-
RTTI for distribution
Which of the following statements provide a valid reason not to use RTTI for distributed (i.e. networked between different platforms) applications in C++?
A. RTTI is too slow.
B. RTTI uses too much memory.
C. RTTI does not have standardized run-time behavior.
D. RTTI's performance is unpredictable/non-deterministic.
E. RTTI frequently fails to function correctly at run-time.
Ok, in terms of not using RTTI for distribution, I'm completely unsure. Maybe D (but that is just a guess). I don't think it would be E, because to my knowledge the point of RTTI is to improve function at run-time. I'm rather sure A, B are correct for RTTI, but not a good reason why not to use RTTI for distribution. Could someone point me to some good links on this, and/or comment on my thoughts. As I can't find much about the reasons not to use RTTI (for distribution, in particular).
Thx.
-
I'd say the best answer is that it's too slow, but I've also seen some RTTI functions work unpredictably also. The only one I trust is dynamic_cast.
-
There is absolutely no reason not to use RTTI in distributed programs.
However, there is a reason not to use RTTI information in distribution itself: the format of the information (the only information you can get anyway is type_info::name()) is not standardized, so you cannot rely on it across compilers. In other words, you can't take type_info::name() on one node, send the result to another, and expect it to match type_info::name() of any type on that node.
But then, RTTI doesn't provide a means of dynamic instantiation anyway.
-
-
Re: Rtti
This question is exact word-for-word that I had recently at an on-line test I took while applying for a job
Too bad I didn't use google while taking the test
I believe that C is the correct answer
Thanks
-
Thanks for letting us know, binaryman. Wonder if he got the job?
-
Yeah, shame you didn't read this before wading in with the "me too" post far too late to be of any use to anyone.
C Board - Announcements in Forum : C++ Programming
