Re: Questions

[David Durham]

David Sugar wrote:

> Older and poorer compilers have different problems.  For example, if 
> you define a class foo in a common header that uses STL, and have an 
> instance of foo in test1.cpp and in test2.cpp, you may have template 
> generated object code in both test1.o and test2.o even when it has 
> been subroutined and is identical/duplicate function calls.  Newer 
> compilers and linkers throw away duplicate code like this.  Then there 
> was a time not all c++ compilers could compile STL... 

Yeah,  I'm at least saying that my app should be compiled with an ANSI 
compliant C++ compiler... And can finally say that with an ANSI standard 
in existance...

> My mistake, I meant "ThreadLock" when I said RWLock :).  That's why 
> you can't find it.... 


Ahh.. found it.. yes.. I see that it's just a simple mutex, locking out 
all but one reader... My implementation should be much more efficient.. 
it uses two mutexes and a counter...

> As for realpath, I think it would have to provide an implimentation if 
> one is not provided.  realpath is far too useful for enforcing path 
> restricted access policies.

Hmm... okay.. well I may leave a '#error sorry, unimplemented' at first 
for systems without this function...  ;)  But, hopefully all the 
platforms that CommonC++ is said to run on has this function whatever 
its name may be.

> > And how should I send this?  just as an attachment or a patch file?  
> > I've never submitted a patch before, but I suppose there are plenty 
> > of little how-tos on this.
???


-- Davy