This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
[PATCH] Avoid redundant shift character in iconv output at block boundary (bug #17197)
- From: Andreas Schwab <schwab at suse dot de>
- To: libc-alpha at sourceware dot org
- Date: Thu, 24 Jul 2014 18:11:41 +0200
- Subject: [PATCH] Avoid redundant shift character in iconv output at block boundary (bug #17197)
- Authentication-results: sourceware.org; auth=none
The IBM-93[03579] converters fail to update the saved shift state if
they output a shift in character just at the end of the output buffer,
so that the next character doesn't fit any more. This causes the next
call to iconv to output a second shift character.
Andreas.
[BZ #17197]
* iconvdata/ibm930.c (BODY for TO_LOOP): Record current DBCS state
immediately after emitting SI.
* iconvdata/ibm933.c (BODY for TO_LOOP): Likewise.
* iconvdata/ibm935.c (BODY for TO_LOOP): Likewise.
* iconvdata/ibm937.c (BODY for TO_LOOP): Likewise.
* iconvdata/ibm939.c (BODY for TO_LOOP): Likewise.
* iconvdata/bug-iconv10.c: New file.
* iconvdata/Makefile (tests): Add bug-iconv10.
($(objpfx)bug-iconv10.out): New rule.
---
iconvdata/Makefile | 5 ++++-
iconvdata/bug-iconv10.c | 60 +++++++++++++++++++++++++++++++++++++++++++++++++
iconvdata/ibm930.c | 2 +-
iconvdata/ibm933.c | 2 +-
iconvdata/ibm935.c | 2 +-
iconvdata/ibm937.c | 2 +-
iconvdata/ibm939.c | 2 +-
7 files changed, 69 insertions(+), 6 deletions(-)
create mode 100644 iconvdata/bug-iconv10.c
diff --git a/iconvdata/Makefile b/iconvdata/Makefile
index 0a410a1..7b2e525 100644
--- a/iconvdata/Makefile
+++ b/iconvdata/Makefile
@@ -67,7 +67,8 @@ 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
+ tst-iconv6 bug-iconv5 bug-iconv6 tst-iconv7 bug-iconv8 bug-iconv9 \
+ bug-iconv10
ifeq ($(have-thread-library),yes)
tests += bug-iconv3
endif
@@ -293,6 +294,8 @@ $(objpfx)tst-iconv4.out: $(objpfx)gconv-modules \
$(addprefix $(objpfx),$(modules.so))
$(objpfx)tst-iconv7.out: $(objpfx)gconv-modules \
$(addprefix $(objpfx),$(modules.so))
+$(objpfx)bug-iconv10.out: $(objpfx)gconv-modules \
+ $(addprefix $(objpfx),$(modules.so))
$(objpfx)iconv-test.out: run-iconv-test.sh $(objpfx)gconv-modules \
$(addprefix $(objpfx),$(modules.so)) \
diff --git a/iconvdata/bug-iconv10.c b/iconvdata/bug-iconv10.c
new file mode 100644
index 0000000..904c510
--- /dev/null
+++ b/iconvdata/bug-iconv10.c
@@ -0,0 +1,60 @@
+/* bug 17197: check for redundant shift character at block boundary. */
+#include <iconv.h>
+#include <locale.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <errno.h>
+
+static int
+do_test (void)
+{
+ iconv_t cd = iconv_open ("IBM930", "UTF-8");
+ if (cd == (iconv_t) -1)
+ {
+ puts ("iconv_open failed");
+ return 1;
+ }
+
+ char instr1[] = "\xc2\xa6.";
+ const char expstr1[4] = "\016Bj\017";
+ const char expstr2[] = "K";
+ char outstr[4];
+ size_t inlen = sizeof (instr1);
+ size_t outlen = sizeof (outstr);
