In case of failures by MS's VC++ 7.1, and in order to get a better sense
whether it is the compiler
or C++ specifications (which I am mostly suspecious of), I would like to try
it on different platforms.

I am talking of course in cases the compilation+linking passes succesfully
(on MS (and Borland C++BuilderX)) and the output is unacceptable.
I am sure (99.99%,... I must leave room to an error on my part...) that the
problem is either in the compiler and / or the C++ standard specifications.

For example, in one case, ALL the program does is outputing typeid(s) of
pointers of 3 instanciated templates where some of the pointers members
point to objects that have not been even declared at all... at the expense
of missing pointers types to object that have been instanciated.
Could not go much wrong with such a program. For commercial reasons, I am
not at liberty to publish the code.

Thanks in advance.

Re: Are there sites to remotely test very small pieces of C++ on cross platforms in case MS' C++/ specs are suspicious? by adebaene

adebaene
Tue May 11 03:29:41 CDT 2004

"David F" <David-White@earthlink.net> wrote in message news:<MVWnc.15914$Hs1.3579@newsread2.news.pas.earthlink.net>...
> In case of failures by MS's VC++ 7.1, and in order to get a better sense
> whether it is the compiler
> or C++ specifications (which I am mostly suspecious of), I would like to try
> it on different platforms.

You can't run the resulting code, but
http://www.comeaucomputing.com/tryitout/ is a reference as far as
standard conformance goes.

>
> I am talking of course in cases the compilation+linking passes succesfully
> (on MS (and Borland C++BuilderX)) and the output is unacceptable.
> I am sure (99.99%,... I must leave room to an error on my part...) that the
> problem is either in the compiler and / or the C++ standard specifications.
>
> For example, in one case, ALL the program does is outputing typeid(s) of
> pointers of 3 instanciated templates where some of the pointers members
> point to objects that have not been even declared at all... at the expense
> of missing pointers types to object that have been instanciated.
> Could not go much wrong with such a program. For commercial reasons, I am
> not at liberty to publish the code.
Couldn't you write a small example that reproduce the behaviour?


Arnaud
MVP - VC