Patchwork fix 19432 - iconv rejects redundant escape sequences in IBM90{0,3,5,7,9}

login
register
mail settings
Submitter Martin Sebor
Date Jan. 6, 2016, 4:07 a.m.
Message ID <568C92E5.8070206@gmail.com>
Download mbox | patch
Permalink /patch/10239/
State New
Headers show

Comments

Martin Sebor - Jan. 6, 2016, 4:07 a.m.
Quoting the bug Description:

   In the problem report corresponding to bug 17197 submitted
   against the RHEL GLIBC, even though the focus of the customer's
   complaint is on GLIBC emitting redundant shift sequences,
   the provided test case implies the customer is actually having
   a problem interpreting already encoded data (i.e., with iconv()
   treating the duplicate/redundant escape sequences as errors).
   If iconv() accepted the duplicate sequences the customer wouldn't
   be affected.

   The solution implemented in response to bug 17197, changing
   the converter to avoid emitting the redundant escape sequences,
   while a worthwhile improvement, is not sufficient to resolve
   the complaint because there may already be IBM930-encoded data
   containing redundant shift sequences that need to be converted.
   The current fix will not help users avoid the errors when
   converting such data.

   Looking for guidance, I checked POSIX to see if it provides
   support for the current behavior (rejecting redundant escape
   sequences) in POSIX.  I don't see anything in POSIX to sanction
   rejecting such input.  Even if POSIX did allow for such behavior,
   since other GLIBC converters accept redundant escape sequences,
   it seems that the IBM930 converter should accept them as well
   for consistency.  Doing otherwise makes the iconv interfaces
   needlessly error-prone to use.

The attached patch removes the checking for redundant escape
sequences from the converters for the five encoders to allow
them to successfully convert data containing such sequences.

Tested on powerpc64le-redhat-linux-gnu.

Martin
Martin Sebor - Jan. 12, 2016, 9:03 p.m.
Ping: I'm still looking for a review/feedback on the following
patch:

   https://sourceware.org/ml/libc-alpha/2016-01/msg00074.html

Thanks
Martin

On 01/05/2016 09:07 PM, Martin Sebor wrote:
> Quoting the bug Description:
>
>    In the problem report corresponding to bug 17197 submitted
>    against the RHEL GLIBC, even though the focus of the customer's
>    complaint is on GLIBC emitting redundant shift sequences,
>    the provided test case implies the customer is actually having
>    a problem interpreting already encoded data (i.e., with iconv()
>    treating the duplicate/redundant escape sequences as errors).
>    If iconv() accepted the duplicate sequences the customer wouldn't
>    be affected.
>
>    The solution implemented in response to bug 17197, changing
>    the converter to avoid emitting the redundant escape sequences,
>    while a worthwhile improvement, is not sufficient to resolve
>    the complaint because there may already be IBM930-encoded data
>    containing redundant shift sequences that need to be converted.
>    The current fix will not help users avoid the errors when
>    converting such data.
>
>    Looking for guidance, I checked POSIX to see if it provides
>    support for the current behavior (rejecting redundant escape
>    sequences) in POSIX.  I don't see anything in POSIX to sanction
>    rejecting such input.  Even if POSIX did allow for such behavior,
>    since other GLIBC converters accept redundant escape sequences,
>    it seems that the IBM930 converter should accept them as well
>    for consistency.  Doing otherwise makes the iconv interfaces
>    needlessly error-prone to use.
>
> The attached patch removes the checking for redundant escape
> sequences from the converters for the five encoders to allow
> them to successfully convert data containing such sequences.
>
> Tested on powerpc64le-redhat-linux-gnu.
>
> Martin
Carlos O'Donell - Jan. 13, 2016, 12:47 a.m.
On 01/05/2016 11:07 PM, Martin Sebor wrote:
> Quoting the bug Description:
> 
>   In the problem report corresponding to bug 17197 submitted
>   against the RHEL GLIBC, even though the focus of the customer's
>   complaint is on GLIBC emitting redundant shift sequences,
>   the provided test case implies the customer is actually having
>   a problem interpreting already encoded data (i.e., with iconv()
>   treating the duplicate/redundant escape sequences as errors).
>   If iconv() accepted the duplicate sequences the customer wouldn't
>   be affected.
> 
>   The solution implemented in response to bug 17197, changing
>   the converter to avoid emitting the redundant escape sequences,
>   while a worthwhile improvement, is not sufficient to resolve
>   the complaint because there may already be IBM930-encoded data
>   containing redundant shift sequences that need to be converted.
>   The current fix will not help users avoid the errors when
>   converting such data.
> 
>   Looking for guidance, I checked POSIX to see if it provides
>   support for the current behavior (rejecting redundant escape
>   sequences) in POSIX.  I don't see anything in POSIX to sanction
>   rejecting such input.  Even if POSIX did allow for such behavior,
>   since other GLIBC converters accept redundant escape sequences,
>   it seems that the IBM930 converter should accept them as well
>   for consistency.  Doing otherwise makes the iconv interfaces
>   needlessly error-prone to use.
> 
> The attached patch removes the checking for redundant escape
> sequences from the converters for the five encoders to allow
> them to successfully convert data containing such sequences.
> 
> Tested on powerpc64le-redhat-linux-gnu.
> 
> Martin
> 
> glibc-19432.patch
> 
> 
> 2016-01-05  Martin Sebor  <msebor@redhat.com>
> 
> 	[BZ #19432]
> 	* iconvdata/Makefile: Add bug-iconv11. 
> 	* iconvdata/bug-iconv11.c: New test.
> 	* iconvdata/ibm930.c: Do not reject redundant shift sequences.
> 	* iconvdata/ibm933.c: Same.
> 	* iconvdata/ibm935.c: Same.
> 	* iconvdata/ibm937.c: Same.
> 	* iconvdata/ibm939.c: Same.

