Fix handling of conversion problem in CP932 module

5 files changed, 90 insertions, 43 deletions

diff --git a/ChangeLog b/ChangeLog
index 70a4b86..549a23d 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,11 @@
 2011-05-14  Ulrich Drepper  <drepper@gmail.com>
 
+	[BZ #12601]
+	* iconvdata/cp932.c (BODY to UCS4): Fix incrementing inptr in case of
+	two-byte sequence errors.
+	* iconvdata/Makefile (tests): Add bug-iconv8.
+	* iconvdata/bug-iconv8.c: New file.
+
 	[BZ #12626]
 	* sysdeps/generic/elf/backtracesymsfd.c (__backtrace_symbols_fd): Move
 	buf2 definition.
diff --git a/NEWS b/NEWS
index 4abf3f0..3db9aea 100644
--- a/NEWS
+++ b/NEWS
@@ -12,9 +12,9 @@ Version 2.14
   386, 11257, 11258, 11487, 11532, 11578, 11653, 11668, 11724, 11945, 11947,
   12052, 12158, 12178, 12200, 12346, 12393, 12420, 12432, 12445, 12449,
   12454, 12460, 12469, 12489, 12509, 12510, 12511, 12518, 12527, 12541,
-  12545, 12551, 12583, 12587, 12597, 12611, 12625, 12626, 12631, 12650,
-  12653, 12655, 12660, 12681, 12685, 12711, 12713, 12714, 12717, 12723,
-  12724, 12734, 12738
+  12545, 12551, 12583, 12587, 12597, 12601, 12611, 12625, 12626, 12631,
+  12650, 12653, 12655, 12660, 12681, 12685, 12711, 12713, 12714, 12717,
+  12723, 12724, 12734, 12738
 
 * The RPC implementation in libc is obsoleted.  Old programs keep working
   but new programs cannot be linked with the routines in libc anymore.
diff --git a/iconvdata/Makefile b/iconvdata/Makefile
index d79756c..e0fe46a 100644
--- a/iconvdata/Makefile
+++ b/iconvdata/Makefile
@@ -68,7 +68,7 @@ include ../Makeconfig
 
 ifeq (yes,$(build-shared))
 tests = bug-iconv1 bug-iconv2 tst-loading tst-e2big tst-iconv4 bug-iconv4 \
-	tst-iconv6 bug-iconv5 bug-iconv6 tst-iconv7
+	tst-iconv6 bug-iconv5 bug-iconv6 tst-iconv7 bug-iconv8
 ifeq ($(have-thread-library),yes)
 tests += bug-iconv3
 endif
diff --git a/iconvdata/bug-iconv8.c b/iconvdata/bug-iconv8.c
new file mode 100644
index 0000000..54c6be2
--- /dev/null
+++ b/iconvdata/bug-iconv8.c
@@ -0,0 +1,43 @@
+// BZ 12601
+#include <stdio.h>
+#include <errno.h>
+#include <iconv.h>
+
+static int
+do_test (void)
+{
+   iconv_t cd;
+   char in[] = "\x83\xd9";
+   char out[256];
+   char *inbuf;
+   size_t inbytesleft;
+   char *outbuf;
+   size_t outbytesleft;
+   size_t ret;
+
+   inbuf = in;
+   inbytesleft = sizeof(in) - 1;
+   outbuf = out;
+   outbytesleft = sizeof(out);
+
+   cd = iconv_open("utf-8", "cp932");
+   ret = iconv(cd, &inbuf, &inbytesleft, &outbuf, &outbytesleft);
+   iconv_close(cd);
+
+   printf("result: %ld %d %ld %d\n", ret, errno, inbytesleft, inbuf[0]);
+
+   /*
+    * result: -1 84 0 0        (84=EILSEQ)
+    *
+    * Error is returnd but inbuf is consumed.
+    *
+    * \x83\xd9 is valid shift-jis sequence but no character is assigned
+    * to it.
+    */
+
+   return (ret != -1 || errno != EILSEQ
+	   || inbytesleft != 2 || inbuf[0] != in[0]);
+}
+
+#define TEST_FUNCTION do_test ()
+#include "../test-skeleton.c"
diff --git a/iconvdata/cp932.c b/iconvdata/cp932.c
index 90ecad9..00d83d4 100644
--- a/iconvdata/cp932.c
+++ b/iconvdata/cp932.c
@@ -1,5 +1,5 @@
 /* Mapping tables for CP932 handling.
-   Copyright (C) 1997,1998,1999,2000,2001,2003 Free Software Foundation, Inc.
+   Copyright (C) 1997-2001,2003,2011 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
    Contributed by MORIYAMA Masayuki <msyk@mtg.biglobe.ne.jp>, 2003.
 
