Craig Silverstein wrote:
> } This is the output of pcre_unittest:
> }
> } pcrecpp_unittest.cc:865: Check failed:
> } RE("(.*)").FullMatch("1234567890123456", (long long*)NULL)
>
> My guess is that HAVE_LONG_LONG is not defined in your config.h --
> that will cause this match to always fail. Can you post the contents
> of your config.h? config.log would be helpful too, if it's not too
> big.
>
> It looks like this is a test that should be protected by '#ifdef
> HAVE_LONG_LONG' but isn't In theory this code shouldn't even compile,
> which leads me to believe it's a configuration problem where
> HAVE_LONG_LONG isn't set for you even though the compiler supports
> "long long" just fine. But it's definitely a bugfix on my part, and
> maybe a bugfix in the configure step for msys as well. I'll have a
> patch shortly.
>
> craig
>
>
>
Hi Craig,
With your patch, all is well. In fact I think it has the most verbose
output I've ever seen from the unittest, a list of test names followed
by "OK".
I have been configuring with cmake, which has configuration files
maintained separately. We may need Daniel, although Philip managed to
make a few recent changes ok.
Anyhoo, here is the config.h that ended up in my build directory. I
don't see a config.log.
Regards,
Sheri
/* config.h for CMake builds */
#define HAVE_DIRENT_H
#define HAVE_UNISTD_H
#define HAVE_SYS_STAT_H
#define HAVE_SYS_TYPES_H
/* #undef HAVE_TYPE_TRAITS_H */
#define HAVE_BITS_TYPE_TRAITS_H
/* #undef HAVE_BCOPY */
#define HAVE_MEMMOVE
#define HAVE_STRERROR
/* #undef PCRE_STATIC */
#define SUPPORT_UTF8
#define SUPPORT_UCP
/* #undef EBCDIC */
#define BSR_ANYCRLF
#define NO_RECURSE
/* #undef SUPPORT_LIBBZ2 */
/* #undef SUPPORT_LIBZ */
/* #undef SUPPORT_LIBREADLINE */
#define NEWLINE -2
#define POSIX_MALLOC_THRESHOLD 10
#define LINK_SIZE 2
#define MATCH_LIMIT 10000000
#define MATCH_LIMIT_RECURSION MATCH_LIMIT
#define MAX_NAME_SIZE 32
#define MAX_NAME_COUNT 10000
/* end config.h for CMake builds */
