Module Name: src
Committed By: rillig
Date: Tue May 16 13:26:26 UTC 2023
Modified Files:
src/usr.bin/indent: indent.c indent.h io.c lexi.c pr_comment.c
Log Message:
indent: directly access the input buffer
No functional change.
To generate a diff of this commit:
cvs rdiff -u -r1.289 -r1.290 src/usr.bin/indent/indent.c
cvs rdiff -u -r1.145 -r1.146 src/usr.bin/indent/indent.h
cvs rdiff -u -r1.174 -r1.175 src/usr.bin/indent/io.c
cvs rdiff -u -r1.196 -r1.197 src/usr.bin/indent/lexi.c
cvs rdiff -u -r1.143 -r1.144 src/usr.bin/indent/pr_comment.c
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
Modified files:
Index: src/usr.bin/indent/indent.c
diff -u src/usr.bin/indent/indent.c:1.289 src/usr.bin/indent/indent.c:1.290
--- src/usr.bin/indent/indent.c:1.289 Tue May 16 11:32:01 2023
+++ src/usr.bin/indent/indent.c Tue May 16 13:26:26 2023
@@ -1,4 +1,4 @@
-/* $NetBSD: indent.c,v 1.289 2023/05/16 11:32:01 rillig Exp $ */
+/* $NetBSD: indent.c,v 1.290 2023/05/16 13:26:26 rillig Exp $ */
/*-
* SPDX-License-Identifier: BSD-4-Clause
@@ -38,7 +38,7 @@
*/
#include <sys/cdefs.h>
-__RCSID("$NetBSD: indent.c,v 1.289 2023/05/16 11:32:01 rillig Exp $");
+__RCSID("$NetBSD: indent.c,v 1.290 2023/05/16 13:26:26 rillig Exp $");
#include <sys/param.h>
#include <err.h>
@@ -291,7 +291,7 @@ main_prepare_parsing(void)
inp_read_line();
int ind = 0;
- for (const char *p = inp_p();; p++) {
+ for (const char *p = inp.st;; p++) {
if (*p == ' ')
ind++;
else if (*p == '\t')
@@ -886,10 +886,10 @@ read_preprocessing_line(void)
buf_add_char(&lab, '#');
- while (ch_isblank(inp_peek()))
- buf_add_char(&lab, inp_next());
+ while (ch_isblank(inp.st[0]))
+ buf_add_char(&lab, *inp.st++);
- while (inp_peek() != '\n' || (state == COMM && !had_eof)) {
+ while (inp.st[0] != '\n' || (state == COMM && !had_eof)) {
buf_add_char(&lab, inp_next());
switch (lab.mem[lab.len - 1]) {
case '\\':
@@ -897,9 +897,9 @@ read_preprocessing_line(void)
buf_add_char(&lab, inp_next());
break;
case '/':
- if (inp_peek() == '*' && state == PLAIN) {
+ if (inp.st[0] == '*' && state == PLAIN) {
state = COMM;
- buf_add_char(&lab, inp_next());
+ buf_add_char(&lab, *inp.st++);
}
break;
case '"':
@@ -915,9 +915,9 @@ read_preprocessing_line(void)
state = CHR;
break;
case '*':
- if (inp_peek() == '/' && state == COMM) {
+ if (inp.st[0] == '/' && state == COMM) {
state = PLAIN;
- buf_add_char(&lab, inp_next());
+ buf_add_char(&lab, *inp.st++);
}
break;
}
Index: src/usr.bin/indent/indent.h
diff -u src/usr.bin/indent/indent.h:1.145 src/usr.bin/indent/indent.h:1.146
--- src/usr.bin/indent/indent.h:1.145 Tue May 16 11:32:01 2023
+++ src/usr.bin/indent/indent.h Tue May 16 13:26:26 2023
@@ -1,4 +1,4 @@
-/* $NetBSD: indent.h,v 1.145 2023/05/16 11:32:01 rillig Exp $ */
+/* $NetBSD: indent.h,v 1.146 2023/05/16 13:26:26 rillig Exp $ */
/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
@@ -133,6 +133,15 @@ struct buffer {
extern FILE *input;
extern FILE *output;
+/*
+ * The current line from the input file, used by the lexer to generate tokens.
+ * To read from the line, start at inp.st and continue up to and including the
+ * next '\n'. To read beyond the '\n', call inp_skip or inp_next, which will
+ * make the next line available, invalidating any pointers into the previous
+ * line.
