refactor: move tokenization code from preprocess.c to tokenize.c

1 files changed, 2 insertions, 393 deletions

diff --git a/src/preprocess.c b/src/preprocess.c
index 34c2fe0..0af146c 100644
--- a/src/preprocess.c
+++ b/src/preprocess.c
@@ -1,12 +1,12 @@
 #include "preprocess.h"
 #include <assert.h>
-#include <ctype.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include "common.h"
 #include "parse.h"
 #include "sys.h"
+#include "tokenize.h"
 
 typedef enum {
     MacroKind_undef,
@@ -165,397 +165,6 @@ void macroargs_build_json(JsonBuilder* builder, MacroArgArray* macroargs) {
 }
 
 typedef struct {
-    InFile* src;
-    bool at_bol;
-    bool expect_header_name;
-    TokenArray* pp_tokens;
-} PpLexer;
-
-static PpLexer* pplexer_new(InFile* src) {
-    PpLexer* ppl = calloc(1, sizeof(PpLexer));
-
-    ppl->src = src;
-    ppl->at_bol = true;
-    ppl->expect_header_name = false;
-    ppl->pp_tokens = calloc(1, sizeof(TokenArray));
-    tokens_init(ppl->pp_tokens, 1024 * 16);
-
-    return ppl;
-}
-
-static void pplexer_tokenize_pp_directive(PpLexer* ppl, Token* tok) {
-    // Skip whitespaces after '#'.
-    char c;
-    while (isspace((c = infile_peek_char(ppl->src)))) {
-        if (c == '\n')
-            break;
-        infile_next_char(ppl->src);
-    }
-    // '#' new-line
-    if (c == '\n') {
-        tok->kind = TokenKind_pp_directive_nop;
-        return;
-    }
-
-    StrBuilder builder;
-    strbuilder_init(&builder);
-    while (isalnum(infile_peek_char(ppl->src))) {
-        strbuilder_append_char(&builder, infile_peek_char(ppl->src));
-        infile_next_char(ppl->src);
-    }
-    const char* pp_directive_name = builder.buf;
-
-    if (builder.len == 0) {
-        tok->kind = TokenKind_hash;
-    } else if (strcmp(pp_directive_name, "define") == 0) {
-        tok->kind = TokenKind_pp_directive_define;
-    } else if (strcmp(pp_directive_name, "elif") == 0) {
-        tok->kind = TokenKind_pp_directive_elif;
-    } else if (strcmp(pp_directive_name, "elifdef") == 0) {
-        tok->kind = TokenKind_pp_directive_elifdef;
-    } else if (strcmp(pp_directive_name, "elifndef") == 0) {
-        tok->kind = TokenKind_pp_directive_elifndef;
-    } else if (strcmp(pp_directive_name, "else") == 0) {
-        tok->kind = TokenKind_pp_directive_else;
-    } else if (strcmp(pp_directive_name, "embed") == 0) {
-        tok->kind = TokenKind_pp_directive_embed;
-    } else if (strcmp(pp_directive_name, "endif") == 0) {
-        tok->kind = TokenKind_pp_directive_endif;
-    } else if (strcmp(pp_directive_name, "error") == 0) {
-        tok->kind = TokenKind_pp_directive_error;
-    } else if (strcmp(pp_directive_name, "if") == 0) {
-        tok->kind = TokenKind_pp_directive_if;
-    } else if (strcmp(pp_directive_name, "ifdef") == 0) {
-        tok->kind = TokenKind_pp_directive_ifdef;
-    } else if (strcmp(pp_directive_name, "ifndef") == 0) {
-        tok->kind = TokenKind_pp_directive_ifndef;
-    } else if (strcmp(pp_directive_name, "include") == 0) {
-        ppl->expect_header_name = true;
-        tok->kind = TokenKind_pp_directive_include;
-    } else if (strcmp(pp_directive_name, "line") == 0) {
-        tok->kind = TokenKind_pp_directive_line;
-    } else if (strcmp(pp_directive_name, "pragma") == 0) {
-        tok->kind = TokenKind_pp_directive_pragma;
-    } else if (strcmp(pp_directive_name, "undef") == 0) {
-        tok->kind = TokenKind_pp_directive_undef;
-    } else if (strcmp(pp_directive_name, "warning") == 0) {
-        tok->kind = TokenKind_pp_directive_warning;
-    } else {
-        tok->kind = TokenKind_pp_directive_non_directive;
-        tok->value.string = pp_directive_name;
-    }
-}
-
-static void pplexer_tokenize_all(PpLexer* ppl) {
-    while (!infile_eof(ppl->src)) {
-        Token* tok = tokens_push_new(ppl->pp_tokens);
-        tok->loc = ppl->src->loc;
-        char c = infile_peek_char(ppl->src);
-
-        if (ppl->expect_header_name && c == '"') {
-            infile_next_char(ppl->src);
-            StrBuilder builder;
-            strbuilder_init(&builder);
-            strbuilder_append_char(&builder, '"');
-            while (1) {
-                char ch = infile_peek_char(ppl->src);
-                if (ch == '"')
-                    break;
-                strbuilder_append_char(&builder, ch);
-                if (ch == '\\') {
-                    infile_next_char(ppl->src);
-                    strbuilder_append_char(&builder, infile_peek_char(ppl->src));
-                }
-                infile_next_char(ppl->src);
-            }
-            strbuilder_append_char(&builder, '"');
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_header_name;
-            tok->value.string = builder.buf;
-            ppl->expect_header_name = false;
-        } else if (ppl->expect_header_name && c == '<') {
-            infile_next_char(ppl->src);
-            StrBuilder builder;
-            strbuilder_init(&builder);
-            strbuilder_append_char(&builder, '<');
-            while (1) {
-                char ch = infile_peek_char(ppl->src);
