[git commit] awk: fix splitting with default FS

[Denys Vlasenko]

commit: https://git.busybox.net/busybox/commit/?id=84ff1825dd82e8de45020e3def34d1430d8e5a99
branch: https://git.busybox.net/busybox/commit/?id=refs/heads/master

function                                             old     new   delta
awk_split                                            543     544      +1

Signed-off-by: Denys Vlasenko <vda.linux at googlemail.com>
---
 editors/awk.c       | 13 ++++++++-----
 testsuite/awk.tests |  7 +++++++
 2 files changed, 15 insertions(+), 5 deletions(-)

diff --git a/editors/awk.c b/editors/awk.c
index 2af823808..b3748b502 100644
--- a/editors/awk.c
+++ b/editors/awk.c
@@ -2049,13 +2049,17 @@ static int awk_split(const char *s, node *spl, char **slist)
 		}
 		return n;
 	}
-	/* space split */
+	/* space split: "In the special case that FS is a single space,
+	 * fields are separated by runs of spaces and/or tabs and/or newlines"
+	 */
 	while (*s) {
-		s = skip_whitespace(s);
+		/* s = skip_whitespace(s); -- WRONG (also skips \v \f \r) */
+		while (*s == ' ' || *s == '\t' || *s == '\n')
+			s++;
 		if (!*s)
 			break;
 		n++;
-		while (*s && !isspace(*s))
+		while (*s && !(*s == ' ' || *s == '\t' || *s == '\n'))
 			*s1++ = *s++;
 		*s1++ = '\0';
 	}
@@ -2304,7 +2308,6 @@ static int awk_getline(rstream *rsm, var *v)
 			setvar_i(intvar[ERRNO], errno);
 		}
 		b[p] = '\0';
-
 	} while (p > pp);
 
 	if (p == 0) {
@@ -3145,7 +3148,7 @@ static var *evaluate(node *op, var *res)
 			/* make sure that we never return a temp var */
 			if (L.v == TMPVAR0)
 				L.v = res;
-			/* if source is a temporary string, jusk relink it to dest */
+			/* if source is a temporary string, just relink it to dest */
 			if (R.v == TMPVAR1
 			 && !(R.v->type & VF_NUMBER)
 				/* Why check !NUMBER? if R.v is a number but has cached R.v->string,
diff --git a/testsuite/awk.tests b/testsuite/awk.tests
index ddc51047b..8ab1c6891 100755
--- a/testsuite/awk.tests
+++ b/testsuite/awk.tests
@@ -540,4 +540,11 @@ testing 'awk assign while assign' \
 │    trim/eff : 57.02%/26, 0.00%                     │          [cpu000:100%]
 └────────────────────────────────────────────────────┘^C"
 
+# If field separator FS=' ' (default), fields are split only on
+# space or tab or linefeed, NOT other whitespace.
+testing 'awk does not split on CR (char 13)' \
+	"awk '{ \$1=\$0; print }'" \
+	'word1 word2 word3\r word2 word3\r\n' \
+	'' 'word1 word2 word3\r'
+
 exit $FAILCOUNT