+ */
+extern struct buffer inp;
+
extern struct buffer token; /* the current token to be processed, is
* typically copied to the buffer 'code', or
* in some cases to 'lab'. */
@@ -409,10 +418,6 @@ int compute_code_indent(void);
int compute_label_indent(void);
int ind_add(int, const char *, size_t);
-const char *inp_p(void);
-const char *inp_line_start(void);
-char inp_peek(void);
-char inp_lookahead(size_t);
void inp_skip(void);
char inp_next(void);
void clear_indent_off_text(void);
Index: src/usr.bin/indent/io.c
diff -u src/usr.bin/indent/io.c:1.174 src/usr.bin/indent/io.c:1.175
--- src/usr.bin/indent/io.c:1.174 Tue May 16 11:32:01 2023
+++ src/usr.bin/indent/io.c Tue May 16 13:26:26 2023
@@ -1,4 +1,4 @@
-/* $NetBSD: io.c,v 1.174 2023/05/16 11:32:01 rillig Exp $ */
+/* $NetBSD: io.c,v 1.175 2023/05/16 13:26:26 rillig Exp $ */
/*-
* SPDX-License-Identifier: BSD-4-Clause
@@ -38,47 +38,19 @@
*/
#include <sys/cdefs.h>
-__RCSID("$NetBSD: io.c,v 1.174 2023/05/16 11:32:01 rillig Exp $");
+__RCSID("$NetBSD: io.c,v 1.175 2023/05/16 13:26:26 rillig Exp $");
#include <stdio.h>
#include <string.h>
#include "indent.h"
-/*
- * The current line, ready to be split into tokens, terminated with '\n'. The
- * current read position is inp.s, and the invariant inp.s < inp.e holds.
- */
-static struct buffer inp;
+struct buffer inp;
static struct buffer indent_off_text;
static int paren_indent;
-const char *
-inp_p(void)
-{
- return inp.st;
-}
-
-const char *
-inp_line_start(void)
-{
- return inp.mem;
-}
-
-char
-inp_peek(void)
-{
- return *inp.st;
-}
-
-char
-inp_lookahead(size_t i)
-{
- return inp.st[i];
-}
-
void
inp_skip(void)
{
@@ -90,7 +62,7 @@ inp_skip(void)
char
inp_next(void)
{
- char ch = inp_peek();
+ char ch = inp.st[0];
inp_skip();
return ch;
}
Index: src/usr.bin/indent/lexi.c
diff -u src/usr.bin/indent/lexi.c:1.196 src/usr.bin/indent/lexi.c:1.197
--- src/usr.bin/indent/lexi.c:1.196 Tue May 16 12:46:43 2023
+++ src/usr.bin/indent/lexi.c Tue May 16 13:26:26 2023
@@ -1,4 +1,4 @@
-/* $NetBSD: lexi.c,v 1.196 2023/05/16 12:46:43 rillig Exp $ */
+/* $NetBSD: lexi.c,v 1.197 2023/05/16 13:26:26 rillig Exp $ */
/*-
* SPDX-License-Identifier: BSD-4-Clause
@@ -38,7 +38,7 @@
*/
#include <sys/cdefs.h>
-__RCSID("$NetBSD: lexi.c,v 1.196 2023/05/16 12:46:43 rillig Exp $");
+__RCSID("$NetBSD: lexi.c,v 1.197 2023/05/16 13:26:26 rillig Exp $");
#include <stdlib.h>
#include <string.h>
@@ -178,9 +178,9 @@ static void
lex_number(void)
{
for (unsigned char s = 'A'; s != 'f' && s != 'i' && s != 'u';) {
- unsigned char ch = (unsigned char)inp_peek();
- if (ch == '\\' && inp_lookahead(1) == '\n') {
- inp_skip();
+ unsigned char ch = (unsigned char)inp.st[0];
+ if (ch == '\\' && inp.st[1] == '\n') {
+ inp.st++;
inp_skip();
line_no++;
continue;
@@ -218,10 +218,10 @@ static void
lex_word(void)
{
for (;;) {
- if (is_identifier_part(inp_peek()))
- token_add_char(inp_next());
- else if (inp_peek() == '\\' && inp_lookahead(1) == '\n') {
- inp_skip();
+ if (is_identifier_part(inp.st[0]))
+ token_add_char(*inp.st++);
+ else if (inp.st[0] == '\\' && inp.st[1] == '\n') {
+ inp.st++;
inp_skip();
line_no++;
} else
@@ -233,17 +233,17 @@ static void
lex_char_or_string(void)
{
for (char delim = token.mem[token.len - 1];;) {
- if (inp_peek() == '\n') {
+ if (inp.st[0] == '\n') {
diag(1, "Unterminated literal");
return;
}
- token_add_char(inp_next());
+ token_add_char(*inp.st++);
if (token.mem[token.len - 1] == delim)
return;
if (token.mem[token.len - 1] == '\\') {
- if (inp_peek() == '\n')
