Re: backtrace

On Fri, 1 Sep 2000, Sascha Schumann wrote: > > I didn't say everything works if it compiles. I said copying va_list's > > works if it compiles. > > See? It is not so hard to write what you mean and it avoids > misunderstandings. You've got to be kidding me. You truly assumed I wasn't aware of the fact that binaries (that obviously got compiled successfully) can crash? Do you want me to talk in formal English each time I talk to you? I can do that, but in my opinion, it would be an insult to your intelligence. Loosening up and figuring out the obvious like most other people would be much better. > > > > > 1.: "The object ap may be passed as an argument to another > > > > > function; if that function invokes the va_arg macro > > > > > with parameter ap, the value of ap in the calling > > > > > function is indeterminate and shall be passed to the > > > > > va_end macro prior to any further reference to ap. > > > > > > > > Where are you reading this? It sounds very odd and I assume it would be > > > > wrong (or otherwise useless to do) on every implementation. > > > > > > "7.15 Variable arguments <stdarg.h>," paragraph 3, page 262, > > > ISO/IEC 9899:1999. > > > > So it falls under the 'otherwise useless to do' IMO. > > "it"? The part I quoted? Yes. > Most of these rules have been established, because some > implementations already behaved that way. The above rule has > no obvious inner logic, but it reflects common practice (at > least at the time of writing; this part dates back to 1990). I would guess so, even though I can't find a good reason for it, and I've yet to meet an implementation on which it would make sense. I guess there's always the possibility that such an implementation exists. > > That's mostly right. Essentially it assumes almost as much as va_list foo > > = va_list bar; assumes. The only difference is that in this case, > > implementations may also use arrays to represent va_list (because arrays > > cannot be lvalues). > > The other significant difference is that the proposal is > aligned with the requirements of ISO C99, because it passes > ap by reference, not by value (which would have the earlier > described side effects). Ok. Zeev -- Zeev Suraski <zeev@zend.com> http://www.zend.com/