Re: gEDA-user: On the nitty-gritty of user-experienced problems

[Dan McMahill <dan@xxxxxxxxxxxx>]

On Sat, Jan 15, 2005 at 06:34:53PM +0100, Mario Klebsch wrote:
 
> IMHO the real problem is not tke lack of standards but the unability to 
> cope with it. UNIX had a long tradition in being inhomogenous. Open 
> Source programs written for UNIX were usually developed and used on a 
> wide variety of different systems having a wider variety of different 
> configurations. Portability issues were soon discovered.
> 
> Linux (and especially major Linux distributions) radically changes 
> this. Today the vast majority od users and developers use Linux based 
> systems, often having very simmilar configurations. Dependencies on 
> details easily creep into the projects and often are unnoticed for quit 
> some time.
 
speaking as one who has ported quite a bit of software to NetBSD/{alpha,
sparc,pmax,vax,etc.} you are correct.  I can't count how many bugs
I've fixed where there were braindead tests for operating system
(__sparc => you must be on sunos), endianness (!m68k => you must be
little endian), sloppy use of pointers (storing in ints, not 64 bit
clean), sloppy casting of structs ignoring alignment requirements,
just lots of junk like that.  

-Dan

--