Module Name: src
Committed By: rillig
Date: Sun Dec 5 11:40:03 UTC 2021
Modified Files:
src/usr.bin/make: for.c str.h
src/usr.bin/make/unit-tests: directive-for-escape.mk
Log Message:
make: inline Str_Words into .for loop handling
This saves one memory allocation and a bit of copying, per .for loop.
No functional change.
To generate a diff of this commit:
cvs rdiff -u -r1.147 -r1.148 src/usr.bin/make/for.c
cvs rdiff -u -r1.9 -r1.10 src/usr.bin/make/str.h
cvs rdiff -u -r1.11 -r1.12 \
src/usr.bin/make/unit-tests/directive-for-escape.mk
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/make/for.c
diff -u src/usr.bin/make/for.c:1.147 src/usr.bin/make/for.c:1.148
--- src/usr.bin/make/for.c:1.147 Thu Sep 2 07:02:07 2021
+++ src/usr.bin/make/for.c Sun Dec 5 11:40:03 2021
@@ -1,4 +1,4 @@
-/* $NetBSD: for.c,v 1.147 2021/09/02 07:02:07 rillig Exp $ */
+/* $NetBSD: for.c,v 1.148 2021/12/05 11:40:03 rillig Exp $ */
/*
* Copyright (c) 1992, The Regents of the University of California.
@@ -58,7 +58,7 @@
#include "make.h"
/* "@(#)for.c 8.1 (Berkeley) 6/6/93" */
-MAKE_RCSID("$NetBSD: for.c,v 1.147 2021/09/02 07:02:07 rillig Exp $");
+MAKE_RCSID("$NetBSD: for.c,v 1.148 2021/12/05 11:40:03 rillig Exp $");
/* One of the variables to the left of the "in" in a .for loop. */
@@ -70,7 +70,7 @@ typedef struct ForVar {
typedef struct ForLoop {
Buffer body; /* Unexpanded body of the loop */
Vector /* of ForVar */ vars; /* Iteration variables */
- Words items; /* Substitution items */
+ SubstringWords items; /* Substitution items */
Buffer curBody; /* Expanded body of the current iteration */
unsigned int sub_next; /* Where to continue iterating */
} ForLoop;
@@ -87,8 +87,7 @@ ForLoop_New(void)
Buf_Init(&f->body);
Vector_Init(&f->vars, sizeof(ForVar));
- f->items.words = NULL;
- f->items.freeIt = NULL;
+ SubstringWords_Init(&f->items);
Buf_Init(&f->curBody);
f->sub_next = 0;
@@ -106,7 +105,7 @@ ForLoop_Free(ForLoop *f)
}
Vector_Done(&f->vars);
- Words_Free(f->items);
+ SubstringWords_Free(f->items);
Buf_Done(&f->curBody);
free(f);
@@ -171,10 +170,10 @@ ForLoop_ParseItems(ForLoop *f, const cha
return false;
}
- f->items = Str_Words(items, false);
+ f->items = Substring_Words(items, false);
free(items);
- if (f->items.len == 1 && f->items.words[0][0] == '\0')
+ if (f->items.len == 1 && Substring_IsEmpty(f->items.words[0]))
f->items.len = 0; /* .for var in ${:U} */
if (f->items.len != 0 && f->items.len % f->vars.len != 0) {
@@ -278,17 +277,16 @@ For_Accum(const char *line)
static size_t
-ExprLen(const char *expr)
+ExprLen(const char *s, const char *e)
{
- char ch, expr_open, expr_close;
+ char expr_open, expr_close;
int depth;
- size_t len;
+ const char *p;
- expr_open = expr[0];
- if (expr_open == '\0')
- /* just escape the $ */
- return 0;
+ if (s == e)
+ return 0; /* just escape the '$' */
+ expr_open = s[0];
if (expr_open == '(')
expr_close = ')';
else if (expr_open == '{')
@@ -297,11 +295,11 @@ ExprLen(const char *expr)
return 1; /* Single char variable */
depth = 1;
- for (len = 1; (ch = expr[len++]) != '\0';) {
- if (ch == expr_open)
+ for (p = s + 1; p != e; p++) {
+ if (*p == expr_open)
depth++;
- else if (ch == expr_close && --depth == 0)
- return len;
+ else if (*p == expr_close && --depth == 0)
+ return (size_t)(p + 1 - s);
}
/* Expression end not found, escape the $ */
@@ -313,11 +311,11 @@ ExprLen(const char *expr)
* that characters that break this syntax must be backslash-escaped.
