eggex 'a'{N *} crashes, needs a proper error

[glyh]

$ = / 'a'{N *} / 
ysh: _gen/bin/oils_for_unix.mycpp.cc:50177: syntax_asdl::re_repeat_t* expr_to_ast::Transformer::_RepeatOp(pnode::PNode*): Assertion `0' failed.
fish: Job 1, 'ysh' terminated by signal SIGABRT (Abort)

Oils 0.21.0 https://www.oilshell.org/

git commit = cc2f683

[glyh]

Looks like this is simply unimplemented.

oil/frontend/syntax.asdl

Line 646 in 3aa3839

# Haven't implemented the modifier, e.g. x{+ P}

Also representing something like "term repeating 1 ~ 4 times non-greedily" doesn't seem to have an elegant solution for now.

[glyh]

If there'll be a refactor I guess it would be nice to have something like https://docs.raku.org/language/grammars

[andychu]

In Oils, we use libc regcomp() / regexec() which only supports POSIX ERE, which doesn't have greedy like Perl/Python

So this should definitely not crash, but it will still be a parse error.

Or we can parse it, and allow people to export to the AST, so they can translate it to Python/Perl syntax

• Write eggex translators to POSIX BRE and PCRE / Perl / Python #488

[glyh]

I think at least user should be able to use eggex to match in a if statement.

[andychu]

This is now a proper parse error

Not sure what you mean, the problem here is that Oils uses libc, which supports POSIX ERE, not Perl-style repetition

'a'{N *} is supposed to be a*? -- nongreedy repetition

In theory eggex can be translated to Perl-style engines, but we don't have that

[andychu]

Released - https://www.oilshell.org/blog/2024/06/release-0.22.0.html