Added rule references to AST parser

This commit is contained in:
Eric-Paul Ickhorn 2023-11-28 16:01:46 +01:00
parent 138e361c54
commit d9aa79239d
2 changed files with 53 additions and 18 deletions

View File

@ -3,6 +3,7 @@
#define PARCEL_AST_H
#include <utility.h>
#include <tokenizer.h>
typedef struct pac_ast pac_ast_s;
typedef struct pac_ast_rule pac_ast_rule_s;
@ -58,14 +59,26 @@ typedef enum
} pac_outline_e;
typedef enum
{
PAC_AST_ITEM_INVALID = 0x00,
PAC_AST_ITEM_REFERENCE,
PAC_AST_ITEM_LITERAL,
PAC_AST_ITEM_OUTLINE
} pac_ast_item_e;
struct pac_ast_item
{
bool_t is_literal;
pac_ast_item_e type;
union pac_item_data
{
pac_ast_literal_s literal;
pac_outline_e outline;
pac_ast_reference_s reference;
} data;
};
pac_ast_s pac_grow_ast (pac_tlist_s tokens);
#endif // PARCEL_AST_H

View File

@ -6,7 +6,7 @@
#include <stdio.h>
#include <stdlib.h>
#define CURRENT_STRING (&tlist->source[tlist->tokens[tlist->cursor].offset])
#define CURRENT_STRING &tlist->source[tlist->tokens[tlist->cursor].offset]
#define CURRENT (tlist->tokens[tlist->cursor])
#define SKIP_TOKEN ++tlist->cursor
#define TOKEN_AT(index) (tlist->tokens[index])
@ -14,18 +14,44 @@
i32_t pac_grow_reference(pac_tlist_s *tlist, pac_ast_reference_s *reference)
{
if(CURRENT.type != PAC_TOKEN_SIGN_OPEN_TAG)
{
return -1;
}
SKIP_TOKEN;
if(CURRENT.type != PAC_TOKEN_WORD)
{
return -1;
}
usz_t len_name = CURRENT.length;
char *name = CURRENT_STRING;
SKIP_TOKEN;
if(CURRENT.type != PAC_TOKEN_SIGN_CLOSE_TAG)
{
return -1;
}
SKIP_TOKEN;
reference->len_name = len_name;
reference->name = malloc(len_name+1);
memcpy(reference->name, name, len_name);
reference->name[len_name] = 0x00;
return 3;
}
i32_t pac_grow_item(pac_tlist_s *tlist, pac_ast_item_s *item)
{
item->is_literal = FALSE;
item->type = PAC_AST_ITEM_INVALID;
memset(item, 0x00, sizeof(pac_ast_item_s));
switch(CURRENT.type)
{
case PAC_TOKEN_LIT_STRING:
{
item->is_literal = TRUE;
item->type = PAC_AST_ITEM_LITERAL;
item->data.literal.length = CURRENT.length;
item->data.literal.string = malloc(item->data.literal.length + 1);
memcpy(item->data.literal.string, CURRENT_STRING, CURRENT.length);
@ -33,14 +59,22 @@ i32_t pac_grow_item(pac_tlist_s *tlist, pac_ast_item_s *item)
SKIP_TOKEN;
} return 1;
case PAC_TOKEN_SIGN_OPEN_TAG:
{
item->type = PAC_AST_ITEM_REFERENCE;
return pac_grow_reference(tlist, &item->data.reference);
}
case PAC_TOKEN_KEYWORD_WORD:
{
item->type = PAC_AST_ITEM_OUTLINE;
item->data.outline = PAC_OUTLINE_WORD;
SKIP_TOKEN;
} return 1;
case PAC_TOKEN_KEYWORD_INTEGER:
{
item->type = PAC_AST_ITEM_OUTLINE;
item->data.outline = PAC_OUTLINE_INTEGER;
SKIP_TOKEN;
} return 1;
@ -136,7 +170,7 @@ i32_t pac_grow_rule(pac_tlist_s *tlist, pac_ast_rule_s *rule)
if(success < 0)
{
printf("Failed parsing variant %u of rule '%s'. ");
printf("Failed parsing variant %u of rule '%d'. ", rule->num_variants);
while(!END_REACHED)
{
if(CURRENT.type == PAC_TOKEN_SIGN_VERTICAL_BAR)
@ -147,11 +181,11 @@ i32_t pac_grow_rule(pac_tlist_s *tlist, pac_ast_rule_s *rule)
if(CURRENT.type == PAC_TOKEN_SIGN_SEMICOLON)
{
printf("Continuing with next rule.\n");
SKIP_TOKEN;
return 2;
}
SKIP_TOKEN;
}
continue;
}
if(CURRENT.type == PAC_TOKEN_SIGN_SEMICOLON)
@ -192,15 +226,3 @@ pac_ast_s pac_grow_ast(pac_tlist_s tokens)
}
return ast;
}
pac_grammar_s pac_convert_grammar(char *source)
{
usz_t len_source = strlen(source);
pac_tlist_s tokens = pac_tokenize_grammar(source, len_source);
pac_display_tlist(tokens);
pac_ast_s ast = pac_grow_ast(tokens);
pac_grammar_s grammar;
return grammar;
}