From 1ff7002b1d229c678fdffebec602fb4c54439a31 Mon Sep 17 00:00:00 2001
From: Ron Yorston <rmy@pobox.com>
Date: Fri, 24 Jan 2020 13:16:45 +0000
Subject: [PATCH] xargs: fix handling of quoted arguments, closes 11441

As reported in bug 11441 when presented with a large number of quoted
arguments xargs can return 'argument line too long':

   seq 10000 29999 | sed -e 's/^/"/' -e 's/$/"/' | busybox xargs echo

This happens because the variant of process_stdin() which handles quoted
arguments doesn't preserve state between calls.  If the allowed number
of characters is exceeded part way through a quoted argument the next
call to process_stdin() incorrectly treats the terminating quote as a
starting quote, thus quoting all of the argument separators.

function                                             old     new   delta
process_stdin                                        274     303     +29
xargs_main                                           731     745     +14
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 2/0 up/down: 43/0)               Total: 43 bytes

Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
---
 findutils/xargs.c     | 27 ++++++++++++++++++---------
 testsuite/xargs.tests |  9 +++++++++
 2 files changed, 27 insertions(+), 9 deletions(-)

diff --git a/findutils/xargs.c b/findutils/xargs.c
index 726315803..4fb306bb8 100644
--- a/findutils/xargs.c
+++ b/findutils/xargs.c
@@ -114,17 +114,28 @@ struct globals {
 	int max_procs;
 #endif
 	smalluint xargs_exitcode;
+#if ENABLE_FEATURE_XARGS_SUPPORT_QUOTES
+#define NORM      0
+#define QUOTE     1
+#define BACKSLASH 2
+#define SPACE     4
+	smalluint process_stdin__state;
+	char process_stdin__q;
+#endif
 } FIX_ALIASING;
 #define G (*(struct globals*)bb_common_bufsiz1)
 #define INIT_G() do { \
 	setup_common_bufsiz(); \
-	G.eof_str = NULL; /* need to clear by hand because we are NOEXEC applet */ \
+	IF_FEATURE_XARGS_SUPPORT_REPL_STR(G.repl_str = "{}";) \
+	IF_FEATURE_XARGS_SUPPORT_REPL_STR(G.eol_ch = '\n';) \
+	/* Even zero values are set because we are NOEXEC applet */ \
+	G.eof_str = NULL; \
 	G.idx = 0; \
 	IF_FEATURE_XARGS_SUPPORT_PARALLEL(G.running_procs = 0;) \
 	IF_FEATURE_XARGS_SUPPORT_PARALLEL(G.max_procs = 1;) \
 	G.xargs_exitcode = 0; \
-	IF_FEATURE_XARGS_SUPPORT_REPL_STR(G.repl_str = "{}";) \
-	IF_FEATURE_XARGS_SUPPORT_REPL_STR(G.eol_ch = '\n';) \
+	IF_FEATURE_XARGS_SUPPORT_QUOTES(G.process_stdin__state = NORM;) \
+	IF_FEATURE_XARGS_SUPPORT_QUOTES(G.process_stdin__q = '\0';) \
 } while (0)
 
 
@@ -257,12 +268,8 @@ static void store_param(char *s)
 #if ENABLE_FEATURE_XARGS_SUPPORT_QUOTES
 static char* FAST_FUNC process_stdin(int n_max_chars, int n_max_arg, char *buf)
 {
-#define NORM      0
-#define QUOTE     1
-#define BACKSLASH 2
-#define SPACE     4
-	char q = '\0';             /* quote char */
-	char state = NORM;
+#define q     G.process_stdin__q
+#define state G.process_stdin__state
 	char *s = buf;             /* start of the word */
 	char *p = s + strlen(buf); /* end of the word */
 
@@ -339,6 +346,8 @@ static char* FAST_FUNC process_stdin(int n_max_chars, int n_max_arg, char *buf)
 	/* store_param(NULL) - caller will do it */
 	dbg_msg("return:'%s'", s);
 	return s;
+#undef q
+#undef state
 }
 #else
 /* The variant does not support single quotes, double quotes or backslash */
diff --git a/testsuite/xargs.tests b/testsuite/xargs.tests
index 2d0a201b7..855b33bc2 100755
--- a/testsuite/xargs.tests
+++ b/testsuite/xargs.tests
@@ -41,4 +41,13 @@ testing "xargs -sNUM test 2" \
 	"echo 1 2 3 4 5 6 7 8 9 0\n""echo 1 2 3 4 5 6 7 8 9\n""echo 1 00\n" \
 	"" "2 3 4 5 6 7 8 9 0 2 3 4 5 6 7 8 9 00\n"
 
+# see that we don't get "argument line too long",
+# but do see the last word, 99999, instead
+optional FEATURE_XARGS_SUPPORT_QUOTES
+testing "xargs argument line too long" \
+	"seq 10000 99999 | sed -e 's/^/\"/' -e 's/$/\"/' | xargs echo | grep -o 99999; echo \$?" \
+	"99999\n0\n" \
+	"" ""
+SKIP=
+
 exit $FAILCOUNT