The strategy and implementation look good to me.

Fix the formatting in the test and check this in.

It's an important part of fixing the loading of these data files
which have redundant shift sequences. Users might not even be aware
of this until they go to load the files later.

> diff --git a/iconvdata/Makefile b/iconvdata/Makefile
> index ae5dde6..357530b 100644
> --- a/iconvdata/Makefile
> +++ b/iconvdata/Makefile
> @@ -68,7 +68,7 @@ modules.so := $(addsuffix .so, $(modules))
>  ifeq (yes,$(build-shared))
>  tests = bug-iconv1 bug-iconv2 tst-loading tst-e2big tst-iconv4 bug-iconv4 \
>  	tst-iconv6 bug-iconv5 bug-iconv6 tst-iconv7 bug-iconv8 bug-iconv9 \
> -	bug-iconv10
> +	bug-iconv10 bug-iconv11

OK.

>  ifeq ($(have-thread-library),yes)
>  tests += bug-iconv3
>  endif
> diff --git a/iconvdata/bug-iconv11.c b/iconvdata/bug-iconv11.c
> new file mode 100644
> index 0000000..3f39780
> --- /dev/null
> +++ b/iconvdata/bug-iconv11.c
> @@ -0,0 +1,116 @@
> +/* bug 19432: iconv rejects redundant escape sequences in IBM903,
> +              IBM905, IBM907, and IBM909 
> +
> +   Copyright (C) 2016 Free Software Foundation, Inc.
> +   This file is part of the GNU C Library.
> +
> +   The GNU C Library is free software; you can redistribute it and/or
> +   modify it under the terms of the GNU Lesser General Public
> +   License as published by the Free Software Foundation; either
> +   version 2.1 of the License, or (at your option) any later version.
> +
> +   The GNU C Library is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +   Lesser General Public License for more details.
> +
> +   You should have received a copy of the GNU Lesser General Public
> +   License along with the GNU C Library; if not, see
> +   <http://www.gnu.org/licenses/>.  */
> +
> +#include <iconv.h>
> +#include <stdio.h>
> +#include <stdlib.h>
> +#include <string.h>
> +#include <errno.h>
> +
> +// The longest test input sequence.

Please use C-style comments.

/* Foo.  */

> +#define MAXINBYTES    8
> +#define MAXOUTBYTES   (MAXINBYTES * MB_LEN_MAX)
> +
> +// Verify that a conversion of the INPUT sequence consisting of
> +// INBYTESLEFT bytes in the encoding specified by the codeset
> +// named by FROM_SET is successful.
> +// Return 0 on success, non-zero on iconv() failure.

Please use C-style block comment.

