Hello Dominik, On Mon, Dec 02, 2013 at 12:53:32PM +0000, Rottsches, Dominik wrote: > I just came across this post by Adam on implementing a new API as - what he > calls - a "Blink Module": > > https://groups.google.com/a/chromium.org/d/msg/blink-dev/R5yYtHEPiw4/-eVaKo6unq4J
Thanks for the pointer. Sorry for the late response. > Did anyone else see this? Does this approach affect how we implement > extensions in the future? Could it be useful for us the use the > "Blink module" approach - without having looked into it in detail > myself. I don't expect this to change things a lot for us in the short term. Currently in Crosswalk, opting to make a change to Blink (instead of doing it without change Blink), have the following tradeoff Positive + Inside the web engine -- this is necessary to implement certain functionality. + Easier to upstream the change, use same infrastructure as upstream. Negative - In many cases we also need to expose the feature in the Blink API and then through Content API, in our case this means patching Chromium as well. - Harder maintenance of Chromium/Blink branches. The apporach Adam Barth proposed doesn't save us from the negative points. In a sense, "making Blink changes" was always on the table anyway (because some changes need to be there). In practice this doesn't change. Chromium is working on a system (mojo) that can be used to solve the first negative point. It will make communication between Blink side and BP side not necessarily having to go through Content API. As for the branches: I think we can do a lot more to ease the impact changes to them cause to us. I have some thoughts on this, I know Alexis has as well, but nothing concrete yet. Cheers, Caio _______________________________________________ Crosswalk-dev mailing list Crosswalk-dev@lists.crosswalk-project.org https://lists.crosswalk-project.org/mailman/listinfo/crosswalk-dev
