"C" Interpreter

[Geoff Clemm]

In article <2033 at scolex> seanf at scolex.UUCP (Sean Fagan) writes:
>In article <8688 at bloom-beacon.MIT.EDU> raeburn at athena.mit.edu (Ken Raeburn) writes:
>>Mmm... make that three complaints.  #2: If you're debugging something
>>large from source code, it gets to be slow and requires lots of
>>memory.  Sure, it's to be expected, but I can wish. :-)

This I agree with, BUT ...

>Somebody gave a talk about Saber-C at the Summer 88 UseNIX (unfortunately,
>it was all sales fluff.  Disgusting).

This is blatantly false.  There were no technical details because most people
attending didn't want to know how to write an interpreter -- they wanted to know
what would be the advantages of using one, in particular, Sabre's.

>Anyway, the point of this is that
>somebody from AT&T wanted to know how well Saber-C would handle *really*
>large functions, say on the order of 3 million lines.
>The Saber-C Official Sales Fluff Man replied something about 3 million line
>functions, and getting what you deserved...

If this is in fact what the AT&T person asked, then the answer was appropriate -
the idea of a 3 million line function is absurd.  Hopefully (for AT&T's sake),
the conversation is being misquoted, and the question was about 3 million line
programs.  In Sabre-C you are allowed to link in standard object code with
the code you are interpreting -- the fairly reasonable assumption is that you
have 50,000 lines of suspect code that you want to interpret, and the other
2.95 million runs as standard object code.

I'm not sure what motivates this flaming, since my experience with the Sabre
folks is that they are civilized, polite, and interested in satisfying their
customers (I have no connection with the company). 

Geoff Clemm