I would like to make the connection between our html tags and the js objects
considerably more efficient,
so when duktape frees an object we quickly manage the corresponding tag.
Right now there are linear searches through every frame in every window on
edbrowse, and other such ugliness.
This is part of the reason edbrowse is slow.
A better connection almost demands one process.
Almost, I know how to do it with 2, but it's a lot more work, along a path that
I'm pretty sure is a dead end, and I really don't want to bother with that.
So I propose a new branch, proc1, wherein I am committed to one process, and
can realize all the efficiency gains therein, and remove the messaging code for
and passing curl space information back and forth and all that.
You might want to follow this branch to help me test and debug.
If and when we're all comfortable with it we can merge back to master, just as
we did with duktape.
And like duktape, it will be a decision that is very hard to undo, so we
definitely want to feel good about it.
We want to feel like js hardly ever seg faults, for instance.
I already feel pretty good about interrupting js with ^c in a graceful manner;
that seems to work.
So keep an eye out for branch proc1, and follow along.
Edbrowse-dev mailing list