Re: [Cpptool-develop] Linux build problems

[Baptiste L. <gai...@fr...>]

----- Original Message -----
From: "Sven Reichard" <rei...@ma...>
To: "CppTool Mailing List" <Cpp...@li...>
Sent: Friday, October 25, 2002 6:35 AM
Subject: [Cpptool-develop] Linux build problems


> Folks,
>
> the code submitted by Baptiste looks nice.. from the outside.
> Unfortunately, the extended use of the Boost library is more than gcc 2.96
> (the most widespread version of the GNU Compiler Collection for Linux) can
> handle. This is neither Boost's nor Baptiste's fault; gcc 2 implemented
> just an "approximation" of the standard library; in fact it was developed
> while the standard wasn't fixed yet.

Hmm, are you refering to lexical_cast ? I just looked over the compiler
status and it seems to be the only one not to compile with gcc 2.95 with
STLPort. I expect the compilation issue for Boost.FileSystem to be fixed
once the development start again.

Though, it would be nice if we could compile on gcc 2.95 which it a widely
spread compiler I believe.

> So I moved to gcc 3. After figuring out how to run both versions on the
> same machine, this lead to an interesting other problem. This compiler
> tries to implement Standard C++ -- and is very picky about non-standard
> code, much more so than MSVC6, which Baptiste used for development. One
> example was in the blanker; the code looked something like
> std::string str;
> const char* _textBegin = str.begin();

Left over of the original implementation which used std::string::iterator.
But I switched back to using plain char, which works just as well.

> What's wrong with that? Well, str.begin() is a string::iterator, and there
> is no standard way to convert that into a character pointer. However, this
> is easy to fix, since str.c_str() returns exactly what we need.
> Other things that happens are throw-specifications being too general. This
> all is just to let you know that it will take a few days to have a running
> program under Linux. I will fix the necessary files and commit the changes
> once it runs. After that I'll go back to testing.

Are you refering to ParseError destructor and what() methods which should be
marked with throw(), or something else ?

> BTW, maybe my edition of Stroustrup is a bit outdated. Has <stdexception>
> been renamed to <stdexcept>? Both compilers seem to agree on that one.

I think it does. At least I know that it works for VC++ STL, gcc STL,
CodeWarrior MTL...

Baptiste.

> Have a good night (even if the Europeans are about to get up :) )
>
> Sven.
>
> --
> Sven Reichard
> Dept. of Math. Sci.
> University of Delaware
> rei...@ma...