/* foo.
   bar.  */
  
> +
> +static int
> +test_ibm93x (const char *from_set, const char *input, size_t inbytesleft)
> +{
> +  const char to_set[] = "UTF-8";
> +  iconv_t cd = iconv_open (to_set, from_set);
> +  if (cd == (iconv_t) -1)
> +    {
> +      printf ("iconv_open(\"%s\", \"%s\"): %s\n",
> +              from_set, to_set, strerror (errno));
> +      return 1;
> +    }
> +
> +  char output [MAXOUTBYTES];
> +  size_t outbytesleft = sizeof output;
> +
> +  char *inbuf = (char*)input;
> +  char *outbuf = output;
> +
> +  printf ("iconv(cd, %p, %zu, %p, %zu)\n",
> +          inbuf, inbytesleft, outbuf, outbytesleft);
> +
> +  errno = 0;
> +  size_t ret = iconv (cd, &inbuf, &inbytesleft, &outbuf, &outbytesleft);
> +  printf ("  ==> %td: %s\n"
> +          "  inbuf%+td, inbytesleft=%zu, outbuf%+td, outbytesleft=%zu\n",
> +          ret, strerror (errno),
> +          inbuf - input, inbytesleft, outbuf - output, outbytesleft);
> +
> +  // Return 0 on success, non-zero on iconv() failure.
> +  return ret == (size_t)-1 || errno;
> +}
> +
> +static int
> +do_test (void)
> +{
> +  // State-dependent encodings to exercise.
> +  static const char* const to_code[] = {
> +    "IBM930", "IBM933", "IBM935", "IBM937", "IBM939"
> +  };
> +
> +  static const size_t ncodesets = sizeof to_code / sizeof *to_code;
> +
> +  // A source sequence containing a redundant shift sequence.
> +  //              <SO><SI>G<SI>
> +  static const struct {
> +    char txt[MAXINBYTES];
> +    size_t len;
> +  } input[] = {
> +#define DATA(s) { s, sizeof s - 1 }
> +    // <SI>: denotes the shift-in 1-byte escape sequence, changing
> +    //       the encoder from a sigle-byte encoding to multibyte
> +    // <SO>: denotes the shift-out 1-byte escape sequence, switching
> +    //       the encoder from a multibyte to a single-byte state
> +
> +    DATA ("\x0e"),               // <SI> (not redundant)
> +    DATA ("\x0f"),               // <S0> (redundant with initial state)
> +    DATA ("\x0e\x0e"),           // <SI><SI>
> +    DATA ("\x0e\x0f\x0f"),       // <SI><SO><SO>
> +    DATA ("\x0f\x0f"),           // <SO><SO>
> +    DATA ("\x0f\x0e\x0e"),       // <SO><SI><SI>
> +    DATA ("\x0e\x0f\xc7\x0f"),   // <SI><SO><G><SO>
> +    DATA ("\xc7\x0f")            // <G><SO> (redundant with initial state)
> +  };
> +
> +  static const size_t ninputs = sizeof input / sizeof *input;
> +
> +  int ret = 0;
> +
> +  size_t i, j;
> +
> +  // Iterate over the IBM93x codesets above and exercise each with
> +  // the input sequences above.
> +  for (i = 0; i != ncodesets; ++i)
> +    for (j = 0; j != ninputs; ++j)
> +      ret += test_ibm93x (to_code [i], input [i].txt, input [i].len);
> +
> +  return ret;
> +}

OK.

> +
> +#define TEST_FUNCTION do_test ()
> +#include "../test-skeleton.c"
> diff --git a/iconvdata/ibm930.c b/iconvdata/ibm930.c
> index 1f60962..9d35734 100644
> --- a/iconvdata/ibm930.c
> +++ b/iconvdata/ibm930.c
> @@ -110,24 +110,14 @@ enum
>  									      \
>      if (__builtin_expect (ch, 0) == SO)					      \
>        {									      \
> -	/* Shift OUT, change to DBCS converter.  */			      \
> -	if (curcs == db)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \

OK.

>  	curcs = db;							      \
>  	++inptr;							      \
>  	continue;							      \
>        }									      \
>      else if (__builtin_expect (ch, 0) == SI)				      \
>        {									      \
> -	/* Shift IN, change to SBCS converter */			      \
> -	if (curcs == sb)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \

OK.

