collective / icalendar / 18758844215

98%

push

github

Fix #355: Preserve URL-encoded characters in content line parsing (#935)

* Add CHANGES.rst entry for issue #355 URL escaping fix

* Fix #355: Properly handle URL-encoded characters in content lines

The previous implementation used escape_string() and unescape_string()
as a workaround to temporarily hide backslash-escaped characters during
parsing. This corrupted already URL-encoded values like %3A (colon) by
incorrectly unescaping them.

This fix implements proper parsing that respects:
- Double-quote delimited sections (case-sensitive)
- Backslash escaping outside quotes
- Delimiter detection only when not quoted or escaped

This prevents URL-encoded characters in values (like Facebook event URLs)
from being corrupted during parsing.

Added type hints to parts() method.

Based on original work by NicoHood in PR #356.

Co-authored-by: NicoHood <NicoHood@users.noreply.github.com>

Fixes #355
Closes #356
Closes #426

* Add test events for issue #355 URL escaping

Test cases with URL-encoded characters in DESCRIPTION fields to verify
that URL-encoded values are not corrupted during parsing.

Co-authored-by: Nicco Kunzmann <niccokunzmann@rambler.ru>

* Add tests for issue #355 URL escaping

Tests verify that URL-encoded characters in DESCRIPTION fields
are not incorrectly unescaped during parsing.

Co-authored-by: Nicco Kunzmann <niccokunzmann@rambler.ru>

* Add unescape_backslash() for RFC 5545 backslash escaping

Adds a new function that only unescapes backslash sequences (\, \; \: \)
without touching URL-encoded values (%XX). This prevents corruption of
URL-encoded characters in values while still properly handling RFC 5545
backslash escaping.

The original unescape_string() conflates backslash escaping with URL
encoding, which was the root cause of issue #355.

* Fix ruff linting: use lowercase for local variable

* Fix error messages to match test expectations

When content lines have no delimiters or invalid names, the error
messages now match wha... (continued)

947 of 976 branches covered (97.03%)

Branch coverage included in aggregate %.

39 of 39 new or added lines in 2 files covered. (100.0%)

8106 of 8275 relevant lines covered (97.96%)

7.79 hits per line