ash: change CRLF handling

As noted in commit 2d848eba5 (ash: fix CRLF handling) all CRs are removed when scripts are read. Allow an alternative approach (configurable at build-time, enabled by default): - Do not strip CRs from input. - Treat CR as similar to space or tab in base syntax mode. - Adjust pgetc_eatbnl() to handle backslash-CRLF in the same way as backslash-LF. With these changes scripts containing CRLF line endings are more likely to work. Adds 48-56 bytes. (GitHub issue #285)
author: Ron Yorston <rmy@pobox.com> 2023-02-07 10:03:04 +0000
committer: Ron Yorston <rmy@pobox.com> 2023-02-07 10:03:04 +0000
commit: 4b894b60ae377ee2a04ae209c14bb1e714ecf6f4 (patch)
tree: 8f42f36eb0a9dad2c712f7b0e07aff7dee059217 /shell
parent: c49b8b7052bb1df3ed5f9aeba61a7965d9c77f83 (diff)
download: busybox-w32-4b894b60ae377ee2a04ae209c14bb1e714ecf6f4.tar.gz
busybox-w32-4b894b60ae377ee2a04ae209c14bb1e714ecf6f4.tar.bz2
busybox-w32-4b894b60ae377ee2a04ae209c14bb1e714ecf6f4.zip
1 files changed, 41 insertions, 3 deletions
diff --git a/shell/ash.c b/shell/ash.c
index a24ed0ac8..63eb4947c 100644
--- a/shell/ash.c
+++ b/shell/ash.c
@@ -192,6 +192,14 @@
 //config:       Enable support for the 'nocaseglob' option, which allows
 //config:       case-insensitive filename globbing.
 //config:
+//config:config ASH_IGNORE_CR
+//config:       bool "Ignore CR in scripts"
+//config:       default y
+//config:       depends on (ASH || SH_IS_ASH || BASH_IS_ASH) && PLATFORM_MINGW32
+//config:       help
+//config:       Allow CRs to be ignored when shell scripts are parsed.  This
+//config:       makes it possible to run scripts with CRLF line endings.
+//config:
 //config:endif # ash options
 //applet:IF_ASH(APPLET(ash, BB_DIR_BIN, BB_SUID_DROP))
@@ -762,7 +770,11 @@ struct strpush {
        struct strpush *spfree;
        /* Remember last two characters for pungetc. */
+#if ENABLE_ASH_IGNORE_CR
+        int lastc[3];   /* Ignoring CRs needs more pungetc. */
+#else
        int lastc[2];
+#endif
        /* Number of outstanding calls to pungetc. */
        int unget;
@@ -787,7 +799,11 @@ struct parsefile {
        struct strpush *spfree;
        /* Remember last two characters for pungetc. */
+#if ENABLE_ASH_IGNORE_CR
+        int lastc[3];   /* Ignoring CRs needs more pungetc. */
+#else
        int lastc[2];
+#endif
        /* Number of outstanding calls to pungetc. */
        int unget;
@@ -3596,7 +3612,11 @@ static const uint8_t syntax_index_table[] ALIGN1 = {
        /*  10 "\n" */ CNL_CNL_CNL_CNL,
        /*  11      */ CWORD_CWORD_CWORD_CWORD,
        /*  12      */ CWORD_CWORD_CWORD_CWORD,
+#if ENABLE_ASH_IGNORE_CR
+        /*  13      */ CSPCL_CWORD_CWORD_CWORD,
+#else
        /*  13      */ CWORD_CWORD_CWORD_CWORD,
+#endif
        /*  14      */ CWORD_CWORD_CWORD_CWORD,
        /*  15      */ CWORD_CWORD_CWORD_CWORD,
        /*  16      */ CWORD_CWORD_CWORD_CWORD,
@@ -11846,7 +11866,9 @@ preadbuffer(void)
                more--;
                c = *q;
-                if (c == '\0' || (ENABLE_PLATFORM_MINGW32 && c == '\r')) {
+                /* Remove CR from input buffer as an alternative to ASH_IGNORE_CR. */
+                if (c == '\0' || (c == '\r' &&
+                                        ENABLE_PLATFORM_MINGW32 && !ENABLE_ASH_IGNORE_CR)) {
