summary |
shortlog |
log |
commit | commitdiff |
tree
raw |
patch |
inline | side by side (from parent 1:
4104cd4)
getcwd(NULL, 1) mallocs a larger buffer on BSD, rather than failing
with ERANGE as on glibc. This behavior difference is not worth
coding around, as it is an uncommon use of getcwd in the first place.
* doc/posix-functions/getcwd.texi (getcwd): Document portability
issue.
* tests/test-getcwd-lgpl.c (main): Relax test.
Reported by Matthias Bolte.
Signed-off-by: Eric Blake <eblake@redhat.com>
+2011-05-13 Eric Blake <eblake@redhat.com>
+
+ getcwd-lgpl: relax test for FreeBSD
+ * doc/posix-functions/getcwd.texi (getcwd): Document portability
+ issue.
+ * tests/test-getcwd-lgpl.c (main): Relax test.
+ Reported by Matthias Bolte.
+
2011-05-11 Eric Blake <eblake@redhat.com>
test-fflush: silence compiler warning
2011-05-11 Eric Blake <eblake@redhat.com>
test-fflush: silence compiler warning
@itemize
@item
On glibc platforms, @code{getcwd (NULL, n)} allocates memory for the result.
@itemize
@item
On glibc platforms, @code{getcwd (NULL, n)} allocates memory for the result.
-On some other platforms, this call is not allowed. Conversely, mingw fails
-to honor non-zero @code{n}.
+On some other platforms, this call is not allowed.
@item
On some platforms, the prototype for @code{getcwd} uses @code{int}
instead of @code{size_t} for the size argument:
@item
On some platforms, the prototype for @code{getcwd} uses @code{int}
instead of @code{size_t} for the size argument:
Portability problems not fixed by Gnulib:
@itemize
Portability problems not fixed by Gnulib:
@itemize
+@item
+When using @code{getcwd(NULL, nonzero)}, some platforms, such as glibc
+or cygwin, allocate exactly @code{nonzero} bytes and fail with
+@code{ERANGE} if it was not big enough, while other platforms, such as
+FreeBSD or mingw, ignore the size argument and allocate whatever size
+is necessary. If this call succeeds, an application cannot portably
+access beyond the string length of the result.
pwd2 = malloc (len + 2);
for ( ; i <= len; i++)
{
pwd2 = malloc (len + 2);
for ( ; i <= len; i++)
{
errno = 0;
ASSERT (getcwd (pwd2, i) == NULL);
ASSERT (errno == ERANGE);
errno = 0;
ASSERT (getcwd (pwd2, i) == NULL);
ASSERT (errno == ERANGE);
+ /* Allow either glibc or BSD behavior, since POSIX allows both. */
- ASSERT (getcwd (NULL, i) == NULL);
- ASSERT (errno == ERANGE);
+ tmp = getcwd (NULL, i);
+ if (tmp)
+ {
+ ASSERT (strcmp (pwd1, tmp) == 0);
+ free (tmp);
+ }
+ else
+ {
+ ASSERT (errno == ERANGE);
+ }
}
ASSERT (getcwd (pwd2, len + 1) == pwd2);
pwd2[len] = '/';
}
ASSERT (getcwd (pwd2, len + 1) == pwd2);
pwd2[len] = '/';
ASSERT (strstr (pwd2, "/../") == NULL);
ASSERT (strstr (pwd2 + 1 + (pwd2[1] == '/'), "//") == NULL);
ASSERT (strstr (pwd2, "/../") == NULL);
ASSERT (strstr (pwd2 + 1 + (pwd2[1] == '/'), "//") == NULL);
+ /* Validate a POSIX requirement on size. */
+ errno = 0;
+ ASSERT (getcwd(pwd2, 0) == NULL);
+ ASSERT (errno == EINVAL);
+
free (pwd1);
free (pwd2);
free (pwd1);
free (pwd2);