Hi, judging these docs and multiple changelog entries mentioning `%define api.pure` as a replacement for `%pure-parser`, and the latter being turned to wrapper on the former at some point, etc.;
I would expect that tests/input.at recommend, for `%pure_parser`, `%define api.pure` optionally `%define api.pure full`, instead of `%pure-parser`? Around line 2372, that’s the test side. I don't know where is the non-test code side of that, and I don't feel that much momentum to find out and propose a patch. Cheers, > On 8 Dec 2018, at 07:19, Akim Demaille <a...@lrde.epita.fr> wrote: > > Hi! > >> Le 7 déc. 2018 à 20:30, Uxio Prego <uxio.pr...@gmail.com> a écrit : >> >> I don't know of this division of parsers in pure and impure. >> >> Are pure approaches like GLR while LALR(k) are impure? > > No, it's unrelated to the parsing technology, it's only a question > of API: whether you exchange information with the scanner via global > variables, or additional arguments. > > https://www.gnu.org/software/bison/manual/html_node/Pure-Decl.html