Re: [xmlblaster-devel] More TestFailsafe

[Michele Laghi <laghi at swissinfo.org>]

Hi Martin,

Martin Johnson wrote:
> Hi Michele,
>
> I not sure of the timer length being the issue.
>
> I guess I question the usefulness of the reached* callbacks. Maybe its WIP?
>
> 10 sec seems reasonable. You get a callback on the reached* functions you
> should be catching this I would have thought.
These callbacks are there to notify the client about state changes in 
the client library. Some clients could ignore them, and some others 
would like to take full control (for example by handling the queued 
messages by themselves). The right place to do that stuff is in these 
reached* methods.

> I have a couple of problems. I was trying to set logging on for the for
> tracing purposes. I dont have the code in front of me now though. There is a
> std::map<const char*, Log> I recall. Shouldnt that be
> std::map<string, Log>. On win32 this was not find() ing anything when
> searching. It was using pointer comparison (memory address equality via
> '<').

You'r right. Thanks I will fix this.

> There also seems to be some problem with the clean up of the static Global.
> The various semaphores etc are invalid shortly after main is left (so it
> would seem) and waiters are not woken as such. This just make the Timeout
> cleanup go into a tight loop forever.
I am not able to reproduce that error. Could you have a closer look at it ?

> I have attempted to see what the semantics where on Linux RH7.3. I have
> built everything ok but just get a lexical_cast exception and crash, are you
> able to help me.
>
> I'm using gcc 2.96-100 and ddd (gdb). What tools do you use. I reall there
> being some 2.96 problems though I recall this was with earlier version of
> RH.
I have no experience with that compiler but I am using gcc 2.95.3 and 
Marcel gcc 3.2. Could you see which lexical_cast was failing ?

Are you trusting ddd ? I believe it is not handling threads right.

Regards
Michele


> Regards
>
> Martin