Re: Tasks (exception round-up)

[Phil Muldoon <pmuldoon at redhat dot com>]

Tom Tromey wrote:
> * Exceptions.  Look at the existing longjmp stuff, come up with an
>   estimate for the new work.  

Can you elaborate a little more on this. I was hoping it would become 
clearer as I investigated.

The only issue I have encountered is attempting to step in the catch 
exception handler code causes GDB to lose control of the inferior in 
certain circumstances:

See:

http://sourceware.org/ml/archer/2008-q3/msg00037.html
http://sourceware.org/ml/archer/2008-q3/msg00038.html

Though this I am guessing a stepping/debug-info bug? Stepping when there 
is debug-info available works ok.

> Also look at the current state of the
>   various "catch" commands, maybe look though gnats to see if anything
>   is known to be broken
>   
I've only been able to get catch throw and catch catch to work with C++. 
Try to provide arguments to either returns a junk argument message to be 
returned.

catch throw foo and catch catch foo with named exceptions would totally 
rock. I can attempt to estimate the work there, if needed. I would like 
them, would it be a generally useful thing for others?

Another nit is the catch catch and catch throw breakpoint is in the 
libstdc++ code, not in the inferior throw foo or catch foo code. This is 
personally not useful; I've no interest in the libstdc++ code and beyond 
a libstdc++ developer I'm not sure someone else would. Daniel noted that 
this had been hacked on a bit, but breaks Eclipse (and I guess by 
extension) other clients.

There are several other long-term outstanding issues:

- If you call a function interactively (I guess via inferior function 
calls) and that function raises an exception, control may not be 
returned at the end of that function. (And I guess control not returned 
until a signal is delivered or another random breakpoint encountered).

- You cannot install an exception handler, or raise an exception in an 
interactive session.

I have not tested these yet - they may be already fixed and this is 
outdated info. I'll attempt to call and cause an exception to trigger in 
an inferior function call next.

Regards

Phil