Broken grammar - Githubissues

dolik-rce commented 2 years ago

When I ask pegof to optimize the the grammar shown above and pass the result to packcc I'm getting this error:

packcc -o lpegrex lpegrex.peg 
packcc: lpegrex.peg:45:1: Illegal rule syntax
packcc: lpegrex.peg:45:21: Invalid directive
packcc: lpegrex.peg:45:22: Illegal rule syntax
packcc: lpegrex.peg:45:53: Invalid directive
packcc: lpegrex.peg:45:54: Illegal rule syntax
packcc: lpegrex.peg:45:96: Invalid directive
packcc: lpegrex.peg:45:99: Illegal rule syntax
packcc: lpegrex.peg:25:7: No definition of rule 'class'

Optimized grammar:

%prefix "lpeg_re"

%header {
    typedef struct  {
    int doOutput;
    } lpeg_re_st;
}

%auxil "lpeg_re_st *"

pattern <- S (definition+ / alternative) !. { printf("\n\n"); }

alternative <- prefix* (<"/"> { printf("\n\t%s ", $1); } S prefix*)*

prefix
    <- "&" S prefix
    / "!" S prefix
    / suffix

suffix <- primary S ((<[*+?]> { printf(" %s ", $1); } / "^" [+-]? [0-9]+ / "->" { auxil->doOutput=0; } S (string / "{}" / name / [0-9]+) { auxil->doOutput=1; } / "=>" { auxil->doOutput=0; } S name) S)* { auxil->doOutput=1; }

primary
    <- "(" { printf(" ( "); } S (definition+ / alternative) ")" { printf(" ) "); }
    / (string / keyword)
    / class
    / "%" name
    / "{:" (name ":")? S (definition+ / alternative) ":}"
    / "=" name
    / "@" S (definition+ / alternative)
    / "{}"
    / "{~" S (definition+ / alternative) "~}"
    / "{|" S (definition+ / alternative) "|}"
    / "{" S (definition+ / alternative) "}"
    / "~?"
    / "~>" S ("foldleft" / "foldright" / "rfoldleft" / "rfoldright")
    / "$" (string / name / [0-9]+)
    / "."
    / name S !(":" S name / ("<--" / "<==" / "<-|" / "<-"))
    / "<" name ">"
    / "^" name
    / "%{" S name S "}"

definition <-  { printf("\n\n"); } S (":" S name S)? ("<--" / "<==" / "<-|" / "<-") S (definition+ / alternative)

class <- <"[" "^"? "%" name / . "-" [^]] / . (!"]" "%" name / . "-" [^]] / .)* "]"> { printf(" %s ", $1); }

S <- ([\t\n\r f] / "--" [^\n\r]*)*

name <- <[A-Z_a-z] ([0-9A-Z_a-z] / "-" !">")*> { if(auxil->doOutput) printf(" %s ", $1); }

string <- <"\"" [^"]* "\"" / "'" [^']* "'"> { if(auxil->doOutput) printf(" %s ", $1); }

keyword <- "`" <[^`]+> "`" { printf(" '%s' ", $1); }

%%
int main() {
    lpeg_re_st state;
    state.doOutput = 1;
    lpeg_re_context_t *ctx = lpeg_re_create(&state);
    while (lpeg_re_parse(ctx, NULL));
    lpeg_re_destroy(ctx);
    return 0;
}

Originally posted by @mingodad in https://github.com/dolik-rce/pegof/issues/1#issuecomment-1103707756

dolik-rce / pegof

Broken grammar #6