+ char *inptr = instr1;
+ char *outptr = outstr;
+ size_t r = iconv (cd, &inptr, &inlen, &outptr, &outlen);
+ if (r != -1
+ || errno != E2BIG
+ || inlen != sizeof (instr1) - 2
+ || inptr != instr1 + 2
+ || outlen != 0
+ || memcmp (outstr, expstr1, sizeof (expstr1)) != 0)
+ {
+ puts ("wrong first conversion");
+ return 1;
+ }
+
+ outlen = sizeof (outstr);
+ outptr = outstr;
+ r = iconv (cd, &inptr, &inlen, &outptr, &outlen);
+ if (r != 0
+ || inlen != 0
+ || outlen != sizeof (outstr) - sizeof (expstr2)
+ || memcmp (outstr, expstr2, sizeof (expstr2)) != 0)
+ {
+ puts ("wrong second conversion");
+ return 1;
+ }
+
+ if (iconv_close (cd) != 0)
+ {
+ puts ("iconv_close failed");
+ return 1;
+ }
+ return 0;
+}
+
+#define TEST_FUNCTION do_test ()
+#include "../test-skeleton.c"
diff --git a/iconvdata/ibm930.c b/iconvdata/ibm930.c
index 768a444..9d4b011 100644
--- a/iconvdata/ibm930.c
+++ b/iconvdata/ibm930.c
@@ -256,6 +256,7 @@ enum
break; \
} \
*outptr++ = SI; \
+ curcs = sb; \
} \
\
if (__glibc_unlikely (outptr + 1 > outend)) \
@@ -269,7 +270,6 @@ enum
*outptr++ = 0x5b; \
else \
*outptr++ = cp[0]; \
- curcs = sb; \
} \
\
/* Now that we wrote the output increment the input pointer. */ \
diff --git a/iconvdata/ibm933.c b/iconvdata/ibm933.c
index f46dfb5..d086bc9 100644
--- a/iconvdata/ibm933.c
+++ b/iconvdata/ibm933.c
@@ -255,6 +255,7 @@ enum
break; \
} \
*outptr++ = SI; \
+ curcs = sb; \
} \
\
if (__glibc_unlikely (outptr + 1 > outend)) \
@@ -263,7 +264,6 @@ enum
break; \
} \
*outptr++ = cp[0]; \
- curcs = sb; \
} \
\
/* Now that we wrote the output increment the input pointer. */ \
diff --git a/iconvdata/ibm935.c b/iconvdata/ibm935.c
index a8e4e6c..1af3ff7 100644
--- a/iconvdata/ibm935.c
+++ b/iconvdata/ibm935.c
@@ -255,6 +255,7 @@ enum
break; \
} \
*outptr++ = SI; \
+ curcs = sb; \
} \
\
if (__glibc_unlikely (outptr + 1 > outend)) \
@@ -263,7 +264,6 @@ enum
break; \
} \
*outptr++ = cp[0]; \
- curcs = sb; \
} \
\
/* Now that we wrote the output increment the input pointer. */ \
diff --git a/iconvdata/ibm937.c b/iconvdata/ibm937.c
index 239be61..ecf8f54 100644
--- a/iconvdata/ibm937.c
+++ b/iconvdata/ibm937.c
@@ -255,6 +255,7 @@ enum
break; \
} \
*outptr++ = SI; \
+ curcs = sb; \
} \
\
if (__glibc_unlikely (outptr + 1 > outend)) \
@@ -263,7 +264,6 @@ enum
break; \
} \
*outptr++ = cp[0]; \
- curcs = sb; \
} \
\
/* Now that we wrote the output increment the input pointer. */ \
diff --git a/iconvdata/ibm939.c b/iconvdata/ibm939.c
index 5d0db36..ae443aa 100644
--- a/iconvdata/ibm939.c
+++ b/iconvdata/ibm939.c
@@ -255,6 +255,7 @@ enum
break; \
} \
*outptr++ = SI; \
+ curcs = sb; \
} \
\
if (__glibc_unlikely (outptr + 1 > outend)) \
@@ -268,7 +269,6 @@ enum
*outptr++ = 0xb2; \
else \
*outptr++ = cp[0]; \
- curcs = sb; \
} \
\
/* Now that we wrote the output increment the input pointer. */ \
--
2.0.2
--
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."