On Thu, Feb 20, 2014 at 04:25:29PM -0500, Karl Dahlke wrote:
> Well I just needed to be a bit more patient.
> It was actually creating those 160,000 objects over again, very slowly.
> Each time it had just enough memory to make the new object,
> then knowing it was out of space it ran gc, and deleted the old one,
> and slowly step through and replaced all those objects.
> Very slow but understandable.

Yeah.

> That's most of the mystery solved.
> But why does a printf, outside of javascript,
> tweak the memory pool inside javascript just enough
> so it can't create that first object, whence it stops right away?
> I can't imagine the connection.
> They should be independent of each other.

Yeah, particularly as the js memory pool is on the heap.

> 
> I added a free memory button to jsrt,
> so we don't have to quit and restart to get js working again.

Good idea.

> 
> The noscript tags hide everything inside,
> but they could let this pass through if isJSAlive fails.
> This would only have meaning when the page was first parsed.
> So if you turned js off, or it turned itself off because
> of an error in another session, then yes you would see
> the text in <noscript>,
> but it wouldn't jump out at you if js stopped in this session
> for any reason.
> If you still think it's a good feature
> I could implement it without too much fuss.

Yeah, I didn't expect the noscript stuff to appear when js failed to execute,
after all that could leave the page in an unknown (possibly completely messed 
up) state.
However I think displaying the contents of the noscript tags if js is already
dead makes sense, since if we have a js failure,
it'd then be possible to refresh the page to get the noscript contents.

Cheers,
Adam.
_______________________________________________
Edbrowse-dev mailing list
[email protected]
http://lists.the-brannons.com/mailman/listinfo/edbrowse-dev

Reply via email to