[prev in list] [next in list] [prev in thread] [next in thread] 

List:       hylafax
Subject:    Re: Some help getting 3.0beta096 installed & working would be appreciated.
From:       Thomas Tornblom <Thomas.Tornblom () Nexus ! SE>
Date:       1995-02-27 10:24:00
[Download RAW message or body]

> Cc: flexfax@sgi.com
> Reply-To: Sam Leffler <sam@cthulhu.engr.sgi.com>
> Date: Tue, 21 Feb 1995 15:12:05 PST
> From: Sam Leffler <sam@flake.asd.sgi.com>
> Precedence: bulk
> 
>     To:  flexfax@sgi.sgi.com
>     Subject:  Re: Some help getting 3.0beta096 installed & working would be appreciated. 
>     Date: Tue, 21 Feb 1995 12:58:14 PST
>     From:  "Ronald F. Guilmette" <rfg@segfault.us.com>
> 
>     
>     In message <199502211533.HAA13345@hyla.chez.sgi.com>, Sam L. wrote:
>     
>     >Alternatively you can suggest to your compiler vendor that they should
>     >learn from everyone else and add support for accepting files w/ arbitrary
>     >extensions as C++ source code.
>     
>     Right.  Extensions like .o, .s, .a, .y, .l, and .f for instance!  Uh huh. :-)
>     
>     Sam, just who is this ``everyone else'' you are speaking of?
> 
> "everyone else" is every compiler that I've encountered: gcc, SGI
> compilers (cfront-based and non-cfront-based), HPUX compiler, IBM
> xlC, CenterLine, and SunPRO C++ (and one other that I can't recall the
> name of right now).  Obviously I haven't tried any PC or Mac-based
> compilers 'cuz I'm not targeting those platforms.
>     
>     I happen to test C and C++ compilers for a living and my own experience
>     is that (a) there are _few_ UNIX-based C++ compilers that allow primary
>     C++ source files to have any old suffix and (b) the _only_ universally
>     recognized C++ source file name under UNIX is `.C'.
> 
> As I said, I've yet to encounter a C++ product for UNIX that
> does not provide some mechanism for stuffing a file w/ an
> arbitrary file extension down the compiler's throat.  Granted
> using an extension like .c++ can be a nuisance sometimes but
> in general the idea that the content of a file is dependent
> on it's name is just something I don't accept.
>     

I have actually come across one, Lucids Energize (RIP). We had it on
evaluation and I tried to compile flexfax with it but it failed
miserably as it didn't recognize .c++. It had some switch that allowed
one to tell the compiler what kind of source it was fed, but the
extension still had to be one of the more used, .C, .cc, .cxx ...

Thomas

[prev in list] [next in list] [prev in thread] [next in thread]