unresolved symbols from value.cpp/.h

Christian Ehrlicher Ch.Ehrlicher at gmx.de
Tue Jan 24 19:18:44 GMT 2006

Hash: SHA1

Maks Orlovich schrieb:
> On Tuesday 24 January 2006 13:01, Christian Ehrlicher wrote:
>>Christian Ehrlicher schrieb:
>>>Here two patches for kjs - splitted into errors and warnings.
>>>The warnigns are mostly irrelevant but I won't disable them by default.
>>Are those changes ok or don't you like to support msvc anymore?
> Sorry, not much time ATM. I think they're fine, but likely the latest sync 
> with the Apple tree will cover at least some portion of them. So feel free
> to commit, but some stuff may get lost at the latest sync, etc.
> Sorry, but we don't have a good solution for pushing things around ATM.
Hmm. How can I ensure that (at least some) of my changes go into apple's
> And whether I like to support msvc is irrelevant, what's relevant is whether 
> someone is willing to make stuff work with it w/o breaking things for the 
> rest of us :-)
My current problem is that all those small things prevent me from
compiling kdelibs so I can go further trying to execute an app on windows.
>>And what about the ugly hack with class ValueImp?
> There is no ugly hack there. Just a standard language idiom. These methods 
> should never be used. The only way to ensure that is to declare them and 
> mark them private, as otherwise public defaults will be generated. And as they 
> can't ever do anything sensible, there is no definition. (The only acceptable 
> one would be abort(), basically). The point is, ValueImp's are not copyable, 
> even though they're exported outside the library.
> And again, the only solution is a #define to disable those for the broken 
> compilers out there. 
Ok, I'll stop calling it ugly and make a define :)

Version: GnuPG v1.4.2 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


More information about the kde-core-devel mailing list