[697] in Release_7.7_team
Re: Suggestion for a future release
daemon@ATHENA.MIT.EDU (Craig Fields)
Tue Aug 6 20:55:01 1996
Date: Tue, 6 Aug 1996 20:54:55 -0400
From: Craig Fields <cfields@MIT.EDU>
To: reidmp@MIT.EDU
Cc: release-team@MIT.EDU
You didn't quite get my explanation right.
> Craig explained that this was done once-upon-a-time to fix a key
> reversal problem on an older platform
(No, older platforms tended to do what was expected. It was to bring
Solaris into line with expectations... and in that sense, to be like
the older platforms.)
The problem was that the software worked reverse from what you
expected, _under Solaris_. At the time, it looked like Sun was
boneheaded and had gotten things backwards (or Motif had), and while
things worked as expected on older platforms, they were all backwards
under Solaris. Since it looked like Sun was the exception rather than
the rule, we hacked it.
In hindsight, it turns out that Sun was to become the rule, and third
party software would eventually deal correctly. Now that it is, we are
getting bitten by the problem we were trying to avoid.
> why continue to do it in the future?
It sounds like at this point that doing it costs more than not. It is
possible that putting things back the way they were will break other
things. But it is probably preferable to fix those other things and
"unfix" the software we've been "fixing" now.
Craig