>  	curcs = sb;							      \
>  	++inptr;							      \
>  	continue;							      \
> diff --git a/iconvdata/ibm933.c b/iconvdata/ibm933.c
> index c984cd1..669e357 100644
> --- a/iconvdata/ibm933.c
> +++ b/iconvdata/ibm933.c
> @@ -108,24 +108,14 @@ enum
>  									      \
>      if (__builtin_expect (ch, 0) == SO)					      \
>        {									      \
> -	/* Shift OUT, change to DBCS converter.  */			      \
> -	if (curcs == db)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \
>  	curcs = db;							      \

OK.

>  	++inptr;							      \
>  	continue;							      \
>        }									      \
>      else if (__builtin_expect (ch, 0) == SI)				      \
>        {									      \
> -	/* Shift IN, change to SBCS converter.  */			      \
> -	if (curcs == sb)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \

OK.

>  	curcs = sb;							      \
>  	++inptr;							      \
>  	continue;							      \
> diff --git a/iconvdata/ibm935.c b/iconvdata/ibm935.c
> index 07f4bf2..2d3065e 100644
> --- a/iconvdata/ibm935.c
> +++ b/iconvdata/ibm935.c
> @@ -109,24 +109,14 @@ enum
>  									      \
>      if (__builtin_expect(ch, 0) == SO)					      \
>        {									      \
> -	/* Shift OUT, change to DBCS converter.  */			      \
> -	if (curcs == db)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \

OK.

>  	curcs = db;							      \
>  	++inptr;							      \
>  	continue;							      \
>        }									      \
>      else if (__builtin_expect (ch, 0) == SI)				      \
>        {									      \
> -	/* Shift IN, change to SBCS converter.  */			      \
> -	if (curcs == sb)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \

OK.

>  	curcs = sb;							      \
>  	++inptr;							      \
>  	continue;							      \
> diff --git a/iconvdata/ibm937.c b/iconvdata/ibm937.c
> index 88344a4..3ed6479 100644
> --- a/iconvdata/ibm937.c
> +++ b/iconvdata/ibm937.c
> @@ -109,24 +109,14 @@ enum
>  									      \
>      if (__builtin_expect (ch, 0) == SO)					      \
>        {									      \
> -	/* Shift OUT, change to DBCS converter.  */			      \
> -	if (curcs == db)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \

OK.

>  	curcs = db;							      \
>  	++inptr;							      \
>  	continue;							      \
>        }									      \
>      else if (__builtin_expect (ch, 0) == SI)				      \
>        {									      \
> -	/* Shift IN, change to SBCS converter.  */			      \
> -	if (curcs == sb)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \

OK.

>  	curcs = sb;							      \
>  	++inptr;							      \
>  	continue;							      \
> diff --git a/iconvdata/ibm939.c b/iconvdata/ibm939.c
> index dbdda62..c0a75f7 100644
> --- a/iconvdata/ibm939.c
> +++ b/iconvdata/ibm939.c
> @@ -109,24 +109,14 @@ enum
>  									      \
>      if (__builtin_expect (ch, 0) == SO)					      \
>        {									      \
> -	/* Shift OUT, change to DBCS converter.  */			      \
> -	if (curcs == db)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \

OK.

>  	curcs = db;							      \
>  	++inptr;							      \
>  	continue;							      \
>        }									      \
>      else if (__builtin_expect (ch, 0) == SI)				      \
>        {									      \
> -	/* Shift IN, change to SBCS converter.  */			      \
> -	if (curcs == sb)						      \
> -	  {								      \
> -	    result = __GCONV_ILLEGAL_INPUT;				      \
> -	    break;							      \
> -	  }								      \
> +	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \

OK.

>  	curcs = sb;							      \
>  	++inptr;							      \
>  	continue;							      \

Cheers,
Carlos.

Patch

2016-01-05  Martin Sebor  <msebor@redhat.com>

	[BZ #19432]
	* iconvdata/Makefile: Add bug-iconv11. 
	* iconvdata/bug-iconv11.c: New test.
	* iconvdata/ibm930.c: Do not reject redundant shift sequences.
	* iconvdata/ibm933.c: Same.
	* iconvdata/ibm935.c: Same.
	* iconvdata/ibm937.c: Same.
	* iconvdata/ibm939.c: Same.