*/
static bool
-NeedsEscapes(const char *value, char endc)
+NeedsEscapes(Substring value, char endc)
{
const char *p;
- for (p = value; *p != '\0'; p++) {
+ for (p = value.start; p != value.end; p++) {
if (*p == ':' || *p == '$' || *p == '\\' || *p == endc ||
*p == '\n')
return true;
@@ -332,27 +330,29 @@ NeedsEscapes(const char *value, char end
* The result is later unescaped by ApplyModifier_Defined.
*/
static void
-Buf_AddEscaped(Buffer *cmds, const char *item, char endc)
+Buf_AddEscaped(Buffer *cmds, Substring item, char endc)
{
+ const char *p;
char ch;
if (!NeedsEscapes(item, endc)) {
- Buf_AddStr(cmds, item);
+ Buf_AddBytesBetween(cmds, item.start, item.end);
return;
}
/* Escape ':', '$', '\\' and 'endc' - these will be removed later by
* :U processing, see ApplyModifier_Defined. */
- while ((ch = *item++) != '\0') {
+ for (p = item.start; p != item.end; p++) {
+ ch = *p;
if (ch == '$') {
- size_t len = ExprLen(item);
+ size_t len = ExprLen(p + 1, item.end);
if (len != 0) {
/*
* XXX: Should a '\' be added here?
* See directive-for-escape.mk, ExprLen.
*/
- Buf_AddBytes(cmds, item - 1, len + 1);
- item += len;
+ Buf_AddBytes(cmds, p, 1 + len);
+ p += len;
continue;
}
Buf_AddByte(cmds, '\\');
Index: src/usr.bin/make/str.h
diff -u src/usr.bin/make/str.h:1.9 src/usr.bin/make/str.h:1.10
--- src/usr.bin/make/str.h:1.9 Sun May 30 21:16:54 2021
+++ src/usr.bin/make/str.h Sun Dec 5 11:40:03 2021
@@ -1,4 +1,4 @@
-/* $NetBSD: str.h,v 1.9 2021/05/30 21:16:54 rillig Exp $ */
+/* $NetBSD: str.h,v 1.10 2021/12/05 11:40:03 rillig Exp $ */
/*
Copyright (c) 2021 Roland Illig <ril...@netbsd.org>
@@ -353,6 +353,14 @@ Words_Free(Words w)
SubstringWords Substring_Words(const char *, bool);
MAKE_INLINE void
+SubstringWords_Init(SubstringWords *w)
+{
+ w->words = NULL;
+ w->len = 0;
+ w->freeIt = NULL;
+}
+
+MAKE_INLINE void
SubstringWords_Free(SubstringWords w)
{
free(w.words);
Index: src/usr.bin/make/unit-tests/directive-for-escape.mk
diff -u src/usr.bin/make/unit-tests/directive-for-escape.mk:1.11 src/usr.bin/make/unit-tests/directive-for-escape.mk:1.12
--- src/usr.bin/make/unit-tests/directive-for-escape.mk:1.11 Thu Sep 2 07:02:08 2021
+++ src/usr.bin/make/unit-tests/directive-for-escape.mk Sun Dec 5 11:40:03 2021
@@ -1,4 +1,4 @@
-# $NetBSD: directive-for-escape.mk,v 1.11 2021/09/02 07:02:08 rillig Exp $
+# $NetBSD: directive-for-escape.mk,v 1.12 2021/12/05 11:40:03 rillig Exp $
#
# Test escaping of special characters in the iteration values of a .for loop.
# These values get expanded later using the :U variable modifier, and this
@@ -13,8 +13,8 @@
ASCII= !"\#$$%&'()*+,-./0-9:;<=>?@A-Z[\]_^a-z{|}~
# XXX: As of 2020-12-31, the '#' is not preserved in the expanded body of
-# the loop since it would not need only the escaping for the :U variable
-# modifier but also the escaping for the line-end comment.
+# the loop. Not only would it need the escaping for the variable modifier
+# ':U' but also the escaping for the line-end comment.
.for chars in ${ASCII}
. info ${chars}
.endfor
@@ -34,7 +34,7 @@ ASCII.2020-12-31= !"\\\#$$%&'()*+,-./0-9
# XXX: It is unexpected that the variable V gets expanded in the loop body.
# The double '$$' should intuitively prevent exactly this. Probably nobody
# was adventurous enough to use literal dollar signs in the values of a .for
-# loop.
+# loop, allowing this edge case to go unnoticed for years.
#
# See for.c, function ExprLen.
V= value
