[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
fruits of clang analysis
From: |
Bruno Haible |
Subject: |
fruits of clang analysis |
Date: |
Sat, 5 Sep 2009 16:55:08 +0200 |
User-agent: |
KMail/1.9.9 |
Jim recommended to use clang's static code analysis yesterday. The results
on my code: Mostly false positives, but also one real bug in gnulib code.
Bugs False harmless or
positives coding style
libiconv 0 0 0
libunistring 1 3 0
libunistring tests 0 7 0
gettext 0 51 6
git-merge-changelog 0 1 0
2009-09-05 Bruno Haible <address@hidden>
Fix conversion behaviour when the input is invalid.
* lib/striconveh.c (mem_cd_iconveh_internal): Fix storing of question
mark occurring in first pass of indirect conversion.
* tests/test-striconveh.c (main): Test conversion of invalid ASCII
input.
Found by clang's static analyzer.
*** lib/striconveh.c.orig 2009-09-05 16:48:30.000000000 +0200
--- lib/striconveh.c 2009-09-05 16:29:54.000000000 +0200
***************
*** 696,702 ****
in1ptr++;
in1size--;
}
! utf8buf[utf8len++] = '?';
}
errno1 = errno;
utf8len = out1ptr - utf8buf;
--- 696,703 ----
in1ptr++;
in1size--;
}
! *out1ptr++ = '?';
! res1 = 0;
}
errno1 = errno;
utf8len = out1ptr - utf8buf;
*** tests/test-striconveh.c.orig 2009-09-05 16:48:30.000000000 +0200
--- tests/test-striconveh.c 2009-09-05 16:48:22.000000000 +0200
***************
*** 66,89 ****
--- 66,103 ----
#if HAVE_ICONV
/* Assume that iconv() supports at least the encodings ASCII, ISO-8859-1,
ISO-8859-2, and UTF-8. */
+ iconv_t cd_ascii_to_88591 = iconv_open ("ISO-8859-1", "ASCII");
iconv_t cd_88591_to_88592 = iconv_open ("ISO-8859-2", "ISO-8859-1");
iconv_t cd_88592_to_88591 = iconv_open ("ISO-8859-1", "ISO-8859-2");
+ iconv_t cd_ascii_to_utf8 = iconv_open ("UTF-8", "ASCII");
iconv_t cd_88591_to_utf8 = iconv_open ("UTF-8", "ISO-8859-1");
iconv_t cd_utf8_to_88591 = iconv_open ("ISO-8859-1", "UTF-8");
iconv_t cd_88592_to_utf8 = iconv_open ("UTF-8", "ISO-8859-2");
iconv_t cd_utf8_to_88592 = iconv_open ("ISO-8859-2", "UTF-8");
iconv_t cd_utf7_to_utf8 = iconv_open ("UTF-8", "UTF-7");
+ iconveh_t cdeh_ascii_to_88591;
+ iconveh_t cdeh_ascii_to_88591_indirectly;
iconveh_t cdeh_88592_to_88591;
iconveh_t cdeh_88592_to_88591_indirectly;
+ iconveh_t cdeh_ascii_to_utf8;
iconveh_t cdeh_88591_to_utf8;
iconveh_t cdeh_utf8_to_88591;
iconveh_t cdeh_utf7_to_utf8;
+ ASSERT (cd_ascii_to_utf8 != (iconv_t)(-1));
ASSERT (cd_88591_to_utf8 != (iconv_t)(-1));
ASSERT (cd_utf8_to_88591 != (iconv_t)(-1));
ASSERT (cd_88592_to_utf8 != (iconv_t)(-1));
ASSERT (cd_utf8_to_88592 != (iconv_t)(-1));
+ cdeh_ascii_to_88591.cd = cd_ascii_to_88591;
+ cdeh_ascii_to_88591.cd1 = cd_ascii_to_utf8;
+ cdeh_ascii_to_88591.cd2 = cd_utf8_to_88591;
+
+ cdeh_ascii_to_88591_indirectly.cd = (iconv_t)(-1);
+ cdeh_ascii_to_88591_indirectly.cd1 = cd_ascii_to_utf8;
+ cdeh_ascii_to_88591_indirectly.cd2 = cd_utf8_to_88591;
+
cdeh_88592_to_88591.cd = cd_88592_to_88591;
cdeh_88592_to_88591.cd1 = cd_88592_to_utf8;
cdeh_88592_to_88591.cd2 = cd_utf8_to_88591;
***************
*** 92,97 ****
--- 106,115 ----
cdeh_88592_to_88591_indirectly.cd1 = cd_88592_to_utf8;
cdeh_88592_to_88591_indirectly.cd2 = cd_utf8_to_88591;
+ cdeh_ascii_to_utf8.cd = cd_ascii_to_utf8;
+ cdeh_ascii_to_utf8.cd1 = cd_ascii_to_utf8;
+ cdeh_ascii_to_utf8.cd2 = (iconv_t)(-1);
+
cdeh_88591_to_utf8.cd = cd_88591_to_utf8;
cdeh_88591_to_utf8.cd1 = cd_88591_to_utf8;
cdeh_88591_to_utf8.cd2 = (iconv_t)(-1);
***************
*** 141,146 ****
--- 159,213 ----
}
}
+ /* Test conversion from ASCII to ISO-8859-1 with invalid input (EILSEQ). */
+ for (indirect = 0; indirect <= 1; indirect++)
+ {
+ for (h = 0; h < SIZEOF (handlers); h++)
+ {
+ enum iconv_ilseq_handler handler = handlers[h];
+ static const char input[] = "Rafa\263 Maszkowski"; /* Rafa?
