[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [DotGNU]Jabber.NET
From: |
Gopal V |
Subject: |
Re: [DotGNU]Jabber.NET |
Date: |
Sun, 16 Feb 2003 20:27:24 +0530 |
User-agent: |
Mutt/1.2.5i |
If memory serves me right, Simon Guindon wrote:
> ./jabber/connection/IQTracker.cs:123: no matching method for call to
> `WaitOne(int, bool)'
> ./jabber/connection/IQTracker.cs:123: invalid operand to unary `!'
/me looks at Tum who did the WaitHandle stuff ...
I think I could fix the methods just for fun and make it compile,
but it won't work without fixing the internal calls .. I need help !.
Because 2 of the WaitOne methods seem to accept a "bool exitContext" as
well according to docs hosted on minddog's box ..
Some of those methods are non-ECMA , could you mark them out in
the classes and tests ?..
> ./jabber/server/JabberService.cs:191: could not coerce constant argument 1
Moving onto fixing this ...
> Most of those seem to be Threading stubs not existing?
Added those ..
> I guess after these compiler issues are resolved, we have to start thinking
> about class lib implementation. Two things come to mind, first is
> System.Xml, but I'm not too worried about that, I will try and organize test
> cases and such for that and try and help develop and work with minddog.
I've almost got it building .... you should maybe figure out a way to
test the jabber/protocol/* first ?.
An initial look there was promising and I think we could fake threading
to test that ;-) ... like executing ThreadPool requests synchronously.
Things like that might affect re-draw of GUI apps in general , but won't
matter much for command line apps right now. Also the interlocked variables
can easily be faked for one single thread :-).
I think you should be able to start testing that region sooner or later.
But well confirm my comments with Joe , I might be just too wrong.
> The other major issue, which is the largest thing lingering over Jabber.NET
> compatibility is Threading, I know Rhys is busy with a great many things
> (including work) and from his statements theres no clear solution yet.
> Maybe we can try and brainstorm a little on this topic and see what arises?
Is there any developer who wants to take this up ? .. It would be a lot of
little things to fix in almost all internal calls .. rather than a single
huge patch to a function.
Maybe we should place a 5 day window for an unstable tree in CVS and
fix segfaults in parallel as we find each place where locks are needed ?.
Otherwise it soon becomes a single man task ...
Gopal
--
The difference between insanity and genius is measured by success