libqtmimetypes ?
Stephen Kelly
steveire at gmail.com
Sun May 13 20:46:10 UTC 2012
Alexander Neundorf wrote:
> On Sunday 13 May 2012, Stephen Kelly wrote:
>> Alexander Neundorf wrote:
> ...
>> > This means either we need to put a lot of work in the buildsystem, to
>> > handle all cases correctly (and clean up afterwards again), or we don't
>> > support building standalone for anything tier>1 at all, also with Qt5,
>> > until either the split has happened or cmake got the new feature from
>> > Yury.
>>
>> Yes, this is what I would suggest is smartest at the moment.
>
> I thought a bit more about it in the meantime.
> We can make it work by ignoring the exported targets (and all their nice
> properties) for the moment and simply do something like the following in
> the Config files:
>
> find_library(_ at PROJECT_NAME@_LIBRARY NAMES @PROJECT_NAME@
> PATHS ${@PROJECT_NAME at _LIBRARY_DIR}
> NO_DEFAULT_PATH)
>
> add_library(@PROJECT_NAME@ IMPORTED)
>
> set_target_properties(@PROJECT_NAME@ PROPERTIES LOCATION ...)
>
> set(@PROJECT_NAME at _LIBRARY ${_ at PROJECT_NAME@_LIBRARY})
>
> It's a pity that it will still not be like it should be, i.e. we still
> cannot point people to it for a good example (I'd like to get to this
> state as soon as possible).
> But it will make it work.
> Not sure I'll find the time to do it the coming week.
Why though? Just to get tier2 libraries building standalone? What you
propose isn't good enough that it gets the tier2 libraries into a final
state as you say, so why bother? It also looks like it might introduce
subtle bugs and be complex to maintain.
For example, why _LIBRARY instead of _LIBRARIES? Someone looking at it will
think they should use _LIBRARY for their own stuff. If it's _LIBRARY because
it is a temporary thing, then the _ at PROJECT_NAME@_LIBRARY should be used by
the caller (with the leading underscore). Also, if this is a tier2 config
file, then where does ${@PROJECT_NAME at _LIBRARY_DIR} come from?
I think it makes more sense to wait until the features are available to make
it possible to get to our final state (Yury's work). Otherwise we'll only
get to some transitional state that can't be used for anything, or am I
wrong?
Thanks,
Steve.
More information about the Kde-frameworks-devel
mailing list