@@ -4549,8 +4549,8 @@ static const char from_ucs4_extra[229][2] =
       ++inptr;								      \
     else if (ch >= 0xa1 && ch <= 0xdf)                                        \
       {                                                                       \
-        ch += 0xfec0;                                                         \
-        ++inptr;                                                              \
+	ch += 0xfec0;                                                         \
+	++inptr;                                                              \
       }									      \
     else if (__builtin_expect (ch, 0) == 0xa0				      \
 	     || __builtin_expect (ch <= 0x80, 0)			      \
@@ -4588,65 +4588,63 @@ static const char from_ucs4_extra[229][2] =
 	if (__builtin_expect (ch2 < 0x40, 0)				      \
 	    || __builtin_expect (ch2 > 0xfc, 0)				      \
 	    || __builtin_expect (ch2 == 0x7f, 0)			      \
-	    || (__builtin_expect (idx > 0x84be, 0) && idx < 0x8740)      \
-	    || (__builtin_expect (idx > 0x879c, 0) && idx < 0x889f)      \
-	    || (__builtin_expect (idx > 0x88fc, 0) && idx < 0x8940)      \
-	    || (__builtin_expect (idx > 0x9ffc, 0) && idx < 0xe040)      \
-	    || (__builtin_expect (idx > 0xeaa4, 0) && idx < 0xed40)      \
-	    || (__builtin_expect (idx > 0xeefc, 0) && idx < 0xf040)      \
+	    || (__builtin_expect (idx > 0x84be, 0) && idx < 0x8740)	      \
+	    || (__builtin_expect (idx > 0x879c, 0) && idx < 0x889f)	      \
+	    || (__builtin_expect (idx > 0x88fc, 0) && idx < 0x8940)	      \
+	    || (__builtin_expect (idx > 0x9ffc, 0) && idx < 0xe040)	      \
+	    || (__builtin_expect (idx > 0xeaa4, 0) && idx < 0xed40)	      \
+	    || (__builtin_expect (idx > 0xeefc, 0) && idx < 0xf040)	      \
 	    || __builtin_expect (idx > 0xfc4b, 0))			      \
 	  {								      \
 	    /* This is illegal.  */					      \
 	    if (! ignore_errors_p ())					      \
 	      {								      \
-	        /* This is an illegal character.  */			      \
-	        result = __GCONV_ILLEGAL_INPUT;				      \
-	        break;							      \
+		/* This is an illegal character.  */			      \
+		result = __GCONV_ILLEGAL_INPUT;				      \
+		break;							      \
 	      }								      \
 									      \
 	    ++inptr;							      \
 	    ++*irreversible;						      \
 	    continue;							      \
 	  }								      \
-	else								      \
-	  {								      \
-	    /* We could pack the data a bit more dense.  The second	      \
-	       byte will never be 0x7f and it will also be never	      \
-	       >0xfc.  But this would mean yet more `if's.  */		      \
-	    if (idx <= 0x84be)						      \
-	      ch = cjk_block1[(ch - 0x81) * 192 + ch2 - 0x40];		      \
-	    else if (idx <= 0x879c)					      \
-	      ch = cjk_block2[(ch - 0x87) * 192 + ch2 - 0x40];		      \