+ if (inp.st[0] == '\n')
++line_no;
token_add_char(inp_next());
}
@@ -260,10 +260,10 @@ probably_typename(void)
return false;
if (ps.in_stmt_or_decl) /* XXX: this condition looks incorrect */
return false;
- if (inp_peek() == '*' && inp_lookahead(1) != '=')
+ if (inp.st[0] == '*' && inp.st[1] != '=')
goto maybe;
/* XXX: is_identifier_start */
- if (ch_isalpha(inp_peek()))
+ if (ch_isalpha(inp.st[0]))
goto maybe;
return false;
maybe:
@@ -316,7 +316,7 @@ static bool
probably_looking_at_definition(void)
{
int paren_level = 0;
- for (const char *p = inp_p(); *p != '\n'; p++) {
+ for (const char *p = inp.st; *p != '\n'; p++) {
if (*p == '(')
paren_level++;
if (*p == ')' && --paren_level == 0) {
@@ -353,15 +353,15 @@ probably_looking_at_definition(void)
static lexer_symbol
lexi_alnum(void)
{
- if (ch_isdigit(inp_peek()) ||
- (inp_peek() == '.' && ch_isdigit(inp_lookahead(1)))) {
+ if (ch_isdigit(inp.st[0]) ||
+ (inp.st[0] == '.' && ch_isdigit(inp.st[1]))) {
lex_number();
- } else if (is_identifier_start(inp_peek())) {
+ } else if (is_identifier_start(inp.st[0])) {
lex_word();
if (token.len == 1 && token.st[0] == 'L' &&
- (inp_peek() == '"' || inp_peek() == '\'')) {
- token_add_char(inp_next());
+ (inp.st[0] == '"' || inp.st[0] == '\'')) {
+ token_add_char(*inp.st++);
lex_char_or_string();
ps.next_unary = false;
return lsym_word;
@@ -369,8 +369,8 @@ lexi_alnum(void)
} else
return lsym_eof; /* just as a placeholder */
- while (ch_isblank(inp_peek()))
- inp_skip();
+ while (ch_isblank(inp.st[0]))
+ inp.st++;
ps.next_unary = ps.prev_token == lsym_tag
|| ps.prev_token == lsym_typedef;
@@ -415,7 +415,7 @@ found_typename:
}
}
- if (inp_peek() == '(' && ps.tos <= 1 && ps.ind_level == 0 &&
+ if (inp.st[0] == '(' && ps.tos <= 1 && ps.ind_level == 0 &&
!ps.in_func_def_params && !ps.block_init) {
if (ps.nparen == 0 && probably_looking_at_definition()) {
@@ -447,14 +447,14 @@ is_asterisk_unary(void)
static void
lex_asterisk_unary(void)
{
- while (inp_peek() == '*' || ch_isspace(inp_peek())) {
- if (inp_peek() == '*')
+ while (inp.st[0] == '*' || ch_isspace(inp.st[0])) {
+ if (inp.st[0] == '*')
token_add_char('*');
inp_skip();
}
if (ps.in_decl) {
- for (const char *tp = inp_p(); *tp != '\n';) {
+ for (const char *tp = inp.st; *tp != '\n';) {
if (ch_isspace(*tp))
tp++;
else if (is_identifier_start(*tp)) {
@@ -491,7 +491,7 @@ skip_string(const char **pp, const char
static void
lex_indent_comment(void)
{
- const char *p = inp_line_start();
+ const char *p = inp.mem;
skip_blank(&p);
if (!skip_string(&p, "/*"))
@@ -528,11 +528,11 @@ lexi(void)
ps.next_col_1 = false;
for (;;) {
- if (ch_isblank(inp_peek())) {
+ if (ch_isblank(inp.st[0])) {
ps.curr_col_1 = false;
- inp_skip();
- } else if (inp_peek() == '\\' && inp_lookahead(1) == '\n') {
- inp_skip();
+ inp.st++;
+ } else if (inp.st[0] == '\\' && inp.st[1] == '\n') {
+ inp.st++;
inp_skip();
line_no++;
} else
@@ -592,19 +592,19 @@ lexi(void)
lsym = ps.next_unary ? lsym_unary_op : lsym_binary_op;
next_unary = true;
- if (inp_peek() == token.mem[token.len - 1]) { /* '++' or '--' */
- token_add_char(inp_next());
+ if (inp.st[0] == token.mem[token.len - 1]) { /* '++' or '--' */
+ token_add_char(*inp.st++);
if (ps.prev_token == lsym_word ||
ps.prev_token == lsym_rparen_or_rbracket) {
lsym = ps.next_unary ? lsym_unary_op : lsym_postfix_op;
next_unary = false;
}
- } else if (inp_peek() == '=') { /* '+=' or '-=' */
- token_add_char(inp_next());
+ } else if (inp.st[0] == '=') { /* '+=' or '-=' */
+ token_add_char(*inp.st++);
