update: added folding into expansion node if possible

This commit is contained in:
maix0 2024-10-03 22:52:01 +02:00
parent 09dbd2de91
commit 142ac9c9e1
11 changed files with 278 additions and 41 deletions

View file

@ -6,7 +6,7 @@
/* By: maiboyer <maiboyer@student.42.fr> +#+ +:+ +#+ */
/* +#+#+#+#+#+ +#+ */
/* Created: 2024/10/02 18:41:16 by maiboyer #+# #+# */
/* Updated: 2024/10/03 21:32:14 by maiboyer ### ########.fr */
/* Updated: 2024/10/03 22:41:08 by maiboyer ### ########.fr */
/* */
/* ************************************************************************** */
@ -33,7 +33,7 @@
/// double LCARRET into DLCARET, ...)
/// - create EXPENSION token when DOLLAR and NQUOTE follow eachother,
/// maybe leaving some stuff after
/// - parse double quote string to see expansion in them, creating a meta
/// - parse double quote string to see expansion in them, creating a meta
/// token consisting of the pieces
// here is the signature easily accessible:
@ -43,7 +43,8 @@
// bellow is some function to apply all the different passes !
static const struct s_ts_pass_def g_ts_passes[] = {
{do_fuck_all, "does nothing lol"},
{ts_double_string_pass, "double string parser"},
{ts_fold_expension, "fold expansion"},
};
t_error ts_apply_passes(t_vec_token ts, t_vec_token *out)
@ -65,3 +66,28 @@ t_error ts_apply_passes(t_vec_token ts, t_vec_token *out)
}
return (*out = ts, NO_ERROR);
}
static const struct s_ts_pass_def g_ts_dq_passes[] = {
{ts_do_fuck_all, "does nothing lol"},
{ts_fold_expension, "fold expansion"},
};
t_error ts_dq_apply_passes(t_vec_token ts, t_vec_token *out)
{
t_usize i;
t_vec_token next;
i = 0;
while (i < sizeof(g_ts_dq_passes) / sizeof(g_ts_dq_passes[0]))
{
if (g_ts_dq_passes[i].fn == NULL)
return (vec_token_free(ts), ERROR);
if ((g_ts_dq_passes[i].fn)(ts, &next))
return (me_eprintf("failed on '%s' dq token pass\n", g_ts_dq_passes[i].name), ERROR);
else
me_printf("Applied '%s' dq_pass\n", g_ts_dq_passes[i].name);
ts = next;
i++;
}
return (*out = ts, NO_ERROR);
}

View file

@ -0,0 +1,98 @@
/* ************************************************************************** */
/* */
/* ::: :::::::: */
/* double_quote_parsing.c :+: :+: :+: */
/* +:+ +:+ +:+ */
/* By: maiboyer <maiboyer@student.42.fr> +#+ +:+ +#+ */
/* +#+#+#+#+#+ +#+ */
/* Created: 2024/10/02 19:04:32 by maiboyer #+# #+# */
/* Updated: 2024/10/03 22:43:27 by maiboyer ### ########.fr */
/* */
/* ************************************************************************** */
#include "me/string/string.h"
#include "parser/passes.h"
#include "me/types.h"
#include "me/vec/vec_token.h"
#include "parser/token.h"
#include "me/char/char.h"
void push_token_and_create_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s);
void push_token_and_set_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s);
t_error _parse_dquote_inner(t_token dquote, t_vec_token *append)
{
t_token ctok;
t_token out;
t_usize i;
char c;
out = token_new_meta(TOK_DQUOTE);
i = 0;
ctok = token_new_none();
while (dquote.string.buf[i] != '\0')
{
c = dquote.string.buf[i++];
if (me_isspace(c))
{
if (ctok.type == TOK_NONE)
ctok = token_new(TOK_WHITESPACE);
if (ctok.type != TOK_WHITESPACE)
{
vec_token_push(&out.subtokens, ctok);
ctok = token_new(TOK_WHITESPACE);
}
string_push_char(&ctok.string, c);
}
else if (c == '$')
push_token_and_create_new(&out.subtokens, &ctok, TOK_DOLLAR, "$");
else
{
if (ctok.type == TOK_NONE)
ctok = token_new(TOK_NQUOTE);
if (ctok.type != TOK_NQUOTE)
{
vec_token_push(&out.subtokens, ctok);
ctok = token_new(TOK_NQUOTE);
}
string_push_char(&ctok.string, c);
}
};
if (ctok.type != TOK_NONE)
vec_token_push(&out.subtokens, ctok);
if (ts_dq_apply_passes(out.subtokens, &out.subtokens))
return (ERROR);
return (vec_token_push(append, out), NO_ERROR);
}
/// There is a few rules the rest of the tokenizer machinery assumes
/// theses function follows:
/// - the input vec WILL be freed when the function return, even in
/// case of error
/// - the output vector isn't populated if the function returns an error,
/// thus it shouldn't be freed in case of error
/// - the output tokens may not be direct copy of the input tokens,
/// but need to be cloned (different allocations for stuff)
t_error ts_double_string_pass(t_vec_token input, t_vec_token *output)
{
t_vec_token out;
t_usize i;
i = 0;
out = vec_token_new(input.len, token_free);
while (i < input.len)
{
if (input.buffer[i].type == TOK_DQUOTE)
{
if (_parse_dquote_inner(input.buffer[i], &out))
return (vec_token_free(input), ERROR);
}
else
vec_token_push(&out, token_clone(&input.buffer[i]));
i++;
}
vec_token_free(input);
return (*output = out, NO_ERROR);
}