diff --git a/iconvdata/Makefile b/iconvdata/Makefile
index ae5dde6..357530b 100644
--- a/iconvdata/Makefile
+++ b/iconvdata/Makefile
@@ -68,7 +68,7 @@  modules.so := $(addsuffix .so, $(modules))
 ifeq (yes,$(build-shared))
 tests = bug-iconv1 bug-iconv2 tst-loading tst-e2big tst-iconv4 bug-iconv4 \
 	tst-iconv6 bug-iconv5 bug-iconv6 tst-iconv7 bug-iconv8 bug-iconv9 \
-	bug-iconv10
+	bug-iconv10 bug-iconv11
 ifeq ($(have-thread-library),yes)
 tests += bug-iconv3
 endif
diff --git a/iconvdata/bug-iconv11.c b/iconvdata/bug-iconv11.c
new file mode 100644
index 0000000..3f39780
--- /dev/null
+++ b/iconvdata/bug-iconv11.c
@@ -0,0 +1,116 @@ 
+/* bug 19432: iconv rejects redundant escape sequences in IBM903,
+              IBM905, IBM907, and IBM909 
+
+   Copyright (C) 2016 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <http://www.gnu.org/licenses/>.  */
+
+#include <iconv.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <errno.h>
+
+// The longest test input sequence.
+#define MAXINBYTES    8
+#define MAXOUTBYTES   (MAXINBYTES * MB_LEN_MAX)
+
+// Verify that a conversion of the INPUT sequence consisting of
+// INBYTESLEFT bytes in the encoding specified by the codeset
+// named by FROM_SET is successful.
+// Return 0 on success, non-zero on iconv() failure.
+
+static int
+test_ibm93x (const char *from_set, const char *input, size_t inbytesleft)
+{
+  const char to_set[] = "UTF-8";
+  iconv_t cd = iconv_open (to_set, from_set);
+  if (cd == (iconv_t) -1)
+    {
+      printf ("iconv_open(\"%s\", \"%s\"): %s\n",
+              from_set, to_set, strerror (errno));
+      return 1;
+    }
+
+  char output [MAXOUTBYTES];
+  size_t outbytesleft = sizeof output;
+
+  char *inbuf = (char*)input;
+  char *outbuf = output;
+
+  printf ("iconv(cd, %p, %zu, %p, %zu)\n",
+          inbuf, inbytesleft, outbuf, outbytesleft);
+
+  errno = 0;
+  size_t ret = iconv (cd, &inbuf, &inbytesleft, &outbuf, &outbytesleft);
+  printf ("  ==> %td: %s\n"
+          "  inbuf%+td, inbytesleft=%zu, outbuf%+td, outbytesleft=%zu\n",
+          ret, strerror (errno),
+          inbuf - input, inbytesleft, outbuf - output, outbytesleft);
+
+  // Return 0 on success, non-zero on iconv() failure.
+  return ret == (size_t)-1 || errno;
+}
+
+static int
+do_test (void)
+{
+  // State-dependent encodings to exercise.
+  static const char* const to_code[] = {
+    "IBM930", "IBM933", "IBM935", "IBM937", "IBM939"
+  };
+
+  static const size_t ncodesets = sizeof to_code / sizeof *to_code;
+
+  // A source sequence containing a redundant shift sequence.
+  //              <SO><SI>G<SI>
+  static const struct {
+    char txt[MAXINBYTES];
+    size_t len;
+  } input[] = {
+#define DATA(s) { s, sizeof s - 1 }
+    // <SI>: denotes the shift-in 1-byte escape sequence, changing
+    //       the encoder from a sigle-byte encoding to multibyte
+    // <SO>: denotes the shift-out 1-byte escape sequence, switching
+    //       the encoder from a multibyte to a single-byte state
+
+    DATA ("\x0e"),               // <SI> (not redundant)
+    DATA ("\x0f"),               // <S0> (redundant with initial state)
+    DATA ("\x0e\x0e"),           // <SI><SI>
+    DATA ("\x0e\x0f\x0f"),       // <SI><SO><SO>
+    DATA ("\x0f\x0f"),           // <SO><SO>
+    DATA ("\x0f\x0e\x0e"),       // <SO><SI><SI>
+    DATA ("\x0e\x0f\xc7\x0f"),   // <SI><SO><G><SO>
+    DATA ("\xc7\x0f")            // <G><SO> (redundant with initial state)
+  };
+
+  static const size_t ninputs = sizeof input / sizeof *input;
+
+  int ret = 0;
+
+  size_t i, j;
+
+  // Iterate over the IBM93x codesets above and exercise each with
+  // the input sequences above.
+  for (i = 0; i != ncodesets; ++i)
+    for (j = 0; j != ninputs; ++j)
+      ret += test_ibm93x (to_code [i], input [i].txt, input [i].len);
+
+  return ret;
+}
+
+#define TEST_FUNCTION do_test ()
+#include "../test-skeleton.c"
