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: [RFC] Set unwindonsignal to on by default

Tom Tromey wrote:
Phil> I propose it is more useful and "safe" for it to be on.

Jan's response made me wonder whether we could detect that this signal
came from the abort call in the terminate code, and ignore it only in
this case.
Sure can - well we can set an internal breakpoint in std::terminate in an inferior function call. It is an option. I thought I would test the waters on changing this option default first. It is the simplest approach.

To play advocate, this change does not do any "damage" to a C programmer. If in an inferior function call they get, in Jan's example, a sigsegv, the C programmer just has to just set the flag to "off" and make the inferior function call again. In the C++ exception example, it really blows the inferior out of the water.

As I mentioned to Jan at the end of my reply, we can start setting breakpoints and deducing the origin of the (abort) signal, but it is a more complex solution. I have a partial patch based on this solution. I'm open to all suggestions.



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