]
                 # Handle third-party exceptions
-                lineno = tb_entries[0].reprfileloc.lineno
+                if isinstance(entry, ReprEntry) and entry.reprfileloc is not None:
+                    lineno = entry.reprfileloc.lineno
+                    filesystempath = entry.reprfileloc.path
+
             elif report.longrepr.reprcrash is not None:
                 lineno = report.longrepr.reprcrash.lineno
         elif isinstance(report.longrepr, tuple):
-            _, lineno, message = report.longrepr
+            filesystempath, lineno, message = report.longrepr
             longrepr += "\n\n" + message
         elif isinstance(report.longrepr, str):
             longrepr += "\n\n" + report.longrepr
 
         workflow_command = _build_workflow_command(
             "error",
-            filesystempath,
+            compute_path(filesystempath),
             lineno,
             message=longrepr,
         )
         print(workflow_command, file=sys.stderr)
 
 
+def compute_path(filesystempath: str) -> str:
+    """Extract and process location information from the report."""
+    runpath = os.environ.get("PYTEST_RUN_PATH")
+    if runpath:
+        filesystempath = os.path.join(runpath, filesystempath)
+
+    # try to convert to absolute path in GitHub Actions
+    workspace = os.environ.get("GITHUB_WORKSPACE")
+    if workspace:
+        full_path = os.path.abspath(filesystempath)
+        try:
+            rel_path = os.path.relpath(full_path, workspace)
+        except ValueError:
+            # os.path.relpath() will raise ValueError on Windows
+            # when full_path and workspace have different mount points.
+            rel_path = filesystempath
+        if not rel_path.startswith(".."):
+            filesystempath = rel_path
+
+    return filesystempath
+
+
 class _AnnotateWarnings:
     def pytest_warning_recorded(self, warning_message, when, nodeid, location):  # noqa: ARG002
         # enable only in a workflow of GitHub Actions
@@ -139,14 +145,14 @@ def pytest_configure(config):
 
 
 def _build_workflow_command(
-    command_name,
-    file,
-    line,
-    end_line=None,
-    column=None,
-    end_column=None,
-    title=None,
-    message=None,
+    command_name: str,
+    file: str,
+    line: int,
+    end_line: int | None = None,
+    column: int | None = None,
+    end_column: int | None = None,
+    title: str | None = None,
+    message: str | None = None,
 ):
     """Build a command to annotate a workflow."""
     result = f"::{command_name} "
@@ -168,5 +174,5 @@ def _build_workflow_command(
     return result
 
 
-def _escape(s):
+def _escape(s: str) -> str:
     return s.replace("%", "%25").replace("\r", "%0D").replace("\n", "%0A")

tox.ini

@@ -20,7 +20,7 @@ PYTEST_MAJOR_VERSION =
 extras = test
 deps =
     pytest6: pytest>=6.0.0,<7.0.0
-    pytest7: pytest>=7.0.0,<7.4.0
+    pytest7: pytest>=7.0.0,<8.0.0
     pytest8: pytest>=8.0.0,<9.0.0
 
 commands = {envpython} -m pytest {posargs}