-                if (ch == '>')
-                    break;
-                strbuilder_append_char(&builder, ch);
-                infile_next_char(ppl->src);
-            }
-            strbuilder_append_char(&builder, '>');
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_header_name;
-            tok->value.string = builder.buf;
-            ppl->expect_header_name = false;
-        } else if (c == '(') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_paren_l;
-        } else if (c == ')') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_paren_r;
-        } else if (c == '{') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_brace_l;
-        } else if (c == '}') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_brace_r;
-        } else if (c == '[') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_bracket_l;
-        } else if (c == ']') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_bracket_r;
-        } else if (c == ',') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_comma;
-        } else if (c == ':') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_colon;
-        } else if (c == ';') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_semicolon;
-        } else if (c == '^') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_assign_xor;
-            } else {
-                tok->kind = TokenKind_xor;
-            }
-        } else if (c == '?') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_question;
-        } else if (c == '~') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_tilde;
-        } else if (c == '+') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_assign_add;
-            } else if (infile_consume_if(ppl->src, '+')) {
-                tok->kind = TokenKind_plusplus;
-            } else {
-                tok->kind = TokenKind_plus;
-            }
-        } else if (c == '|') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_assign_or;
-            } else if (infile_consume_if(ppl->src, '|')) {
-                tok->kind = TokenKind_oror;
-            } else {
-                tok->kind = TokenKind_or;
-            }
-        } else if (c == '&') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_assign_and;
-            } else if (infile_consume_if(ppl->src, '&')) {
-                tok->kind = TokenKind_andand;
-            } else {
-                tok->kind = TokenKind_and;
-            }
-        } else if (c == '-') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '>')) {
-                tok->kind = TokenKind_arrow;
-            } else if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_assign_sub;
-            } else if (infile_consume_if(ppl->src, '-')) {
-                tok->kind = TokenKind_minusminus;
-            } else {
-                tok->kind = TokenKind_minus;
-            }
-        } else if (c == '*') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_assign_mul;
-            } else {
-                tok->kind = TokenKind_star;
-            }
-        } else if (c == '/') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_assign_div;
-            } else if (infile_consume_if(ppl->src, '/')) {
-                while (!infile_eof(ppl->src) && infile_peek_char(ppl->src) != '\n') {
-                    infile_next_char(ppl->src);
-                }
-                tok->kind = TokenKind_whitespace;
-            } else if (infile_consume_if(ppl->src, '*')) {
-                while (infile_peek_char(ppl->src)) {
-                    if (infile_consume_if(ppl->src, '*')) {
-                        if (infile_consume_if(ppl->src, '/')) {
-                            break;
-                        }
-                        continue;
-                    }
-                    infile_next_char(ppl->src);
-                }
-                tok->kind = TokenKind_whitespace;
-            } else {
-                tok->kind = TokenKind_slash;
-            }
-        } else if (c == '%') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_assign_mod;
-            } else {
-                tok->kind = TokenKind_percent;
-            }
-        } else if (c == '.') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '.')) {
-                if (infile_consume_if(ppl->src, '.')) {
-                    tok->kind = TokenKind_ellipsis;
-                } else {
-                    tok->kind = TokenKind_other;
-                    tok->value.string = "..";
-                }
-            } else {
-                tok->kind = TokenKind_dot;
-            }
-        } else if (c == '!') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_ne;
-            } else {
-                tok->kind = TokenKind_not;
-            }
-        } else if (c == '=') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_eq;
-            } else {
-                tok->kind = TokenKind_assign;
-            }
-        } else if (c == '<') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_le;
-            } else if (infile_consume_if(ppl->src, '<')) {
-                if (infile_consume_if(ppl->src, '=')) {
