(forgot to cc the list)
On 9/18/2011 11:29 AM, Herczeg Zoltán wrote:
> Hi Sheri,
>
> unfortunately pcre_jit_test just contains lots of test cases without any real organization, since it was never planned to be an official test program, it just happened that way. Refactoring it is on my TODO list, and its prioirty is considerably increased as you want to use it. The time consuming part is checking whether UTF8 usage is required or optional for each test case.
>
> The \B test has PCRE_MULTILINE | PCRE_NEWLINE_ANYCRLF | PCRE_UCP flags, and obviously cannot be compiled without UCP support (as you probably know you can use UCP without UTF8, although its scope is limited to the first 256 character).
>
> Regards,
> Zoltan
OK, I had wondered if my ANYCRLF newline default could be causing
exceptions in various assertions in the absence of ucp support.
Regards,
Sheri
>
> Sheri<silvermoonwoman@???> írta:
>> On 9/18/2011 10:31 AM, Philip Hazel wrote:>
>> On Sun, 18 Sep 2011, Sheri wrote:>
>>> I think that could be easily fixed in the CMakeLists.txt file, but should the>
>>> building of the pcre_jit_test program be excluded in the absense of utf8>
>>> support?>
>> Yes, I'm afraid it should. The pcre_jit_test program currently expects>
>> UTF-8 support to be available. Zoltan knows about this, but at the>
>> moment wants to concentrate on fixing bugs.>
>> Philip>
> With both utf8 and ucp support, pcre_jit_test has no exceptions.>
> But when built without ucp support, although it reports 89% success>
> there appear to be some unexpected results from pcre_jit_test. For>
> example (among others):>
> Cannot compile pattern: \B>
> Cannot compile pattern: \b>
> Cannot compile pattern: [^b-d^&\s]{3,}>
> Regards,>
> Sheri>
> -- >
> ## List details at https://lists.exim.org/mailman/listinfo/pcre-dev>
>
>
>
