[Edbrowse-dev] root functions

[Edbrowse-dev] root functions

[Karl Dahlke]

> as JS_Add*Root functions are present in smjs 185 as well.

I'm sure they weren't there from the get-go.
I would have noticed them.
They likely came in as of 185, which was itself a substantial change
from what came before.
Support for moz185 came, by the change log, on version 3.4.8.
Now on 3.4.10.
So we haven't been living dangerously for a long time,
but still too long.
Time to figure out the new paradigm.

Karl Dahlke

Re: [Edbrowse-dev] root functions

[Chris Brannon]

Karl Dahlke <eklhad@comcast.net> writes:

> So we haven't been living dangerously for a long time,

Well, I just grabbed the 1.5 sources, and guess what?  JS_AddRoot and
JS_AddNamedRoot have been there since 2004, at least.
But don't worry, I'm *pretty sure* we're ok.  The global object (jwin)
should be rooted.
I'm pretty sure that the JS objects we create and use are reachable from
the global object, aren't they?  If so, then they won't be collected.
But I still can't swear to the fact that we don't have GC bugs.

-- Chris

Re: [Edbrowse-dev] root functions

[Adam Thompson]

On Sun, Jan 05, 2014 at 03:10:48PM -0500, Karl Dahlke wrote:
> > as JS_Add*Root functions are present in smjs 185 as well.
> 
> I'm sure they weren't there from the get-go.
> I would have noticed them.
> They likely came in as of 185, which was itself a substantial change
> from what came before.
> Support for moz185 came, by the change log, on version 3.4.8.
> Now on 3.4.10.
> So we haven't been living dangerously for a long time,
> but still too long.
> Time to figure out the new paradigm.

Yeah, though I'm glad to hear the bugs aren't as old as I'd assumed.
As it looks like we can still do things in a c-like way,
I think we may be able to escape massive rewriting at the moment.
I'm currently trying to work out what needs protecting from the GC and what should be ok.

Cheers,
Adam.