-	    else if (idx <= 0x88fc)					      \
-	      ch = cjk_block3[(ch - 0x88) * 192 + ch2 - 0x9f];		      \
-	    else if (idx <= 0x9ffc)					      \
-	      ch = cjk_block4[(ch - 0x89) * 192 + ch2 - 0x40];		      \
-	    else if (idx <= 0xeaa4)					      \
-	      ch = cjk_block5[(ch - 0xe0) * 192 + ch2 - 0x40];		      \
-	    else if (idx <= 0xeefc)					      \
-	      ch = cjk_block6[(ch - 0xed) * 192 + ch2 - 0x40];		      \
-	    else if (idx <= 0xf9fc)					      \
-	      ch = (ch-0xf0)*188 + ch2-((ch2<0x7f)?0x40:0x41) + 0xe000;	      \
-	    else							      \
-	      ch = cjk_block7[(ch - 0xfa) * 192 + ch2 - 0x40];		      \
 									      \
-	    inptr += 2;							      \
-	  }								      \
+	/* We could pack the data a bit more dense.  The second		      \
+	   byte will never be 0x7f and it will also be never		      \
+	   >0xfc.  But this would mean yet more `if's.  */		      \
+	if (idx <= 0x84be)						      \
+	  ch = cjk_block1[(ch - 0x81) * 192 + ch2 - 0x40];		      \
+	else if (idx <= 0x879c)						      \
+	  ch = cjk_block2[(ch - 0x87) * 192 + ch2 - 0x40];		      \
+	else if (idx <= 0x88fc)						      \
+	  ch = cjk_block3[(ch - 0x88) * 192 + ch2 - 0x9f];		      \
+	else if (idx <= 0x9ffc)						      \
+	  ch = cjk_block4[(ch - 0x89) * 192 + ch2 - 0x40];		      \
+	else if (idx <= 0xeaa4)						      \
+	  ch = cjk_block5[(ch - 0xe0) * 192 + ch2 - 0x40];		      \
+	else if (idx <= 0xeefc)						      \
+	  ch = cjk_block6[(ch - 0xed) * 192 + ch2 - 0x40];		      \
+	else if (idx <= 0xf9fc)						      \
+	  ch = (ch-0xf0)*188 + ch2-((ch2<0x7f)?0x40:0x41) + 0xe000;	      \
+	else								      \
+	  ch = cjk_block7[(ch - 0xfa) * 192 + ch2 - 0x40];		      \
 									      \
 	if (__builtin_expect (ch, 1) == 0)				      \
 	  {								      \
 	    /* This is an illegal character.  */			      \
 	    if (! ignore_errors_p ())					      \
 	      {								      \
-	        /* This is an illegal character.  */			      \
-	        result = __GCONV_ILLEGAL_INPUT;				      \
-	        break;							      \
+		/* This is an illegal character.  */			      \
+		result = __GCONV_ILLEGAL_INPUT;				      \
+		break;							      \
 	      }								      \
 									      \
 	    inptr += 2;							      \
 	    ++*irreversible;						      \
 	    continue;							      \
 	  }								      \
+									      \
+	inptr += 2;							      \
       }									      \
 									      \
     put32 (outptr, ch);							      \
@@ -4674,7 +4672,7 @@ static const char from_ucs4_extra[229][2] =
 	else if (ch >= 0x2010 && ch <= 0x9fa0)				      \
 	  cp = from_ucs4_cjk[ch - 0x2010];				      \
 	else if (ch >= 0xe000 && ch <= 0xe757)				      \
-          {								      \
+	  {								      \
 	    pua[0] = (ch - 0xe000) / 188 + 0xf0;			      \
 	    pua[1] = (ch - 0xe000) % 188 + 0x40;			      \
 	    if (pua[1] > 0x7e)						      \