Re: [pcre-dev] A test for the memory mismatch problem

Philip Hazel wrote:
> On Thu, 29 Mar 2007, Sheri wrote:
>
>
>> pcre.h:75:2: #error _WIN32 is defined
>>
>> pcre.h:75:2: #error WIN32 is defined
>>
>
> Rats! So that doesn't explain it.
>
> Looking at that code again (I don't really understand it), I see that it
> also tests PCRE_DEFINITION and DLL_EXPORT and PCRE_STATIC. Please can
> you run same test again (3 times), in each case, substituting one of
> those names for _WIN32 ?
>
> I'm grasping at straws here - anybody else with Windows programming
> experience on this list care to comment? It appears that the global
> variable pcre_malloc, which is the address of a function, isn't being
> properly shared between two modules.
>
>
> Philip
>
>
pcre.h:75:2: #error PCRE_DEFINITION is defined
pcre.h:75:2: #error DLL_EXPORT is defined
pcre.h:77:2: #error PCRE_STATIC is not defined

There were some notes in 7.0's NON-UNIX USE about the necessity of
defining static if linking to the ".a" file. Wondering if the fact that
the posix library is now separate and interlinked to the pcre library
(and I'm building as shared) has anything to do with anything. Though I
realize the test itself isn't using the posix interface.

Is there anything I could do building 7.0 the old way, and changing
7.0's pcretest.c or pcre.h to help you (for comparison purposes)?

Regards,
Sheri