[RESOLVED] C++ standardization and other issues.

[Nemanja Trifunovic]

Hi Jonathan.

Thanks a lot for your answers.

1) Yes, the standardization of C++ is a purely voluntary process. I, along with Herb Sutter, are the Microsoft representatives on the ISO C++ Committee and we are all volunteers. But if you want a truely independent Standard this is the best way to do it.

Now call me cynical, but I'd prefer big, greedy capitalist boys putting some money behind the effort. If you look at the competition (Java, C#), they are progressing much faster these days. Of course, a healthy dose of conservatism is good, but C++ is pretty much stagnant. 10+ years between the two versions of the Standard is way too long.

Have you had a look at Boost (www.boost.org)? Boost is an incredibly supportive development and proving ground for new C++ libraries - if you feel there should be a library for Unicode (warning: this one is a political minefield), XML, etc., this is probably the best place to start.

Sure I have. In fact, I wish I could submit a library to Boost myself, but can't find time to properly start it in my own time, let alone finish and test it. That's exactly why I would like people improving C++ on their paid time.

Besides, Boost covers the libraries only, not the language itself.

2) Are you compiling managed code? I build the Visual C++ compiler many times a day and I never find the link times a problem.

No, it's good old native code In fact, there is a "sub-thread here": http://groups.google.com/group/comp....8232287473bad5

where some of us accuse gcc of being too slow, but it turns out that VC 8 is getting closer.

3) Ah "export" - less said the better :-) I know the EDG guys well and I can tell you that they still bare the scares from implementing "export". There has got to be a better way to compile templates - problem is I am not certain that "export" is it.

Well, it was my fault to mention "export". In fact, two-phase name lookup is more important, IMHO. I know some people that were really confused by the fact that their code compiles fine with VC++, but fails with gcc. Of course, they were looking for the problem on the wrong side

Thanks.