Fix wsgi invalid request uri

[IonDragos2003]

Description

Fixes #4447

WSGI Instrumentation currently parses RAW_URI / REQUEST_URI to derive url.path & url.query. Malformed request URIs can cause urllib.parse.urlparse to raise a ValueError, which breaks instrumentation and may result in a 500 response.

This change avoids parsing the raw request URI and instead derives url.path and url.query directly from the WSGI environ (PATH_INFO and QUERY_STRING).

Type of change

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

• Added a regression test:
  • test_request_attributes_with_invalid_request_uri_uses_wsgi_environ
  • Verifies that malformed REQUEST_URI does not raise and that URL_PATH / URL_QUERY are derived from WSGI environ.
• Updated existing test:
  • test_request_attributes_with_full_request_uri to reflect new behavior (no parsing of REQUEST_URI).

Ran tests:
uv run python -m pytest instrumentation/opentelemetry-instrumentation-wsgi/tests

All tests pass.

Does This PR Require a Core Repo Change?

• No.

Checklist:

• Followed the style guidelines of this project
• Changelogs have been updated
• Unit tests have been added
• Documentation has been updated

[emdneto]

example.com/ still valid no? I don't see the exception running this test against main

[IonDragos2003]

Hi @emdneto, seems that these changes are now breaking some Falcon tests.

Should I do something like

if target:
    try:
        path, query = _parse_url_query(target)
    except ValueError:
        path = environ.get("PATH_INFO")
        query = environ.get("QUERY_STRING")
    _set_http_target(result, target, path, query, sem_conv_opt_in_mode)

[emdneto]

Hi @emdneto, seems that these changes are now breaking some Falcon tests.

Should I do something like

if target:
    try:
        path, query = _parse_url_query(target)
    except ValueError:
        path = environ.get("PATH_INFO")
        query = environ.get("QUERY_STRING")
    _set_http_target(result, target, path, query, sem_conv_opt_in_mode)

@IonDragos2003

I would fix the Falcon tests instead, since the target is no longer used to parse the path and query. Because of that, _has_fixed_http_target in the Falcon tests no longer has the same effect it used to.

Also, please fix precommit by running tox -e precommit and add a changelog entry! Thanks

[emdneto]

Thanks for the PR!

Just a heads-up: we no longer update CHANGELOG.md directly. The changelog is now generated from changelog fragments using Towncrier.

Please add the appropriate changelog fragment for this change instead of editing CHANGELOG.md manually. You can find the instructions and expected format in CONTRIBUTING.md.

[tammy-baylis-swi]

Lgtm and adheres to PEP 333.

[emdneto]

Last comment: i believe here you'll have to set PATH_INFO to get the correct path instead

[IonDragos2003]

Hi,

I hope I added this correctly. I am on holidays and added this quickly on my phone ..

I can look at this once I am back. Thank you