d_initial is not used by the parser, so no point generating it

[tyomitch]

No description provided.

[pablogsal]

I think is better to pass the index of the start symbol into the list of dfas down to the grammar and use that when generating the parser input files.

[bedevere-bot]

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

[tyomitch]

I think is better to pass the index of the start symbol into the list of dfas down to the grammar and use that when generating the parser input files.

Could you please clarify?

[pablogsal]

Could you please clarify?

Yup, make pgen generate it from the position of the initial symbol. The initial dfa symbol is placed in the grammar in:

https://github.com/python/cpython/blob/e12f56075edc5704c9cfbf69d268474328642fa3/Parser/pgen/pgen.py#L53

The value has to be the position of that symbol in the list of dfas. It will be 0 as single_input is always the first DFA in the list of dfas, but doing programmatically will protect in case the position changes in the future. Basically something on the lines of (for illustrative purposes):

change:

"0, {n_states}, states_{dfa_index},\n".format(

to

start_index = list(self.dfas.keys()).index(self.start)
"{start}, {n_states}, states_{dfa_index},\n".format(start=start_index

[tyomitch]

Now I see what you mean, but d_initial is not about the initial DFA of the grammar, it's about the initial state of a DFA: each DFA has its own d_initial, and they're all zeroes.

[pablogsal]

Now I see what you mean, but d_initial is not about the initial DFA of the grammar, it's about the initial state of a DFA: each DFA has its own d_initial, and they're all zeroes.

Oh, right! Sorry, I have mistaken this with the start rule in the old pgen:

cpython/Parser/pgen.c

Line 658 in 79fbcc5

g = newgrammar(gr->gr_nfa[0]->nf_type);

But as you indicate, d_initial has been initialized always to 0:

cpython/Parser/pgen.c

Line 642 in 79fbcc5

d->d_initial = 0;

[pablogsal]

@tyomitch Can you rebase to regenerate the graminit.c file.

[tyomitch]

@tyomitch Can you rebase to regenerate the graminit.c file.

Done!

[pablogsal]

Thanks @tyomitch for the contribution! :)