- } else if (inp_peek() == '>') { /* '->' */
- token_add_char(inp_next());
+ } else if (inp.st[0] == '>') { /* '->' */
+ token_add_char(*inp.st++);
lsym = lsym_unary_op;
next_unary = false;
ps.want_blank = false;
@@ -614,8 +614,8 @@ lexi(void)
case '=':
if (ps.init_or_struct)
ps.block_init = true;
- if (inp_peek() == '=')
- token_add_char(inp_next());
+ if (inp.st[0] == '=')
+ token_add_char(*inp.st++);
lsym = lsym_binary_op;
next_unary = true;
break;
@@ -623,10 +623,10 @@ lexi(void)
case '>':
case '<':
case '!': /* ops like <, <<, <=, !=, etc */
- if (inp_peek() == '>' || inp_peek() == '<' || inp_peek() == '=')
- token_add_char(inp_next());
- if (inp_peek() == '=')
- token_add_char(inp_next());
+ if (inp.st[0] == '>' || inp.st[0] == '<' || inp.st[0] == '=')
+ token_add_char(*inp.st++);
+ if (inp.st[0] == '=')
+ token_add_char(*inp.st++);
lsym = ps.next_unary ? lsym_unary_op : lsym_binary_op;
next_unary = true;
break;
@@ -637,8 +637,8 @@ lexi(void)
lsym = lsym_unary_op;
next_unary = true;
} else {
- if (inp_peek() == '=')
- token_add_char(inp_next());
+ if (inp.st[0] == '=')
+ token_add_char(*inp.st++);
lsym = lsym_binary_op;
next_unary = true;
}
@@ -646,20 +646,20 @@ lexi(void)
default:
if (token.mem[token.len - 1] == '/'
- && (inp_peek() == '*' || inp_peek() == '/')) {
+ && (inp.st[0] == '*' || inp.st[0] == '/')) {
enum indent_enabled prev = indent_enabled;
lex_indent_comment();
if (prev == indent_on && indent_enabled == indent_off)
clear_indent_off_text();
- token_add_char(inp_next());
+ token_add_char(*inp.st++);
lsym = lsym_comment;
next_unary = ps.next_unary;
break;
}
/* handle '||', '&&', etc., and also things as in 'int *****i' */
- while (token.mem[token.len - 1] == inp_peek() || inp_peek() == '=')
- token_add_char(inp_next());
+ while (inp.st[0] == token.mem[token.len - 1] || inp.st[0] == '=')
+ token_add_char(*inp.st++);
lsym = ps.next_unary ? lsym_unary_op : lsym_binary_op;
next_unary = true;
Index: src/usr.bin/indent/pr_comment.c
diff -u src/usr.bin/indent/pr_comment.c:1.143 src/usr.bin/indent/pr_comment.c:1.144
--- src/usr.bin/indent/pr_comment.c:1.143 Tue May 16 11:32:01 2023
+++ src/usr.bin/indent/pr_comment.c Tue May 16 13:26:26 2023
@@ -1,4 +1,4 @@
-/* $NetBSD: pr_comment.c,v 1.143 2023/05/16 11:32:01 rillig Exp $ */
+/* $NetBSD: pr_comment.c,v 1.144 2023/05/16 13:26:26 rillig Exp $ */
/*-
* SPDX-License-Identifier: BSD-4-Clause
@@ -38,7 +38,7 @@
*/
#include <sys/cdefs.h>
-__RCSID("$NetBSD: pr_comment.c,v 1.143 2023/05/16 11:32:01 rillig Exp $");
+__RCSID("$NetBSD: pr_comment.c,v 1.144 2023/05/16 13:26:26 rillig Exp $");
#include <string.h>
@@ -61,7 +61,7 @@ com_add_delim(void)
static bool
fits_in_one_line(int com_ind, int max_line_length)
{
- for (const char *start = inp_p(), *p = start; *p != '\n'; p++) {
+ for (const char *start = inp.st, *p = start; *p != '\n'; p++) {
if (p[0] == '*' && p[1] == '/') {
int len = ind_add(com_ind + 3, start, (size_t)(p - start));
len += p == start || ch_isblank(p[-1]) ? 2 : 3;
@@ -85,9 +85,9 @@ analyze_comment(bool *p_may_wrap, bool *
ind = 0;
} else {
- if (inp_peek() == '-' || inp_peek() == '*' ||
+ if (inp.st[0] == '-' || inp.st[0] == '*' ||
token.mem[token.len - 1] == '/' ||
- (inp_peek() == '\n' && !opt.format_block_comments)) {
+ (inp.st[0] == '\n' && !opt.format_block_comments)) {
may_wrap = false;
delim = false;
}
@@ -122,20 +122,20 @@ analyze_comment(bool *p_may_wrap, bool *
* Find out how much indentation there was originally, because that
* much will have to be ignored by output_complete_line.
