Comments on the preceeding bugs in System V

[gwyn at brl-smoke.UUCP]

In article <566 at stride.stride.UUCP> bruce at stride.UUCP (Bruce Robertson) writes:
>Just a quick question... these are *serious* bugs in a *production* version
>of System V.  These aren't bugs due to our port to the 68000.  Also, these
>are NEW bugs; none of them existed in Release 1 of System V. I know that
>AT&T does Beta testing.  All of the bugs showed up within days of installing
>the new stdio package, and were tracked down and fixed within a couple of
>weeks.  I don't understand how they could have been missed by a
>thorough Beta test.  Does AT&T just ignore Beta test results, or is it
>just very careless when choosing Beta sites?

It would be interesting to hear how AT&T tests their code.
I have made literally over a thousand bug fixes to SVR2 (V1)
user-mode source code so far, and as you observe, many of
the bugs were newly introduced.  Interestingly, although I'm
one of the major redistributors of UNIX System V source code
to UNIX System vendors, I've never even been approached by
AT&T about being a code or SVID reviewer, test site, etc.

By the way, thanks for sharing your bug fixes.