Bruno Haible <bruno <at> clisp.org> writes:
> Indeed. I was seeing angle brackets instead of parentheses... Together with
> the description of strtod(), it's clear, however. Here is a fix.
Some nits:
> --- m4/printf.m4 22 May 2007 01:02:09 -0000
> --- 138,163 ----
> AC_TRY_RUN([
> #include <stdio.h>
> #include <string.h>
> + static int
> + strisnan (const char *string, size_t start_index, size_t end_index)
> + {
> + if (start_index < end_index)
I know you just copy-n-pasted strisnan across all the files where you added it;
but in the .m4 file, where you only call strisnan with start_index of 0, it
might be simpler to get rid of the check for start_index<end_index since it is
always true, or even remove the parameter start_index.
> --- tests/test-snprintf-posix.h 22 May 2007 01:02:10 -0000
> --- 396,402 ----
> /* "0000000nan 33" is not a valid result; see
> <http://lists.gnu.org/archive/html/bug-gnulib/2007-04/msg00107.html>
*/
> ASSERT (strlen (result) == 20 + 3
> ! && strisnan (result, strspn (result, " "), strlen (result) - 3, 0)
> && strcmp (result + strlen (result) - 3, " 33") == 0);
This assertion will fail if the implementation produces an n-char-sequence
NaN. You may want to assert that strlen(result) == the return value of the
printf, or that strlen(result) > minimum length, but you cannot assert that the
result is a fixed length. This occurs several times throughout your patch.
--
Eric Blake
