Re: [pcre-dev] Yet Another RC

On Mon, 17 Sep 2007, Sheri wrote:

> When building RC3 using CMake as configurer for Visual Studio Express, I
> get these warnings and errors. Note that pcre_scanner_unittest doesn't
> currently build. Are the warnings on the other components nothing to
> worry about?

Hopefully Craig will comment on the data warnings in the C++ sources.

> pcre_scanner_unittest.cc
> ..\pcre-7.4-RC3\pcre_scanner_unittest.cc(136) : error C3861: 'snprintf':
> identifier not found

Aarrgghh!! This module currently has

#ifdef HAVE_WINDOWS_H
# define snprintf _snprintf
#endif

Previously it tested _WIN32 instead of HAVE_WINDOWS_H, but someone (was
it you?) had a problem with that in one or other mode of building, so I
changed it (as per discussion on this list). [This reminds me of my PhD
theses, where one equation had > changed to >= and back so many times,
at the whim of supervisors and examiners, that the buildup of Snopaque
was getting serious...]

I'm afraid I can only do what I'm advised to do on this one... Help!

> pcretest.c
> ..\pcre-7.4-RC3\pcretest.c(1358) : warning C4244: '=' : conversion from
> 'unsigned long' to 'pcre_uint16', possible loss of data

...etc ...

Those could do with the insertion of some casts.

> 'unsigned long' to 'pcre_uint16', possible loss of data
> ..\pcre-7.4-RC3\pcretest.c(1599) : warning C4244: '=' : conversion from
> 'unsigned long' to 'uschar', possible loss of data

... etc ...

Those might as well get some casts too.

Neither of the pcretest groups of warnings is anything to worry about.

Philip

--
Philip Hazel, University of Cambridge Computing Service.