spl/stage1/splc0.c

/* splc0.c — SPL compiler CLI (stage 1, 引导用)
 *
 *   splc0 --dump tokens|ast|all <file>    dump 前端产物
 *   splc0 <in> <out>                      编译 (阶段 B 实现)
 */
#define __SCC_LOG_IMPL_IMPORT_SRC__
#include "../stage0/include/utils.h"

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include "spl_ast.h"
#include "spl_ast2ir.h"
#include "spl_lexer.h"
#include "spl_sema.h"
#include "spl_tok.h"

static char *read_file(const char *path, long *out_len) {
    FILE *f = fopen(path, "rb");
    if (!f) {
        perror("fopen");
        return NULL;
    }
    fseek(f, 0, SEEK_END);
    long len = ftell(f);
    fseek(f, 0, SEEK_SET);
    char *buf = malloc((size_t)len + 1);
    if (!buf) {
        fclose(f);
        return NULL;
    }
    fread(buf, 1, (size_t)len, f);
    fclose(f);
    buf[len] = '\0';
    *out_len = len;
    return buf;
}

static const char *const tok_type_names[] = {
#define X(name, enum_name, dummy) #enum_name,
    KEYWORD_TABLE
#undef X
#define X(name, enum_name, dummy) #enum_name,
        TOKEN_TABLE
#undef X
};

static void dump_tokens(const char *src, const char *fname) {
    spl_tok_vec_t toks = spl_lex(src, fname);
    printf("tokens got (%zu)\n", toks.size);
    for (usize i = 0; i < toks.size; i++) {
        const spl_tok_t *t = &toks.data[i];
        printf("[%s] %.*s (%zu:%zu)\n", tok_type_names[t->type], (int)t->len, t->lexeme, t->line,
               t->col);
    }
    vec_free(toks);
}

static void dump_ast(const char *src, const char *fname) {
    spl_tok_vec_t toks = spl_lex(src, fname);
    spl_ast_t ast;
    spl_ast_init(&ast, &toks);
    spl_ast_prase(&ast);
    spl_ast_valid(&ast);
    spl_ast_dump(&ast, ast.root);
    spl_ast_drop(&ast);
}

static void dump_sema(const char *src, const char *fname) {
    spl_tok_vec_t toks = spl_lex(src, fname);
    spl_ast_t ast;
    spl_ast_init(&ast, &toks);
    spl_ast_prase(&ast);
    spl_ast_valid(&ast);
    spl_sema_t sema;
    spl_sema_init(&sema);
    sema.ast = &ast;
    spl_sema_run(&sema);
    spl_sema_check(&sema);
    printf("Sema  root_scope=%zu  scopes=%zu  errors=%d\n", sema.root_scope, sema.scopes.size,
           sema.error_count);
    for (usize i = 0; i < sema.scopes.size; i++) {
        printf("scope[%zu] parent=%zu\n", i, sema.scopes.data[i].parent);
        map_for(sema.scopes.data[i].symbols, mi) {
            printf("  %s -> def#%zu\n", sema.scopes.data[i].symbols.data[mi].key,
                   sema.scopes.data[i].symbols.data[mi].val);
        }
    }
    printf("TypeTable:\n");
    for (usize i = 0; i < sema.type.type_table.size; i++) {
        printf("  id#%zu type=", i);
        spl_type_pure_dump(&sema.type, i);
        printf("\n");
    }
    printf("DefTable:\n");
    for (usize i = 0; i < sema.type.def_table.size; i++) {
        printf("  def#%zu ", i);
        spl_type_def_dump(&sema.type, i);
        printf("\n");
    }
    spl_sema_drop(&sema);
    spl_ast_drop(&ast);
}

static void dump_ir(const char *src, const char *fname) {
    spl_tok_vec_t toks = spl_lex(src, fname);
    spl_ast_t ast;
    spl_ast_init(&ast, &toks);
    spl_ast_prase(&ast);
    spl_ast_valid(&ast);
    spl_sema_t sema;
    spl_sema_init(&sema);
    sema.ast = &ast;
    spl_sema_run(&sema);
    spl_sema_check(&sema);
    if (sema.error_count) {
        printf("sema errors=%d, skip IR\n", sema.error_count);
        spl_sema_drop(&sema);
        spl_ast_drop(&ast);
        return;
    }
    spl_ast2ir_t a2ir;
    spl_ast2ir_init(&a2ir, &sema);
    spl_ast2ir_run(&a2ir);
    if (a2ir.err_count)
        printf("ast2ir errors=%d\n", a2ir.err_count);
    spl_ir_dump(&a2ir.ir, &sema.type);
    spl_ast2ir_drop(&a2ir);
    spl_sema_drop(&sema);
    spl_ast_drop(&ast);
}

static int cmd_dump(const char *flags, const char *path) {
    long len;
    char *src = read_file(path, &len);
    if (!src)
        return 1;
    int do_tokens = strstr(flags, "tokens") != NULL || strcmp(flags, "all") == 0;
    int do_ast = strstr(flags, "ast") != NULL || strcmp(flags, "all") == 0;
    int do_sema = strstr(flags, "sema") != NULL || strcmp(flags, "all") == 0;
    int do_ir = strstr(flags, "ir") != NULL || strcmp(flags, "all") == 0;
    if (do_tokens)
        dump_tokens(src, path);
    if (do_ast)
        dump_ast(src, path);
    if (do_sema)
        dump_sema(src, path);
    if (do_ir)
        dump_ir(src, path);
    free(src);
    return 0;
}

int main(int argc, char **argv) {
    if (argc < 2) {
        LOG_FATAL("Usage: splc0 [--dump <flags>] <in> [out]\n");
        return 1;
    }
    if (strcmp(argv[1], "--help") == 0 || strcmp(argv[1], "-h") == 0) {
        LOG_INFO("splc0 <in> <out>              compile (.spl -> .sir, 阶段 B)\n");
        LOG_INFO("splc0 --dump <flags> <file>   dump: tokens,ast,sema,ir,all\n");
        return 0;
    }
    int argi = 1;
    if (argi >= argc) {
        LOG_FATAL("Usage: splc0 [--dump <flags>] <in> [out]\n");
        return 1;
    }
    if (strcmp(argv[argi], "--dump") == 0) {
        if (argc < argi + 3) {
            LOG_INFO("splc0: --dump need <flags> <file>\n");
            return 1;
        }
        return cmd_dump(argv[argi + 1], argv[argi + 2]);
    }
    LOG_FATAL("splc0: compile todo\n");
    return 1;
}