egcs help

egcs help

[Steve Stevenson]

A colleague wants to use a piece of software developed in the 1990s,  
but written EGCS. Is there help in getting this going? I understand  
the EGCS was merged back into the g++ fold, but g++ as currently  
developed doesn't like the code.


Best regards,

steve
--------
D. E. Stevenson, School of Computing
Director, Institute for Modeling and Simulation Applications
Clemson University, Clemson, SC 29634-0974
864.656.5880 http://www.cs.clemson.edu/~steve

“The surest way to become a pacifist is to join the infantry.”  Bill  
Mauldin

RE: egcs help

[John (Eljay) Love-Jensen]

Hi Steve,

> A colleague wants to use a piece of software developed in the 1990s, but written EGCS.  Is there help in getting this going?

Most likely, the code needs to be ported from "EGCS's proto-C++" to compliant C++ (ISO 14882).

> I understand the EGCS was merged back into the g++ fold...

Correct.

> ... but g++ as currently developed doesn't like the code.

GCC's C++ has become more-and-more compliant to the ISO 14882 standard over the versions & years.

As I recall, EGCS was a splinter effort due to frustration that GCC wasn't keeping up with the rapid evolution of C++ during the standardization process.  There was substantial churn during the C++ standardization process, such that it was a very difficult moving target to pin down.

Since C++ was standardized, GCC has become much, much more compliant with C++.

Old EGCS is not very C++ compliant.  And code based upon it may also be not very C++ compliant, depending if that code used (at that time) well established core C++-isms which have made it into the standard, or the lunatic fringe of proto-C++-isms which did not make it into the standard.

HTH,
--Eljay

Re: egcs help

[Ingo Krabbe]

Am Freitag, 26. Oktober 2007 16:07:53 schrieb John (Eljay) Love-Jensen:
> Hi Steve,
>
> > A colleague wants to use a piece of software developed in the 1990s, but
> > written EGCS.  Is there help in getting this going?
>
[...]

> As I recall, EGCS was a splinter effort due to frustration that GCC wasn't
> keeping up with the rapid evolution of C++ during the standardization
> process.  There was substantial churn during the C++ standardization
> process, such that it was a very difficult moving target to pin down.
>
> Since C++ was standardized, GCC has become much, much more compliant with
> C++.
>
> Old EGCS is not very C++ compliant.  And code based upon it may also be not
> very C++ compliant, depending if that code used (at that time) well
> established core C++-isms which have made it into the standard, or the
> lunatic fringe of proto-C++-isms which did not make it into the standard.

The best way to use such old code (if I were you, the only way) is to detect 
the problems that rise from breaking the standard and redevelop this old 
Software package.

This might be quite few work, depending on how much models need to be touched.

bye

ingo

>
> HTH,
> --Eljay



-- 

Mit freundlichen Grüßen / kind regards

I n g o   K r a b b e
System Administrator / system administrator
_________________________________________________________________________
Elements of Art GmbH         http://www.eoa.de
An der Eickesmuehle 22       41238 D-Moenchengladbach / Germany
fon: +49 (2166) 91567-3418   Registergericht: Moenchengladbach HRB 5855
fax: +49 (2166) 91567-99     Geschaeftsfuehrer: Erik Winterberg,
                                                Sebastian Leppert