Maszkowski */
+ for (o = 0; o < 2; o++)
+ {
+ size_t *offsets = (o ? new_offsets (strlen (input)) : NULL);
+ char *result = NULL;
+ size_t length = 0;
+ int retval = mem_cd_iconveh (input, strlen (input),
+ (indirect
+ ? &cdeh_ascii_to_88591_indirectly
+ : &cdeh_ascii_to_88591),
+ handler,
+ offsets,
+ &result, &length);
+ switch (handler)
+ {
+ case iconveh_error:
+ ASSERT (retval == -1 && errno == EILSEQ);
+ ASSERT (result == NULL);
+ if (o)
+ free (offsets);
+ break;
+ case iconveh_question_mark:
+ case iconveh_escape_sequence:
+ {
+ static const char expected[] = "Rafa? Maszkowski";
+ ASSERT (retval == 0);
+ ASSERT (length == strlen (expected));
+ ASSERT (result != NULL && memcmp (result, expected, strlen
(expected)) == 0);
+ if (o)
+ {
+ for (i = 0; i < 16; i++)
+ ASSERT (offsets[i] == i);
+ ASSERT (offsets[16] == MAGIC);
+ free (offsets);
+ }
+ free (result);
+ }
+ break;
+ }
+ }
+ }
+ }
+
/* Test conversion from ISO-8859-2 to ISO-8859-1 with EILSEQ. */
for (indirect = 0; indirect <= 1; indirect++)
{
***************
*** 276,281 ****
--- 343,392 ----
}
}
+ /* Test conversion from ASCII to UTF-8 with invalid input (EILSEQ). */
+ for (h = 0; h < SIZEOF (handlers); h++)
+ {
+ enum iconv_ilseq_handler handler = handlers[h];
+ static const char input[] = "Rafa\263 Maszkowski"; /* Rafa? Maszkowski
*/
+ for (o = 0; o < 2; o++)
+ {
+ size_t *offsets = (o ? new_offsets (strlen (input)) : NULL);
+ char *result = NULL;
+ size_t length = 0;
+ int retval = mem_cd_iconveh (input, strlen (input),
+ &cdeh_ascii_to_utf8,
+ handler,
+ offsets,
+ &result, &length);
+ switch (handler)
+ {
+ case iconveh_error:
+ ASSERT (retval == -1 && errno == EILSEQ);
+ ASSERT (result == NULL);
+ if (o)
+ free (offsets);
+ break;
+ case iconveh_question_mark:
+ case iconveh_escape_sequence:
+ {
+ static const char expected[] = "Rafa? Maszkowski";
+ ASSERT (retval == 0);
+ ASSERT (length == strlen (expected));
+ ASSERT (result != NULL && memcmp (result, expected, strlen
(expected)) == 0);
+ if (o)
+ {
+ for (i = 0; i < 16; i++)
+ ASSERT (offsets[i] == i);
+ ASSERT (offsets[16] == MAGIC);
+ free (offsets);
+ }
+ free (result);
+ }
+ break;
+ }
+ }
+ }
+
/* Test conversion from UTF-8 to ISO-8859-1 with EILSEQ. */
for (h = 0; h < SIZEOF (handlers); h++)
{
***************
*** 468,473 ****
--- 579,614 ----
}
}
+ /* Test conversion from ASCII to ISO-8859-1 with invalid input (EILSEQ). */
+ for (indirect = 0; indirect <= 1; indirect++)
+ {
+ for (h = 0; h < SIZEOF (handlers); h++)
+ {
+ enum iconv_ilseq_handler handler = handlers[h];
+ static const char input[] = "Rafa\263 Maszkowski"; /* Rafa?
Maszkowski */
+ char *result = str_cd_iconveh (input,
+ (indirect
+ ? &cdeh_ascii_to_88591_indirectly
+ : &cdeh_ascii_to_88591),
+ handler);
+ switch (handler)
+ {
+ case iconveh_error:
+ ASSERT (result == NULL && errno == EILSEQ);
+ break;
+ case iconveh_question_mark:
+ case iconveh_escape_sequence:
+ {
+ static const char expected[] = "Rafa? Maszkowski";
+ ASSERT (result != NULL);
+ ASSERT (strcmp (result, expected) == 0);
+ free (result);
+ }
+ break;
+ }
+ }
+ }
+
/* Test conversion from ISO-8859-2 to ISO-8859-1 with EILSEQ. */
for (indirect = 0; indirect <= 1; indirect++)
{
***************
*** 533,538 ****
--- 674,704 ----
free (result);
}
+ /* Test conversion from ASCII to UTF-8 with invalid input (EILSEQ). */
+ for (h = 0; h < SIZEOF (handlers); h++)
+ {
+ enum iconv_ilseq_handler handler = handlers[h];
+ static const char input[] = "Rafa\263 Maszkowski"; /* Rafa? Maszkowski
*/
+ char *result = str_cd_iconveh (input,
+ &cdeh_ascii_to_utf8,
+ handler);
+ switch (handler)
+ {
+ case iconveh_error:
+ ASSERT (result == NULL && errno == EILSEQ);
+ break;
+ case iconveh_question_mark:
+ case iconveh_escape_sequence:
+ {
+ static const char expected[] = "Rafa? Maszkowski";
+ ASSERT (result != NULL);
+ ASSERT (strcmp (result, expected) == 0);
+ free (result);
+ }
+ break;
+ }
+ }
+
/* Test conversion from UTF-8 to ISO-8859-1 with EILSEQ. */
for (h = 0; h < SIZEOF (handlers); h++)
{
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- fruits of clang analysis,
Bruno Haible <=