libgcj/117: _Jv_PthreadCheckMonitor test is not portable.

Fri Dec 10 08:36:00 GMT 1999

Tom Tromey wrote:
> My understanding is that this is a thread system and does not use
> native threads. This is also not what we want. (Maybe I'm wrong.
> The URL you give is hardly an easy way to find out...).

Sorry, for that:
 http://www.mozilla.org/docs/refList/refNSPR/
If I understand correctly the situation is as follows: if a native 
thread library (like pthread) exists, a PRThread maps to a native 
thread. If there's no native thread, they emulate threads. 
> Anyway, having control over the thread layer is important to us, I
> think. We want to be able to write new thread ports without having to
> mess around (with negotiations or forking other people's packages).

NSPR has been ported to over 20 OS. I don't know if they plan to port
NSPR to embedded systems. FWIW: Doesn't your argumentation apply to the 
implementation of GC as well?
Kind regards,
JÃƒÂ¶rg