On Tue, Oct 04, 2016 at 11:26:05AM +0200, Bjoern Michaelsen wrote:
> The main trouble with all these IDE integrations seems that the output of 
> "make
> -np" on which it depends is not as stable between versions of make as
> originally expected by that naive guy who did the initial implementation. Thus
> the parser of that is both more fragile and more complex as it should be by 
> now.
> I have a half-finished patch that will make make write out how it intends to
> build stuff in small (json) text files. Parsing _that_ should be really 
> trivial
> and not fragile. Also it might make possible creating:
> 1/ the possiblity of creating a IDE integration without a full build
> 2/ the possiblity to create MS Build solutions for some (non-special) targets,
>    thus building like MSVS would.

Half of this, the "export from gbuild" part, is done now:


It should be rather trivial to read that from gbuildtoide like the Pythontest
does. If anyone wants to other fun stuff with it, feel free. Now that the base
is there, extending this should be trvial.


LibreOffice mailing list

Reply via email to