The `should_ignore_error()` method was added in f191809 to support
context preservation for debugging, but it no longer serves its
original purpose and adds unnecessary overhead.
Issues with the current implementation:
- Always returns False by default
- Called on every single request with no benefit
- The original intention for error ignoring during debugging is not
how context preservation works anymore
- No documentation beyond API reference
- No tests for the functionality
- No evidence of real-world usage
Changes:
- Add deprecation warning to App.should_ignore_error() that will be
removed in Flask 4.0
- Optimize call site to only invoke the method if it's been overridden
by a subclass, eliminating the function call overhead for 99.9% of
requests
- Add comprehensive tests for the deprecation behavior
- Update CHANGES.rst with deprecation notice
Teardown handlers should manage their own error handling instead of
relying on this method.
Fixes#5816
* update pre-commit hook
* upgrade pip with venv
* update description and version
* show url in publish environment
* update versions
* update versions, separate typing job
* use dependabot grouped updates
ignore upload/download-artifact until slsa updates
* use sphinx.ext.extlinks instead of sphinx-issues
* update dev dependencies
* update editorconfig
* update gitignore
* update .readthedocs.yaml
* license is txt, readme is md
* update pyproject.toml
add typed classifier
add pyright config
simplify urls
* tox builds docs in place
* update min test py version
* add tox env to update all dev dependencies
* update issue and pr templates
* rename security docs page to not conflict with org policy file
* simplify matrix
This makes the Flask.url_for self argument positional only (Flask
supports Python 3.8+) thereby restoring the ability to pass self as a
value argument to url_for.
This ensures that a child's subdomain prefixs any parent subdomain
such that the full domain is child.parent.domain.tld and onwards with
further nesting. This makes the most sense to users and mimics how
url_prefixes work (although subdomains suffix).
Remove the `--eager-loading/--lazy-loading` options and the
`DispatchingApp` middleware. The `run` command handles loading
exceptions directly. The reloader always prints out tracebacks
immediately and always defers raising the error.
