What causes "Invalid regex {e.pattern!r}: {e}" in pytest-dev/pytest?

Escape regex metacharacters in your message pattern, or test it with python -c 'import re; re.compile("your pattern")'.

pytest-dev/pytest · error · UsageError

Invalid regex {e.pattern!r}: {e}

Error message

Invalid regex {e.pattern!r}: {e}

What it means

Pytest raises this UsageError when the message or module field of a warning filter specification is not a valid regular expression. Pytest compiles both fields with re.compile() after optional escaping; a re.error indicates the regex syntax is broken.

Source

Thrown at src/_pytest/config/__init__.py:2320

        message = re.escape(message)
    if module and escape:
        module = re.escape(module) + r"\Z"
    if lineno_:
        try:
            lineno = int(lineno_)
            if lineno < 0:
                raise ValueError("number is negative")
        except ValueError as e:
            raise UsageError(
                error_template.format(error=f"invalid lineno {lineno_!r}: {e}")
            ) from None
    else:
        lineno = 0
    try:
        re.compile(message)
        re.compile(module)
    except re.error as e:
        raise UsageError(
            error_template.format(error=f"Invalid regex {e.pattern!r}: {e}")
        ) from None
    return action, message, category, module, lineno


def _resolve_warning_category(category: str) -> type[Warning]:
    """
    Copied from warnings._getcategory, but changed so it lets exceptions (specially ImportErrors)
    propagate so we can get access to their tracebacks (#9218).
    """
    __tracebackhide__ = True
    if not category:
        return Warning

    if "." not in category:
        import builtins as m

        klass = category

View on GitHub (pinned to 98b357f69e)

Solutions

Escape regex metacharacters in your message pattern, or test it with python -c 'import re; re.compile("your pattern")'.
Use plain substring text (no metacharacters) for simple matches.
If you need a literal special character, prefix with backslash (e.g. '\(' to match a literal paren).

Example fix

# before (unbalanced bracket)
[pytest]
filterwarnings = ["ignore:[unbalanced:DeprecationWarning"]

# after
[pytest]
filterwarnings = ["ignore:\[unbalanced:DeprecationWarning"]

Defensive patterns

Strategy: validation

Validate before calling

import re
def validate_filter_regex(message: str, module: str) -> None:
    for label, pattern in (('message', message), ('module', module)):
        if pattern:
            try:
                re.compile(pattern)
            except re.error as e:
                raise ValueError(f'Invalid {label} regex {pattern!r}: {e}') from None

Type guard

import re
def is_valid_regex(pattern: str) -> bool:
    try:
        re.compile(pattern)
        return True
    except re.error:
        return False

Try / catch

import re
try:
    re.compile(user_message)
except re.error as e:
    print(f'message field is not valid regex: {e}'); user_message = re.escape(user_message)

Prevention

Prefer plain substrings over regex when full pattern matching isn't needed.
Run re.compile() on hand-written filter patterns before committing them.
Use re.escape() on dynamic substrings embedded in filter regexes.

When it happens

Trigger: A filter whose message or module field contains invalid regex syntax, e.g. 'ignore:[unbalanced::DeprecationWarning' (unmatched bracket) or 'ignore:msg:*bad::DeprecationWarning'. When escape=False (ini/filterwarnings mark) the raw string is compiled; when escape=True (command-line -W) re.escape is applied first, so this most often triggers from ini/mark filters with hand-written regex.

Common situations: Writing a partial regex in filterwarnings without escaping special chars; unbalanced parentheses/brackets in the message pattern; using a glob-style '*' expecting shell semantics instead of regex.

Related errors

AI-assisted analysis of pytest-dev/pytest@98b357f69e (2026-08-04). Data as JSON: /data/errors/fc62040c9ce5398a.json. Report an issue: GitHub.