Xref: utzoo comp.sys.atari.st:22916 comp.sys.ibm.pc:39926 alt.msdos.programmer:888 misc.wanted:7247 comp.sources.wanted:9764 alt.sources.wanted:80
Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!van-bc!johna
From: johna@van-bc.UUCP (John Altstadt)
Newsgroups: comp.sys.atari.st,comp.sys.ibm.pc,alt.msdos.programmer,misc.wanted,comp.sources.wanted,alt.sources.wanted
Subject: looking for some case tools
Summary: similar to DEC's program coverage analyzer and program test-bed analyzer
Keywords: everybody needs one
Message-ID: <95@van-bc.UUCP>
Date: 14 Dec 89 06:37:24 GMT
Reply-To: johna@van-bc.UUCP (John Altstadt)
Followup-To: email
Organization: Wimsey Associates
Lines: 38

This is my first attempt at cross posting, I hope it actually
got where I wanted it to.

Has anybody seen programs similar to DEC's CASE tools "program
coverage analyzer" and "program test-bed analyzer" or even a
version of the p.d. program "calls" that is not tied to pipes?
C source code is preferred, since these would be running on
several different machines.

For those that don't know, the first two packages scan through
your code and make sure that every test/branch/function call
in that code is exercised.  They allow you to build a minimal
set of test bed input data that runs through every possible part
of your code (even parts that you didn't know existed).  When
you add extra features to your code, you can verify that all the
old data still makes it through, and figure out exactly what
other data must be added to the input to verify the new parts.

Think of them as advanced code profilers that supply quick
feedback on how data flows through your code, allowing you to
find that one-in-a-million bug that was in code never tested,
and add new features without adding new bugs, since you can
re-test using all your old data and quickly generate new data
that tests only the new code.

If nothing like this exists, and somebody thinks it would be a
fun project, a good place to start might be with the source to
cpr, calls, profile, and the C interpreter in chapter 1 of
Herbert Schildt's book "Born to Code in C".

Please reply by mail, I hope the path given below is correct.
(assuming of course that a .signature was included)

John
-- 

johna@wimsey.bc.ca || ...!ubc-cs!van-bc!johna
                       || ...!uunet!van-bc!johna