diff --git a/iconvdata/ibm930.c b/iconvdata/ibm930.c
index 1f60962..9d35734 100644
--- a/iconvdata/ibm930.c
+++ b/iconvdata/ibm930.c
@@ -110,24 +110,14 @@  enum
 									      \
     if (__builtin_expect (ch, 0) == SO)					      \
       {									      \
-	/* Shift OUT, change to DBCS converter.  */			      \
-	if (curcs == db)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \
 	curcs = db;							      \
 	++inptr;							      \
 	continue;							      \
       }									      \
     else if (__builtin_expect (ch, 0) == SI)				      \
       {									      \
-	/* Shift IN, change to SBCS converter */			      \
-	if (curcs == sb)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \
 	curcs = sb;							      \
 	++inptr;							      \
 	continue;							      \
diff --git a/iconvdata/ibm933.c b/iconvdata/ibm933.c
index c984cd1..669e357 100644
--- a/iconvdata/ibm933.c
+++ b/iconvdata/ibm933.c
@@ -108,24 +108,14 @@  enum
 									      \
     if (__builtin_expect (ch, 0) == SO)					      \
       {									      \
-	/* Shift OUT, change to DBCS converter.  */			      \
-	if (curcs == db)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \
 	curcs = db;							      \
 	++inptr;							      \
 	continue;							      \
       }									      \
     else if (__builtin_expect (ch, 0) == SI)				      \
       {									      \
-	/* Shift IN, change to SBCS converter.  */			      \
-	if (curcs == sb)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \
 	curcs = sb;							      \
 	++inptr;							      \
 	continue;							      \
diff --git a/iconvdata/ibm935.c b/iconvdata/ibm935.c
index 07f4bf2..2d3065e 100644
--- a/iconvdata/ibm935.c
+++ b/iconvdata/ibm935.c
@@ -109,24 +109,14 @@  enum
 									      \
     if (__builtin_expect(ch, 0) == SO)					      \
       {									      \
-	/* Shift OUT, change to DBCS converter.  */			      \
-	if (curcs == db)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \
 	curcs = db;							      \
 	++inptr;							      \
 	continue;							      \
       }									      \
     else if (__builtin_expect (ch, 0) == SI)				      \
       {									      \
-	/* Shift IN, change to SBCS converter.  */			      \
-	if (curcs == sb)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \
 	curcs = sb;							      \
 	++inptr;							      \
 	continue;							      \
diff --git a/iconvdata/ibm937.c b/iconvdata/ibm937.c
index 88344a4..3ed6479 100644
--- a/iconvdata/ibm937.c
+++ b/iconvdata/ibm937.c
@@ -109,24 +109,14 @@  enum
 									      \
     if (__builtin_expect (ch, 0) == SO)					      \
       {									      \
-	/* Shift OUT, change to DBCS converter.  */			      \
-	if (curcs == db)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \
 	curcs = db;							      \
 	++inptr;							      \
 	continue;							      \
       }									      \
     else if (__builtin_expect (ch, 0) == SI)				      \
       {									      \
-	/* Shift IN, change to SBCS converter.  */			      \
-	if (curcs == sb)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \
 	curcs = sb;							      \
 	++inptr;							      \
 	continue;							      \
diff --git a/iconvdata/ibm939.c b/iconvdata/ibm939.c
index dbdda62..c0a75f7 100644
--- a/iconvdata/ibm939.c
+++ b/iconvdata/ibm939.c
@@ -109,24 +109,14 @@  enum
 									      \
     if (__builtin_expect (ch, 0) == SO)					      \
       {									      \
-	/* Shift OUT, change to DBCS converter.  */			      \
-	if (curcs == db)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \
 	curcs = db;							      \
 	++inptr;							      \
 	continue;							      \
       }									      \
     else if (__builtin_expect (ch, 0) == SI)				      \
       {									      \
-	/* Shift IN, change to SBCS converter.  */			      \
-	if (curcs == sb)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \
 	curcs = sb;							      \
 	++inptr;							      \
 	continue;							      \