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_ir2vm.h"
#include "spl_lexer.h"
#include "spl_sema.h"
#include "spl_tok.h"

typedef enum {
    DUMP_NONE,
    DUMP_TOKEN,
    DUMP_AST,
    DUMP_SEMA,
    DUMP_IR,
} dump_t;

static char *read_file(const char *path, long *out_len) {
    FILE *f = fopen(path, "rb");
    if (!f) {
        fprintf(stderr, "path `%s` ", path);
        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(spl_tok_vec_t toks) {
    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 int compile_spl(const char *src, const char *fname, const char *outpath, int gen_debug,
                       dump_t dump) {
    spl_tok_vec_t toks = spl_lex(src, fname);
    if (dump == DUMP_TOKEN) {
        dump_tokens(toks);
        // memory leak
        return 0;
    }
    spl_ast_t ast;
    spl_ast_init(&ast, &toks);
    spl_ast_prase(&ast);
    if (ast.parsed < 0) {
        printf("parse failed, no output\n");
        // memory leak
        return 1;
    }
    // spl_ast_valid(&ast); TODO
    if (dump == DUMP_AST) {
        spl_ast_dump(&ast, ast.root);
        // memory leak
        return 0;
    }
    spl_sema_t sema;
    spl_type_t type;
    spl_scope_t scope;
    spl_type_init(&type);
    spl_scope_init(&scope);
    spl_sema_init(&sema, &ast, &type, &scope);
    sema.ast = &ast;
    spl_sema_run(&sema);
    if (sema.error_count) {
        printf("sema errors=%d, no output\n", sema.error_count);
        // memory leak
        return 1;
    }
    if (dump == DUMP_SEMA) {
        spl_sema_dump(&sema);
        spl_ast_dump(&ast, ast.root);
        // memory leak
        return 0;
    }

    spl_ast2ir_t a2ir;
    spl_ir_builder_t ir_builder;
    spl_ir_builder_init(&ir_builder);
    spl_ast2ir_init(&a2ir, &ir_builder, &sema);
    spl_ast2ir_run(&a2ir);
    if (a2ir.err_count) {
        printf("ast2ir errors=%d, no output\n", a2ir.err_count);
        spl_ast2ir_drop(&a2ir);
        spl_sema_drop(&sema);
        spl_ast_drop(&ast);
        return 1;
    }
    if (dump == DUMP_IR) {
        spl_ir_dump(&ir_builder.ir, &type);
        // memory leak
        return 0;
    }

    // spl_ir2vm_t ir2vm;
    // spl_ir2vm_init(&ir2vm, &a2ir.ir, &sema.type);
    // int rc = spl_ir2vm_run(&ir2vm, outpath);
    // if (gen_debug && rc == 0)
    //     gen_debug_map(outpath, &ast, &a2ir.ir, &ir2vm);
    // spl_ir2vm_drop(&ir2vm);
    // spl_ast2ir_drop(&a2ir);
    // spl_sema_drop(&sema);
    // spl_ast_drop(&ast);
    // return rc ? 1 : 0;
    return 0;
}

int main(int argc, char **argv) {
    if (argc < 2) {
        LOG_FATAL("Usage: splc0 [--dump <flags>] <in> [out]");
        return 1;
    }
    if (strcmp(argv[1], "--help") == 0 || strcmp(argv[1], "-h") == 0) {
        LOG_INFO("splc0 <in> <out>              compile (.spl -> .sir, stage B)");
        LOG_INFO("splc0 --dump <flags> <file>   dump: lex,ast,sema,ir");
        return 0;
    }
    int argi = 1;
    if (argi >= argc) {
        LOG_FATAL("Usage: splc0 [--dump <flags>] <in> [out]");
        return 1;
    }

    dump_t dump = DUMP_NONE;
    if (strcmp(argv[argi], "--dump") == 0) {
        if (argc < argi + 3) {
            LOG_INFO("splc0: --dump need <flags> <file>");
            return 1;
        }

        const char *flags = argv[argi + 1];
        if (strstr(flags, "lex") != NULL) {
            dump = DUMP_TOKEN;
        }
        if (strstr(flags, "ast") != NULL) {
            dump = DUMP_AST;
        }
        if (strstr(flags, "sema") != NULL) {
            dump = DUMP_SEMA;
        }
        if (strstr(flags, "ir") != NULL) {
            dump = DUMP_IR;
        }
        argi += 2;
    }
    /* splc0 [-g] <in> <out> */
    if (argc < argi + 2) {
        LOG_FATAL("Usage: splc0 <in> <out> [-g]  or  splc0 --dump <flags> <file>\n");
        return 1;
    }
    int gen_debug = 0;
    if (strcmp(argv[argi], "-g") == 0) {
        gen_debug = 1;
        argi++;
    }
    if (argc < argi + 2) {
        LOG_FATAL("Usage: splc0 <in> <out> [-g]  or  splc0 --dump <flags> <file>\n");
        return 1;
    }
    long len;
    char *src = read_file(argv[argi], &len);
    if (!src)
        return 1;
    int rc = compile_spl(src, argv[argi], argv[argi + 1], gen_debug, dump);
    free(src);
    return rc;
}