gh-142411: Change documentation to reflect the new docstring adjustments in 3.13 #142413
Conversation
|
All commit authors signed the Contributor License Agreement. |
| indented less should not occur, but if they occur all their leading whitespace | ||
| should be stripped. Equivalence of whitespace should be tested after expansion | ||
| of tabs (to 8 spaces, normally). | ||
| The Python parser strips indentation from multi-line string literals when they |
Doc/library/functions.rst
Outdated
|
|
||
| - :exc:`SyntaxError` if the compiled source is invalid | ||
| - :exc:`ValueError` if the source contains null bytes | ||
| - :exc:`UnicodeDecodeError` if the source code contains docstrings that |
There was a problem hiding this comment.
I wouldn't try to keep this list precise and complete. The exceptions change over time, and it's not like people are going to write except blocks for these specific exceptions. Details are in the versionchanged blocks below. Here can we say that SyntaxError happens if the source is invalid, and other exceptions if the text can't be read in various ways?
There was a problem hiding this comment.
I now changed the description to this:
This function raises :exc:`SyntaxError` or :exc:`ValueError` if the compiled source is invalid.
To summarize some things about compile's interface:
- If the source contains null bytes (whether it's a
bytesobject or astrobject),SyntaxErroris raised since Python 3.11, and notValueError, contrary to current documentation. - If the source is an
ast.Module, null bytes and surrogates seem to be allowed (but lead to bizzarre behaviour) - If the source is a
bytes,SyntaxErrorwill be raise if the source is encoded incorrectly - If the source is valid,
-OOis not enabled, and any docstings would have a surrogate,UnicodeDecodeError(subclass ofValueError) is raised
It seems like the error conditions are a bit complicated and may change from version to version, so maybe it's good to keep it a bit opaque? Meaning: if you want to use this function and you expect the input to sometimes be invalid, prepare to catch both SyntaxError and ValueError
decorator-factory
force-pushed
the
df-compile-surrogatedoc
branch
from
1b5431a to
eeb1ceb
Compare
hugovk
added
needs backport to 3.13
|
Thanks @decorator-factory for the PR, and @hugovk for merging it 🌮🎉.. I'm working now to backport this PR to: 3.13, 3.14. |
|
Sorry, @decorator-factory and @hugovk, I could not cleanly backport this to |
|
Sorry, @decorator-factory and @hugovk, I could not cleanly backport this to |
|
@decorator-factory Would you like to have a go at backporting? See instructions above. |
|
@hugovk Yes, I'll try to do that. Backports are just pull requests to branches other than main, correct? And in this case, miss Islington failed to generate a PR due to conflicts, and now I just need to make those PRs to 3.13 and 3.14 branches? |
…ring adjustments in 3.13 (pythonGH-142413) (cherry picked from commit f6b6a99) Co-authored-by: decorator-factory <42166884+decorator-factory@users.noreply.github.com>
|
GH-142689 is a backport of this pull request to the 3.14 branch. |
|
GH-142690 is a backport of this pull request to the 3.13 branch. |
Exactly so. Thanks, merged! |
3 participants
See issue #142411 for details.
Document that
compilecan now raiseUnicodeDecodeErrorFix paragraph about docstring indentation in the tutorial. The REPL output was fixed in Docs: Fix indent in
controlflow.rst#134008, but the preceding paragraph hasn't been adjustedI believe this should be backported to 3.13 and 3.14, not sure how that works.
📚 Documentation preview 📚: https://cpython-previews--142413.org.readthedocs.build/