*/
- size_t len = (size_t)(inp_p() - 2 - inp_line_start());
- ps.n_comment_delta = -ind_add(0, inp_line_start(), len);
+ size_t len = (size_t)(inp.st - 2 - inp.mem);
+ ps.n_comment_delta = -ind_add(0, inp.mem, len);
} else {
ps.n_comment_delta = 0;
- if (!(inp_peek() == '\t' && !ch_isblank(inp_lookahead(1))))
- while (ch_isblank(inp_peek()))
- inp_skip();
+ if (!(inp.st[0] == '\t' && !ch_isblank(inp.st[1])))
+ while (ch_isblank(inp.st[0]))
+ inp.st++;
}
ps.comment_delta = 0;
com_add_char('/');
com_add_char(token.mem[token.len - 1]); /* either '*' or '/' */
- if (may_wrap && !ch_isblank(inp_peek()))
+ if (may_wrap && !ch_isblank(inp.st[0]))
com_add_char(' ');
if (delim && fits_in_one_line(ind, line_length))
@@ -163,7 +163,7 @@ copy_comment_wrap(int line_length, bool
ssize_t last_blank = -1; /* index of the last blank in com.mem */
for (;;) {
- switch (inp_peek()) {
+ switch (inp.st[0]) {
case '\n':
if (had_eof) {
diag(1, "Unterminated comment");
@@ -194,21 +194,21 @@ copy_comment_wrap(int line_length, bool
do { /* flush any blanks and/or tabs at start of
* next line */
inp_skip();
- if (inp_peek() == '*' && skip_asterisk) {
+ if (inp.st[0] == '*' && skip_asterisk) {
skip_asterisk = false;
- inp_skip();
- if (inp_peek() == '/')
+ inp.st++;
+ if (inp.st[0] == '/')
goto end_of_comment;
}
- } while (ch_isblank(inp_peek()));
+ } while (ch_isblank(inp.st[0]));
break; /* end of case for newline */
case '*':
- inp_skip();
- if (inp_peek() == '/') {
+ inp.st++;
+ if (inp.st[0] == '/') {
end_of_comment:
- inp_skip();
+ inp.st++;
if (delim) {
if (com.len > 3)
@@ -244,7 +244,7 @@ copy_comment_wrap(int line_length, bool
last_blank = (ssize_t)com.len;
com_add_char(ch);
now_len++;
- if (memchr("*\n\r\b\t", inp_peek(), 6) != NULL)
+ if (memchr("*\n\r\b\t", inp.st[0], 6) != NULL)
break;
if (now_len >= line_length && last_blank != -1)
break;
@@ -284,7 +284,7 @@ static void
copy_comment_nowrap(void)
{
for (;;) {
- if (inp_peek() == '\n') {
+ if (inp.st[0] == '\n') {
if (token.mem[token.len - 1] == '/')
return;
@@ -302,7 +302,7 @@ copy_comment_nowrap(void)
continue;
}
- com_add_char(inp_next());
+ com_add_char(*inp.st++);
if (com.mem[com.len - 2] == '*' && com.mem[com.len - 1] == '/'
&& token.mem[token.len - 1] == '*')
return;
