Thanks for your outline, @cocobean . For the record, my comparisons with WebKit_GTK would only be to try to find equivalent functionalities between the Linux and Haiku versions of webkit because internet searching for results of Linux problems is easier than finding answers to Haiku problems online. If I have to ask @PulkoMandy every time I run into trouble, I’ll have added to his work instead of taking away from it. He’s got a full plate already.
If i may suggest something, a good place maybe to start looking for usable API’s is the cairo lib, i would try to find out what it has to offer in terms of rendering API’s and take as much as possible advantage of what it has to offer, and then build webkit2 around that (like many other OS’s do), then one could add the haiku specific parts and see what needs to be patched or what changes need to be done to the appserver to make the native look of haiku. it might turn out to be of a less maintenance burden.
Webkit2 can already use cairo.
We can use cairo if we want, but this won’t give us native looking rendering at all.
Maybe I can look into webkit2 early next year : )
Yes that’s why i said we can later try to patch cairo to get the native look, now this is in theory, so it’s an idea one could explore, because retrying with the same the approach again might have the same result where there is alot of things to cover and not enough manpower/time.
If we use cairo we won’t get native rendering. trying to patch ot to immitate native rendering would be more work than using the code we already have for native drawing in webkit. I don’t think a hybrid aproach makes sense : )
I should remember that the graphics rendering in WebKit2 is already working since 2019 using the native APIs. It is the same code as in WebKitLegacy, just used a bit differently, and it solves the pretty bad flickering issues we have in WebKitLegacy.
So I’m not sure we should abandon that now that all the work is done, and switch to Cairo? What would we gain?
5 Likes
It was just a theory i thought that is worth looking at, it’s just that i keep reading complaints about several rendering issues, so i thought it’s possible to try another approach.
What is the purpose of these messages?
1 Like
How is posting versions that we don’t use information about bleeding edge development? The numbers are really un-interesting. Please refrain from such context free posts.
