In article <37afe9de@news.povray.org> , Nieminen Mika <war### [at] cctutfi>  
wrote:

> : Yes, I know. As I said, with a different Mac compiler (= different C
> : libraries) we get problems like that, too. Unfortunately  you never know
> : (without knowing where in the code it happens) if a compiler bug or a code
> : bug causes writing to an invalid location.
>
>   "My code is NOT buggy. The problem must be in the compiler. Or Windows.
> Or in intel's cpu. But my code is NOT buggy!" ;)

:-)  "My computer doesn't work, the screen is always dark. The problem is
the OS, the CPU, Windows, and it is all the programmer's fault.  He won't
plug in the power cable for me!"

But serious: Of course we all know that POV-Ray has a lot of small bugs,
like all software of that size has (sad but true). Sometimes it gets worse
because a specific platform/compiler doesn't like a particular bug which is
mostly "harmless" on other platforms.

I.e. on the Mac I recently had to chase a bug that sounds "impossible": When
inserting a CD the POV-Ray core code would abort with an error when parsing
a macro that is in an include file.
Now you say: What does POV-Ray care about CDs? - Nothing, of course! BUT
when mounting the CD the Mac OS would for some reason for a short time
change the global default directory to read from (there is no rule that it
may not). This wouldn't cause any problems using Mac OS file-io (which
doesn't depend on file and path names to find a file "again").  Now, whose
"bug" is it: Is it POV-Ray using standard C io, the Mac OS written in 1983
in Pascal, or the compiler library mapping the standard C io to the Mac file
handling?  Everything contributes to it resulting in a bug...


    Thorsten


____________________________________________________
Thorsten Froehlich
e-mail: mac### [at] povrayorg

I am a member of the POV-Ray Team.
Visit POV-Ray on the web: http://mac.povray.org

Post a reply to this message