The elements in `Grammar/python.gram` stripped by `peg_lexer.py` still shows in generated PO files

[mattwang44]

My original question has answered by Stan. I leave it here as a record:

Click to expand

I'm not sure if it's done deliberately, but it seems odd to me and I can't find out why it happens so far.

---

The reference/grammar.rst includes Grammar/python.gram

cpython/Doc/reference/grammar.rst

Lines 18 to 19 in 78a50ee

	.. literalinclude:: ../../Grammar/python.gram
	:language: peg

and Grammar/python.gram has a section for invalid rules at the bottom.

cpython/Grammar/python.gram

Lines 1193 to 1202 in 78a50ee

	# ========================= END OF THE GRAMMAR ===========================
	# ========================= START OF INVALID RULES =======================
	# From here on, there are rules for invalid syntax with specialised error messages
	invalid_arguments:
	| ((','.(starred_expression | ( assignment_expression | expression !':=') !'=')+ ',' kwargs) | kwargs) a=',' ','.(starred_expression !'=')+ {

However, the content after START OF INVALID RULES is missing on the page

snapshot

I checked the previous version and found out the doc has this issue since 3.11 (after PEG parser is introduced).

---

Other info:
I translated this page into Traditional Chinese (zh-tw) recently and somehow it shows correctly on the translated page

snapshot

So here's the revised issue description:

Some elements in Grammar/python.gram are stripped by peg_lexer.py, and hence they're hidden on the reference/grammar docs. However, the full content of Grammar/python.gram is still remained in the generated PO files (e.g. the one in zh-tw).

[StanFromIreland]

This is intentional, as the introduction states:

The version here omits details related to code generation and error recovery.

cpython/Doc/tools/extensions/lexers/peg_lexer.py

Lines 8 to 16 in e4d3c83

	This lexer strips the following elements from the grammar:
	- Meta-tags
	- Variable assignments
	- Actions
	- Lookaheads
	- Rule types
	- Rule options
	- Rules named `invalid_*` or `incorrect_*`

Why this is not done for translations is a bug.

[mattwang44]

ok thanks! I'll revise the issue description in-place then.