c++/6001: g++ problems with nested headers from > egcs-2.91.66 until 3.0.0

[Mario Deilmann]

The following reply was made to PR c++/6001; it has been noted by GNATS.

From: Mario Deilmann <Mario.Deilmann@pallas.com>
To: Neil Booth <neil@daikokuya.demon.co.uk>
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: c++/6001: g++ problems with nested headers from > egcs-2.91.66 until 3.0.0
Date: Wed, 20 Mar 2002 10:41:11 +0100 (MET)

 Dear Neil,
 
 thanks for your fast response.
 
 I tried gcc -E -H on both files and there are a LOT of differences (about
 1200). I have looked into the differences and most of them obviously are
 caused by the different include sequence. 
 
 I have no idea where I should look for the important differences ? Any
 idea ?
 
 I have included the differences files ddiff:
 
 The 1st error was:  
 
 /home/mdeil/bin/../include/g++-v3/sparc-sun-solaris2.7/bits/basic_file_model.h:
 In constructor `std::__basic_file<_CharT>::__basic_file(std::__c_lock*)':
 /home/mdeil/bin/../include/g++-v3/sparc-sun-solaris2.7/bits/basic_file_model.h:39:
 parse error before `;' token
 
 poohmak1/orig> grep -i -n "basic_file_model" ddiff
 2085:>  .....
 /home/mdeil/bin/../include/g++-v3/sparc-sun-solaris2.7/bits/basic_file_model.h
 2088:<       __bas....
 /home/mdeil/bin/../include/g++-v3/sparc-sun-solaris2.7/bits/basic_file_model.h
 
 Mmh, what does this tell me ?
 
 Viele Gruesse / Best Regards
 
 -- Mario 
 
 Mario Deilmann
 Senior Sales Engineer
 --------------------------------------------------------------------
 //// pallas 
 Pallas GmbH / Hermuelheimer Str. 10 / 50321 Bruehl / Germany
 Mario.Deilmann@pallas.com / www.pallas.com
 Tel +49-2232-1896-31 / Fax +49-2232-1896-29
 --------------------------------------------------------------------
 
 On Tue, 19 Mar 2002, Neil Booth wrote:
 
 > mdeil@pallas.de wrote:-
 > 
 > > I found some strange behaviour with some of the g++ versions greater than 2.91 on different platforms with a small C++ code (included)
 > > I investigated a little more and found out that in my opinion the program was badly coded. The header files included each others and also some of the system headers. So I cleaned up all headers so that there are no more
 > > includes in the header files. Collected the used system and program header's and added the std namespace. Wow, what a suprise now all gcc versions include 3.0 accepts the code. 
 > > 
 > > Can you give me some insight's what makes this program so PRBLEMATIC are this ANSI/ISO feature problems ? All compilers that produce errors find them in their own include files which confuses me. I thought that the problems should be located in the original source code. 
 > 
 > We cannot debug this easily for you.  However, you can probably get a deeper
 > understanding of what is happening with the -H option.  Please compare
 > the output of gcc -E -H yourfile.c between the two compilers, and see
 > where they diverge.  It generates an include stack trace.
 > 
 > Let us know if we can close this PR.
 > 
 > Neil.
 >