-                    tok->kind = TokenKind_assign_lshift;
-                } else {
-                    tok->kind = TokenKind_lshift;
-                }
-            } else {
-                tok->kind = TokenKind_lt;
-            }
-        } else if (c == '>') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '=')) {
-                tok->kind = TokenKind_ge;
-            } else if (infile_consume_if(ppl->src, '>')) {
-                if (infile_consume_if(ppl->src, '=')) {
-                    tok->kind = TokenKind_assign_rshift;
-                } else {
-                    tok->kind = TokenKind_rshift;
-                }
-            } else {
-                tok->kind = TokenKind_gt;
-            }
-        } else if (c == '#') {
-            infile_next_char(ppl->src);
-            if (infile_consume_if(ppl->src, '#')) {
-                tok->kind = TokenKind_hashhash;
-            } else {
-                if (ppl->at_bol) {
-                    pplexer_tokenize_pp_directive(ppl, tok);
-                } else {
-                    tok->kind = TokenKind_hash;
-                }
-            }
-        } else if (c == '\'') {
-            infile_next_char(ppl->src);
-            StrBuilder builder;
-            strbuilder_init(&builder);
-            strbuilder_append_char(&builder, '\'');
-            strbuilder_append_char(&builder, infile_peek_char(ppl->src));
-            if (infile_peek_char(ppl->src) == '\\') {
-                infile_next_char(ppl->src);
-                strbuilder_append_char(&builder, infile_peek_char(ppl->src));
-            }
-            strbuilder_append_char(&builder, '\'');
-            infile_next_char(ppl->src);
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_character_constant;
-            tok->value.string = builder.buf;
-        } else if (c == '"') {
-            infile_next_char(ppl->src);
-            StrBuilder builder;
-            strbuilder_init(&builder);
-            while (1) {
-                char ch = infile_peek_char(ppl->src);
-                if (ch == '"')
-                    break;
-                strbuilder_append_char(&builder, ch);
-                if (ch == '\\') {
-                    infile_next_char(ppl->src);
-                    strbuilder_append_char(&builder, infile_peek_char(ppl->src));
-                }
-                infile_next_char(ppl->src);
-            }
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_literal_str;
-            tok->value.string = builder.buf;
-        } else if (isdigit(c)) {
-            // TODO: implement tokenization of pp-number.
-            StrBuilder builder;
-            strbuilder_init(&builder);
-            while (isalnum(infile_peek_char(ppl->src))) {
-                strbuilder_append_char(&builder, infile_peek_char(ppl->src));
-                infile_next_char(ppl->src);
-            }
-            tok->kind = TokenKind_literal_int;
-            tok->value.integer = atoi(builder.buf);
-        } else if (isalpha(c) || c == '_') {
-            StrBuilder builder;
-            strbuilder_init(&builder);
-            while (isalnum(infile_peek_char(ppl->src)) || infile_peek_char(ppl->src) == '_') {
-                strbuilder_append_char(&builder, infile_peek_char(ppl->src));
-                infile_next_char(ppl->src);
-            }
-            tok->kind = TokenKind_ident;
-            tok->value.string = builder.buf;
-        } else if (c == '\n') {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_newline;
-        } else if (isspace(c)) {
-            while (isspace((c = infile_peek_char(ppl->src)))) {
-                if (c == '\n')
-                    break;
-                infile_next_char(ppl->src);
-            }
-            if (ppl->at_bol && infile_peek_char(ppl->src) == '#') {
-                infile_next_char(ppl->src);
-                pplexer_tokenize_pp_directive(ppl, tok);
-            } else {
-                tok->kind = TokenKind_whitespace;
-            }
-        } else {
-            infile_next_char(ppl->src);
-            tok->kind = TokenKind_other;
-            char* buf = calloc(2, sizeof(char));
-            buf[0] = c;
-            tok->value.string = buf;
-        }
-        ppl->at_bol = tok->kind == TokenKind_newline;
-    }
-    Token* eof_tok = tokens_push_new(ppl->pp_tokens);
-    eof_tok->loc = ppl->src->loc;
-    eof_tok->kind = TokenKind_eof;
-}
-
-static TokenArray* pp_tokenize(InFile* src) {
-    PpLexer* ppl = pplexer_new(src);
-    pplexer_tokenize_all(ppl);
-    return ppl->pp_tokens;
-}
-
-typedef struct {
     TokenArray* pp_tokens;
     int pos;
     MacroArray* macros;
@@ -1546,7 +1155,7 @@ static char* get_ducc_include_path() {
 
 static TokenArray* do_preprocess(InFile* src, int depth, MacroArray* macros, StrArray* included_files,
                                  StrArray* user_include_dirs) {
-    TokenArray* pp_tokens = pp_tokenize(src);
+    TokenArray* pp_tokens = tokenize(src);
     Preprocessor* pp = preprocessor_new(pp_tokens, depth, macros, included_files);
 
     // Ducc's built-in headers has highest priority.