Re: Boost.Threads draft library submission

From: williamkempf_at_[hidden]
Date: 2001-03-15 15:23:57

--- In boost_at_y..., Doug Gregor <gregod_at_r...> wrote:
> On Thursday 15 March 2001 11:20, you wrote:
> [snip]
> > > - Add an extra line feed at the end of each line (happens
> >
> > everywhere)
> >
> > This one I don't understand.  Is it a problem with the \r\n
> > DOS/Windows line endings?
> >
> 
> Both GCC and Comeau complain if there is no end-of-line marker 
after the last 
> line of the file, if that line is a preprocessor directorive. It's 
a very 
> strict interpretation of the C++ standard which states that the 
newline 
> character is the last character in a preprocessing directive, so 
there must 
> be a newline after every line in the file that has a preprocessing 
directive, 
> and an end-of-file doesn't count.

Ahh... when you said "Add an extra line feed at the end of each line 
(happens everywhere)" is sounded like you meant there should be two 
line feeds after every line.  My editor is getting in the way when it 
comes to the line feed after the last line in a file... I'll take 
care of this.

> > > - Nested "impl" classes have to be declared friends of their
> >
> > enclosing
> >
> > > classes (semaphore.hpp, mutex.hpp, fast_mutex.hpp,
> >
> > recursive_mutex.hpp)
> >
> > This one I don't buy.  The "nested class" does not use 
the "enclosing
> > class" in any way while the "enclosing class" uses only public
> > methods of the "nested class".  I see no need for friendship in
> > either direction.
> >
> 
> Each of the "impl" classes uses "cv_state", which is a private type.

Ahh... this is a late modification that VC++ didn't have a problem 
with.  Until "finalizing" the draft for submission the cv_state was 
public.  When I made it private I missed the fact that the pimpl 
classes used them and VC++ doesn't seem to care.  I'll correct this 
as well.
 
Bill Kempf