Message written at 27 Jan 1999 11:09:17 +0000
--- Copy of mail to rasmuse@hem.passagen.se ---

>>>>> "rasmus" == rasmus ekman  writes:

 rasmus> Still complains about some extra CR's, and claims to be 
 rasmus> running \s3voof23.orc and \s3voof23.sco files. That's 
 rasmus> nothing to nag about however. (There was the shit about 
 rasmus> error message line numbers too, but who cares. Document 
 rasmus> it and it's a feature.)

The strange names is of course what it is doing.  In order to simplify
the coding, to disturb as little as possible, i create temporary files
and copy the text from the .csd file into them. I also set up an
atexit function to delete them.  I did think of suppressing the
printed names but I thought that honesty was better!
  I will look at the problems caused by additional free ^M
characters.  The problems I had before was that the C libraries do
different things on different platforms, and while we all know that \n
is ^J there is no guarantee that that is the case.  Some more thought
might help...
  I did have a complaint/comment that in one of teh source files an
additional ^M had crept in, but none of my compilers noticed, and i am
not totally sure how it got there.  Something to do with emacs over
shared files and Samba I suspect.
  This all reminds me of an examination question in computing from the
1960s or early 1970s; "Why do apparently sensible people make such a
mess when it comes to character codes?"  I think it was reworded in
the printed version but was the original question.  Of course on Titan
we did not have control characters.