This is the mail archive of the mailing list for the Archer project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: C++ Exceptions Status

Tom Tromey wrote:
"Phil" == Phil Muldoon <> writes:

Phil> I've written this up to accompany the call today.

Thanks.  I'm still reading most of this but wanted to address one
particular thing sooner.

Phil> The issue that I am currently wrestling with is how specific is this
Phil> unwind and exception work to C++? Are Java and Python exceptions
Phil> handled in the same way? And how can we avoid language specific
Phil> semantics in very general commands like "next"? Can we?

Our primary focus is C++ and C (though we don't talk about the latter
much since it works well :).

Debugging Python or other languages -- let's simply not worry about
it. If we need this someday (I dunno, we might), it will take an
special effort. Debugging an interpreted language would require far
more than fixes to exception handling -- it would also require changes
to stack tracing, to breakpoint setting, to expression parsing and/or
evaluation, etc.

This has come up a few times, and I think the way I wrote this was not clear. So let me clarify - it was purely a rhetorical question. I'm not looking at any language issues other than C++. But purely from the point of view that the exception support in GDB is pretty generic, and in order to improve that I think we need to move to language specific solutions. I was purely musing along the lines that if we provide a specific solution context regarding C++, how it can be done in harmony with any other language specific support.



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]