                        memmove(q, q + 1, more);
                } else {
                        q++;
@@ -11934,6 +11956,9 @@ static int __pgetc(void)
        else
                c = preadbuffer();
+#if ENABLE_ASH_IGNORE_CR
+        g_parsefile->lastc[2] = g_parsefile->lastc[1];
+#endif
        g_parsefile->lastc[1] = g_parsefile->lastc[0];
        g_parsefile->lastc[0] = c;
@@ -11971,11 +11996,24 @@ pgetc_eatbnl(void)
        int c;
        while ((c = pgetc()) == '\\') {
+#if !ENABLE_ASH_IGNORE_CR
                if (pgetc() != '\n') {
                        pungetc();
                        break;
                }
+#else
+                int c2 = pgetc();
+                if (c2 == '\r') {
+                        if (pgetc() == '\n')
+                                goto eatbnl;
+                        pungetc();
+                }
+                if (c2 != '\n') {
+                        pungetc();
+                        break;
+                }
+ eatbnl:
+#endif
                nlprompt();
        }
@@ -14114,7 +14152,7 @@ xxreadtoken(void)
        setprompt_if(needprompt, 2);
        for (;;) {                      /* until token or start of word found */
                c = pgetc_eatbnl();
-                if (c == ' ' || c == '\t')
+                if (c == ' ' || c == '\t' || (ENABLE_ASH_IGNORE_CR && c == '\r'))
                        continue;
                if (c == '#') {
author	Ron Yorston <rmy@pobox.com>	2023-02-07 10:03:04 +0000
committer	Ron Yorston <rmy@pobox.com>	2023-02-07 10:03:04 +0000
commit	4b894b60ae377ee2a04ae209c14bb1e714ecf6f4 (patch)
tree	8f42f36eb0a9dad2c712f7b0e07aff7dee059217 /shell
parent	c49b8b7052bb1df3ed5f9aeba61a7965d9c77f83 (diff)
download	busybox-w32-4b894b60ae377ee2a04ae209c14bb1e714ecf6f4.tar.gz busybox-w32-4b894b60ae377ee2a04ae209c14bb1e714ecf6f4.tar.bz2 busybox-w32-4b894b60ae377ee2a04ae209c14bb1e714ecf6f4.zip

diff --git a/shell/ash.c b/shell/ash.c index a24ed0ac8..63eb4947c 100644 --- a/shell/ash.c +++ b/shell/ash.c
@@ -192,6 +192,14 @@
192	//config: Enable support for the 'nocaseglob' option, which allows	192	//config: Enable support for the 'nocaseglob' option, which allows
193	//config: case-insensitive filename globbing.	193	//config: case-insensitive filename globbing.
194	//config:	194	//config:
		195	//config:config ASH_IGNORE_CR
		196	//config: bool "Ignore CR in scripts"
		197	//config: default y
		198	//config: depends on (ASH \|\| SH_IS_ASH \|\| BASH_IS_ASH) && PLATFORM_MINGW32
		199	//config: help
		200	//config: Allow CRs to be ignored when shell scripts are parsed. This
		201	//config: makes it possible to run scripts with CRLF line endings.
		202	//config:
195	//config:endif # ash options	203	//config:endif # ash options
196		204
197	//applet:IF_ASH(APPLET(ash, BB_DIR_BIN, BB_SUID_DROP))	205	//applet:IF_ASH(APPLET(ash, BB_DIR_BIN, BB_SUID_DROP))
@@ -762,7 +770,11 @@ struct strpush {
762	struct strpush *spfree;	770	struct strpush *spfree;
763		771
764	/* Remember last two characters for pungetc. */	772	/* Remember last two characters for pungetc. */
		773	#if ENABLE_ASH_IGNORE_CR
		774	int lastc[3]; /* Ignoring CRs needs more pungetc. */
		775	#else
765	int lastc[2];	776	int lastc[2];
		777	#endif
766		778
767	/* Number of outstanding calls to pungetc. */	779	/* Number of outstanding calls to pungetc. */
