Hi, Strk, [EMAIL PROTECTED] wrote:
>> In order to automatically capture diversions between capi/geos_c.cpp and
>> capi/geos_c.h, I applied the attached patch to geos.
>>
>> Apart from being an ugly hack that I don't want to see applied
>> permanently, it tells us that most remaining differences seem to be
>> signed vs. unsigned char, and const declaration.
>>
>> Should those be fixed? And can we come up with a better scheme to
>> automatically catch diversions of those files?
>
> I'd rather see this applied permanently, unless you see any
> drawback (I assume you also get rid of all prototypes in the
> .cpp file).
Yes, getting rid of all those prototypes would be worth the effort, I think.
But we will need get more intrusive this way, including some
preprocessor magic in geos_c.h, to avoid
geos_c.h:72: error: expected unqualified-id before '{' token
This is the first error that's spit out. This and the '#define's are the
main cause why I regard that patch as an ugly hack. :-)
> I think we can fix const vs. unconst and signed vs. unsigned,
> witout fear of breaks. IIRC the problem with signed vs. unsigned
> was with function returning -1 on exception. Those would force
> callers to cast to signed for checking...
We should fix geos_c.cpp to provide whatever geos_c.h tells, to keep
compatibility with clients, possibly using (ugly) casts in geos_c.cpp.
C++ is more picky about those issues than C, I'm afraid.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
signature.asc
Description: OpenPGP digital signature
_______________________________________________ geos-devel mailing list geos-devel@geos.refractions.net http://geos.refractions.net/mailman/listinfo/geos-devel