View file

@ -0,0 +1,54 @@
/* ************************************************************************** */
/* */
/* ::: :::::::: */
/* fold_expansion.c :+: :+: :+: */
/* +:+ +:+ +:+ */
/* By: maiboyer <maiboyer@student.42.fr> +#+ +:+ +#+ */
/* +#+#+#+#+#+ +#+ */
/* Created: 2024/10/02 19:04:32 by maiboyer #+# #+# */
/* Updated: 2024/10/03 22:50:40 by maiboyer ### ########.fr */
/* */
/* ************************************************************************** */
#include "me/types.h"
#include "me/vec/vec_token.h"
#include "parser/passes.h"
#include "parser/token.h"
/// This is a sample pass
///
/// There is a few rules the rest of the tokenizer machinery assumes
/// theses function follows:
/// - the input vec WILL be freed when the function return, even in
/// case of error
/// - the output vector isn't populated if the function returns an error,
/// thus it shouldn't be freed in case of error
/// - the output tokens may not be direct copy of the input tokens,
/// but need to be cloned (different allocations for stuff)
t_error ts_fold_expension(t_vec_token input, t_vec_token *output)
{
t_vec_token out;
t_usize i;
t_token tmp;
i = 0;
out = vec_token_new(input.len, token_free);
while (i < input.len)
{
if (i + 1 >= input.len)
vec_token_push(&out, token_clone(&input.buffer[i]));
else if (input.buffer[i].type == TOK_DOLLAR \
&& (input.buffer[i + 1].type == TOK_DOLLAR \
|| input.buffer[i + 1].type == TOK_NQUOTE))
{
tmp = token_clone(&input.buffer[++i]);
tmp.type= TOK_EXPENSION;
vec_token_push(&out, tmp);
}
else
vec_token_push(&out, token_clone(&input.buffer[i]));
i++;
}
vec_token_free(input);
return (*output = out, NO_ERROR);
}

View file

@ -6,7 +6,7 @@
/* By: maiboyer <maiboyer@student.42.fr> +#+ +:+ +#+ */
/* +#+#+#+#+#+ +#+ */
/* Created: 2024/10/02 19:04:32 by maiboyer #+# #+# */
/* Updated: 2024/10/03 21:37:04 by maiboyer ### ########.fr */
/* Updated: 2024/10/03 22:23:40 by maiboyer ### ########.fr */
/* */
/* ************************************************************************** */
@ -25,7 +25,7 @@
/// thus it shouldn't be freed in case of error
/// - the output tokens may not be direct copy of the input tokens,
/// but need to be cloned (different allocations for stuff)
t_error do_fuck_all(t_vec_token input, t_vec_token *output)
t_error ts_do_fuck_all(t_vec_token input, t_vec_token *output)
{
t_vec_token out;
t_usize i;

View file

@ -6,7 +6,7 @@
/* By: maiboyer <maiboyer@student.42.fr> +#+ +:+ +#+ */
/* +#+#+#+#+#+ +#+ */
/* Created: 2024/09/30 19:39:39 by maiboyer #+# #+# */
/* Updated: 2024/10/03 21:31:57 by maiboyer ### ########.fr */
/* Updated: 2024/10/03 22:44:57 by maiboyer ### ########.fr */
/* */
/* ************************************************************************** */
@ -16,26 +16,10 @@
#include "me/vec/vec_token.h"
#include "parser/token.h"
static void push_token_and_create_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s)
{
t_token tmp;
if (tok->type != TOK_NONE)
vec_token_push(tokens, *tok);
*tok = token_new_none();
tmp = token_new(ttype);
string_push(&tmp.string, s);
vec_token_push(tokens, tmp);
}
static void push_token_and_set_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s)
{
if (tok->type != TOK_NONE)
vec_token_push(tokens, *tok);
*tok = token_new(ttype);
string_push(&tok->string, s);
}
void push_token_and_create_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s);
void push_token_and_set_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s);
static void handle_quote(t_vec_token *ret, char chr, t_token *tok, char *quote)
{
@ -68,7 +52,7 @@ static void handle_noquote(t_vec_token *ret, char chr, t_token *tok, char *quote
else if (chr == ')')
push_token_and_create_new(ret, tok, TOK_RPAREN, ")");
else if (chr == ';')
push_token_and_create_new(ret, tok, TOK_RPAREN, ";");
push_token_and_create_new(ret, tok, TOK_SEMICOLON, ";");
else if (me_isspace(chr))
push_token_and_create_new(ret, tok, TOK_WHITESPACE, " ");
else

View file

@ -0,0 +1,46 @@
/* ************************************************************************** */
/* */
/* ::: :::::::: */
/* tokenizer_utils.c :+: :+: :+: */
/* +:+ +:+ +:+ */
/* By: maiboyer <maiboyer@student.42.fr> +#+ +:+ +#+ */
/* +#+#+#+#+#+ +#+ */
/* Created: 2024/10/03 22:07:25 by maiboyer #+# #+# */
/* Updated: 2024/10/03 22:08:14 by maiboyer ### ########.fr */
/* */
/* ************************************************************************** */
#include "me/char/char.h"
#include "me/string/string.h"
#include "me/types.h"
#include "me/vec/vec_token.h"
#include "parser/token.h"
// These *should* be stable-ish, just copy the definition into the files to
// use them
void push_token_and_create_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s);
void push_token_and_set_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s);
void push_token_and_create_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s)
{
t_token tmp;
if (tok->type != TOK_NONE)
vec_token_push(tokens, *tok);
*tok = token_new_none();
tmp = token_new(ttype);
string_push(&tmp.string, s);
vec_token_push(tokens, tmp);
}
void push_token_and_set_new(\
t_vec_token *tokens, t_token *tok, enum e_token ttype, t_const_str s)
{
if (tok->type != TOK_NONE)
vec_token_push(tokens, *tok);
*tok = token_new(ttype);
string_push(&tok->string, s);
}