gnash-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gnash-dev] Re: unreachable objects at startup


From: strk
Subject: [Gnash-dev] Re: unreachable objects at startup
Date: Sun, 13 Jun 2010 14:56:40 +0200

For the record, the other 2 objects are:

 1. "MovieClip" class declaration (not sure it makes sense
    to use a placeholder for it, since it'll be always needed
    for the root movie).

 2. "enableDebugConsole" global special function, for which we
    clear all the properties which are added by the builtin_function
    constructor. This is unimplemented, btw. Fixing this would require
    adding another constructor to both as_function and builtin_function.

Not a big deal to keep these, but would be nice to run with no
superflous allocations, if not else so next time someone looks at
GC stats doesn't wonder again :)

--strk;

On Sun, Jun 13, 2010 at 01:41:20PM +0200, strk wrote:
> initObjectClass() calls attachObjectInterface() twice.
> This results in 10 objects being replaced during startup,
> which are 10 objects found unreachable at first GC run.
> 
> There's a total of 12 objects unreachable at first GC run,
> which I'm trying to figure out.
> 
> One thing: I did expect these kind of replacements
> to throw an abort or similar, which was one of the reasons
> for using init_member rather than set_member in initialization 
> code. This didn't happen, so would be worth inspecting why
> as that might help figuring out the left-over 2.
> 
> Note that when building with fileio extension enabled the
> number of unreachable objects at first GC run is  higher
> so it does make sense to have a way to catch such programmatic
> errors.
> 
> I'm on a very slow connection so feel free to commit a fix for
> the double attachObjectInterface. Thanks.
> 
> --strk; 
> 
>   ()   Free GIS & Flash consultant/developer
>   /\   http://strk.keybit.net/services.html

-- 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html



reply via email to

[Prev in Thread] Current Thread [Next in Thread]