768	int unget;	780	int unget;
@@ -787,7 +799,11 @@ struct parsefile {
787	struct strpush *spfree;	799	struct strpush *spfree;
788		800
789	/* Remember last two characters for pungetc. */	801	/* Remember last two characters for pungetc. */
		802	#if ENABLE_ASH_IGNORE_CR
		803	int lastc[3]; /* Ignoring CRs needs more pungetc. */
		804	#else
790	int lastc[2];	805	int lastc[2];
		806	#endif
791		807
792	/* Number of outstanding calls to pungetc. */	808	/* Number of outstanding calls to pungetc. */
793	int unget;	809	int unget;
@@ -3596,7 +3612,11 @@ static const uint8_t syntax_index_table[] ALIGN1 = {
3596	/* 10 "\n" */ CNL_CNL_CNL_CNL,	3612	/* 10 "\n" */ CNL_CNL_CNL_CNL,
3597	/* 11 */ CWORD_CWORD_CWORD_CWORD,	3613	/* 11 */ CWORD_CWORD_CWORD_CWORD,
3598	/* 12 */ CWORD_CWORD_CWORD_CWORD,	3614	/* 12 */ CWORD_CWORD_CWORD_CWORD,
		3615	#if ENABLE_ASH_IGNORE_CR
		3616	/* 13 */ CSPCL_CWORD_CWORD_CWORD,
		3617	#else
3599	/* 13 */ CWORD_CWORD_CWORD_CWORD,	3618	/* 13 */ CWORD_CWORD_CWORD_CWORD,
		3619	#endif
3600	/* 14 */ CWORD_CWORD_CWORD_CWORD,	3620	/* 14 */ CWORD_CWORD_CWORD_CWORD,
3601	/* 15 */ CWORD_CWORD_CWORD_CWORD,	3621	/* 15 */ CWORD_CWORD_CWORD_CWORD,
3602	/* 16 */ CWORD_CWORD_CWORD_CWORD,	3622	/* 16 */ CWORD_CWORD_CWORD_CWORD,
@@ -11846,7 +11866,9 @@ preadbuffer(void)
11846	more--;	11866	more--;
11847		11867
11848	c = *q;	11868	c = *q;
11849	if (c == '\0' \|\| (ENABLE_PLATFORM_MINGW32 && c == '\r')) {	11869	/* Remove CR from input buffer as an alternative to ASH_IGNORE_CR. */
		11870	if (c == '\0' \|\| (c == '\r' &&
		11871	ENABLE_PLATFORM_MINGW32 && !ENABLE_ASH_IGNORE_CR)) {
11850	memmove(q, q + 1, more);	11872	memmove(q, q + 1, more);
11851	} else {	11873	} else {
11852	q++;	11874	q++;
@@ -11934,6 +11956,9 @@ static int __pgetc(void)
11934	else	11956	else
11935	c = preadbuffer();	11957	c = preadbuffer();
11936		11958
		11959	#if ENABLE_ASH_IGNORE_CR
		11960	g_parsefile->lastc[2] = g_parsefile->lastc[1];
		11961	#endif
11937	g_parsefile->lastc[1] = g_parsefile->lastc[0];	11962	g_parsefile->lastc[1] = g_parsefile->lastc[0];
11938	g_parsefile->lastc[0] = c;	11963	g_parsefile->lastc[0] = c;
11939		11964
@@ -11971,11 +11996,24 @@ pgetc_eatbnl(void)
11971	int c;	11996	int c;
11972		11997
11973	while ((c = pgetc()) == '\\') {	11998	while ((c = pgetc()) == '\\') {
		11999	#if !ENABLE_ASH_IGNORE_CR
11974	if (pgetc() != '\n') {	12000	if (pgetc() != '\n') {
11975	pungetc();	12001	pungetc();
11976	break;	12002	break;
11977	}	12003	}
11978		12004	#else
		12005	int c2 = pgetc();
		12006	if (c2 == '\r') {
		12007	if (pgetc() == '\n')
		12008	goto eatbnl;
		12009	pungetc();
		12010	}
		12011	if (c2 != '\n') {
		12012	pungetc();
		12013	break;
		12014	}
		12015	eatbnl:
		12016	#endif
11979	nlprompt();	12017	nlprompt();
11980	}	12018	}
11981		12019
@@ -14114,7 +14152,7 @@ xxreadtoken(void)
14114	setprompt_if(needprompt, 2);	14152	setprompt_if(needprompt, 2);
14115	for (;;) { /* until token or start of word found */	14153	for (;;) { /* until token or start of word found */
14116	c = pgetc_eatbnl();	14154	c = pgetc_eatbnl();
14117	if (c == ' ' \|\| c == '\t')	14155	if (c == ' ' \|\| c == '\t' \|\| (ENABLE_ASH_IGNORE_CR && c == '\r'))
14118	continue;	14156	continue;
14119		14157
14120	if (c == '#') {	14158	if (c == '#') {