Report data file errors in more detail: include file and directory paths, and make helpful suggestions

Author: bartfeenstra

coverage/cmdline.py

@@ -13,6 +13,7 @@
 import sys
 import textwrap
 import traceback
+from contextlib import suppress
 
 from typing import cast, Any, NoReturn
 
@@ -24,7 +25,8 @@
 from coverage.control import DEFAULT_DATAFILE
 from coverage.data import combinable_files, debug_data_file
 from coverage.debug import info_header, short_stack, write_formatted_info
-from coverage.exceptions import _BaseCoverageException, _ExceptionDuringRun, NoSource
+from coverage.exceptions import _BaseCoverageException, _ExceptionDuringRun, NoSource, \
+    NoDataFilesFoundError
 from coverage.execfile import PyRunner
 from coverage.results import display_covered, should_fail_under
 from coverage.version import __url__
@@ -882,9 +884,10 @@ def do_debug(self, args: list[str]) -> int:
             print(info_header("data"))
             data_file = self.coverage.config.data_file
             debug_data_file(data_file)
-            for filename in combinable_files(data_file):
-                print("-----")
-                debug_data_file(filename)
+            with suppress(NoDataFilesFoundError):
+                for filename in combinable_files(data_file):
+                    print("-----")
+                    debug_data_file(filename)
         elif args[0] == "config":
             write_formatted_info(print, "config", self.coverage.config.debug_info())
         elif args[0] == "premain":

coverage/data.py

@@ -18,7 +18,8 @@
 
 from typing import Callable, Iterable
 
-from coverage.exceptions import CoverageException, NoDataError
+from coverage.exceptions import CoverageException, NoDataError, DataFileNotFoundError, \
+    NoDataFilesFoundError
 from coverage.files import PathAliases
 from coverage.misc import Hasher, file_be_gone, human_sorted, plural
 from coverage.sqldata import CoverageData
@@ -82,12 +83,15 @@ def combinable_files(data_file: str, data_paths: Iterable[str] | None = None) ->
             pattern = glob.escape(os.path.join(os.path.abspath(p), local)) +".*"
             files_to_combine.extend(glob.glob(pattern))
         else:
-            raise NoDataError(f"Couldn't combine from non-existent path '{p}'")
+            raise DataFileNotFoundError.new_for_data_file(p, True)
 
     # SQLite might have made journal files alongside our database files.
     # We never want to combine those.
     files_to_combine = [fnm for fnm in files_to_combine if not fnm.endswith("-journal")]
 
+    if not files_to_combine:
+        raise NoDataFilesFoundError.new_for_data_directory(data_dir)
+
     # Sorting isn't usually needed, since it shouldn't matter what order files
     # are combined, but sorting makes tests more predictable, and makes
     # debugging more understandable when things go wrong.
@@ -129,10 +133,12 @@ def combine_parallel_data(
     `message` is a function to use for printing messages to the user.
 
     """
-    files_to_combine = combinable_files(data.base_filename(), data_paths)
-
-    if strict and not files_to_combine:
-        raise NoDataError("No data to combine")
+    try:
+        files_to_combine = combinable_files(data.base_filename(), data_paths)
+    except NoDataFilesFoundError:
+        if strict:
+            raise
+        return None
 
     file_hashes = set()
     combined_any = False
@@ -190,6 +196,7 @@ def combine_parallel_data(
             file_be_gone(f)
 
     if strict and not combined_any:
+        # @todo Files found but no usable data in them. Parse error?
         raise NoDataError("No usable data files")
 
 

coverage/exceptions.py

@@ -5,6 +5,9 @@
 
 from __future__ import annotations
 
+import os.path
+
+
 class _BaseCoverageException(Exception):
     """The base-base of all Coverage exceptions."""
     pass
@@ -24,11 +27,29 @@ class DataError(CoverageException):
     """An error in using a data file."""
     pass
 
+
 class NoDataError(CoverageException):
     """We didn't have data to work with."""
     pass
 
 
+class DataFileNotFoundError(NoDataError):
+    """A data file or data directory could be found."""
+    @classmethod
+    def new_for_data_file(cls, data_file_path: str, combined: bool = False) -> 'Self':
+        message = f"The data file or directory `{os.path.abspath(data_file_path)}` could not be found."
+        if not combined:
+            message += " Perhaps `coverage combine` must be run first."
+        return cls(message)
+
+
+class NoDataFilesFoundError(NoDataError):
+    """No data files could be found in a data directory."""
+    @classmethod
+    def new_for_data_directory(cls, data_directory_path: str) -> 'Self':
+        return cls(f"The data directory `{os.path.abspath(data_directory_path)}` does not contain any data files.")
+
+
 class NoSource(CoverageException):
     """We couldn't find the source for a module."""
     pass

tests/test_api.py

@@ -446,7 +446,7 @@ def test_combining_twice(self) -> None:
         self.assert_exists(".coverage")
 
         cov2 = coverage.Coverage()
-        with pytest.raises(NoDataError, match=r"No data to combine"):
+        with pytest.raises(NoDataError, match=r"^The data directory `(.+?)` does not contain any data files.$"):
             cov2.combine(strict=True, keep=False)
 
         cov3 = coverage.Coverage()
@@ -1326,7 +1326,7 @@ def test_combine_parallel_data(self) -> None:
         # Running combine again should fail, because there are no parallel data
         # files to combine.
         cov = coverage.Coverage()
-        with pytest.raises(NoDataError, match=r"No data to combine"):
+        with pytest.raises(NoDataError):
             cov.combine(strict=True)
 
         # And the originally combined data is still there.

tests/test_data.py

@@ -915,7 +915,7 @@ def test_combining_from_files(self) -> None:
 
     def test_combining_from_nonexistent_directories(self) -> None:
         covdata = DebugCoverageData()
-        msg = "Couldn't combine from non-existent path 'xyzzy'"
+        msg = r"^The data file or directory `(.+?)` could not be found.$"
         with pytest.raises(NoDataError, match=msg):
             combine_parallel_data(covdata, data_paths=['xyzzy'])