On Monday, November 12, 2012 19:04:25 Stuart Dickson wrote: > A single library for managing EMF, SVM and WMF files. > > I have submitted commits to the branch: libs-vectorimage-stuartmd-forreview > Only the last three are relevant to the libs/vectorimage work - the diff > may be generated with the call: > git diff c55ec6499c5ead6d6a9b434651ab31f92d8b0b8b
Hmm, this generated a completely unrelated diff that adds KDGANTT_EXPORT to the exported classes in kdgantt/. Did you or I do something wrong? As a side note, notice that kdgantt/ is an imported library from KDAB just like plugins/vectorshape/kdchart/*. If you patch it, your patches will be lost when kdgantt is updated with a new version unless you save the patches and reapply them to the new version. We do this in the chart shape; see plugins/chartshape/kdchartpatches/*. Perhaps you can do something similar for kdgantt. (As a side side note, I'm not sure why kdgantt/ is under / in Calligra instead of under /libs/). > > In reviewing master branch, it was determined that the problem building > the epub export filter lay in the way it was trying to use the vector > image libraries already present in plugins/vectorshape and > filters/kolibwmf. (This problem may have only really manifested on > Windows, owing to the way it handles shared libraries). As far as I know there is no vector image library in plugins/vectorshape. But there are a number of classes that implement support for these formats. However they are not used to create a library but are recompiled everywhere they are used. This is ugly but apparently it never causes any problem on Linux. It is good that the classes are moved to a real library. > There had already been a suggestion that, as there was already some > awkward coupling of plugins, and filters and knowledge of the location > of the sources required, that moving the library code for EMF, SVM and > WMF vector file formats into their own library would be the best way > forward. Yeah, indeed. > This patch creates the library and makes the necessary changes to its > dependents and has been tested to function on the Windows platform. I couldn't see that (see above), but I am interested in reviewing the patch. > > Regards > Stuart > > > > -- Stuart Dickson > // KO GmbH http://kogmbh.com/legal/ > _______________________________________________ > calligra-devel mailing list > calligra-devel@kde.org > https://mail.kde.org/mailman/listinfo/calligra-devel _______________________________________________ calligra-devel mailing list calligra-devel@kde.org https://mail.kde.org/mailman/listinfo/calligra-devel
