[conspire] (forw) Testing New Keyboard

[Rick Moen]

Quoting Dire Red (deirdre at deirdre.net):

> 
> > On Jun 27, 2015, at 11:22 PM, Rick Moen <rick at linuxmafia.com> wrote:
> > 
> > Robert Johnstone dropped by with his keyboard.  I _think_ I know what's
> > going on and gave him two candidate solutions.  Robert brought just his
> > (new) keyboard and an adapter (about which, more in a minute).  He did
> > _not_ bring his PC system unit.
> 
> So, amazingly, it turns out that the symptoms were as described, just incomplete information.

Yeah, go figure, eh?

This is one of the reasons why I carefully stressed to Robert that I
absolutely believed his account, even while also telling him that the
claimed systems were incongruous, verging on impossible.  Also, Robert
struck me as a painstakingly accurate sort of fellow, even over e-mail,
and it turns out he is:  You should have seen the binder of carefully
printed-out materials he brought along concerning this problem.  The
problem had its own case file!

However, earlier, hearing an 'impossible' set of claimed facts,
meticulously documented, I had something of the same sinking feeling
Trey Harris felt when the '500 mile e-mail' problem was dropped in his
lap.  And, there, too, the 'impossible' facts made sense when a missing
piece of data arrived.

http://linuxmafia.com/pub/humour/500-mile-e-mail
http://www.ibiblio.org/harris/500milemail.html

> Kind of reminds me of how I got involved in Bay Area Linux: I had a
> machine that would segfault on man pages (also certain du/ds
> settings). That problem turned out to be a counterfeit FPU.

Punch line for those following along at home:  /usr/bin/man uses
floating point functions.  I've forgotten exactly why.

So, this was a case where the symptom was '/usr/bin/man and certain
incantations of du and df (I think that's what you mean by 'ds') cause
consistent segfaults on my Linux machine' -- which is a head-scratcher
-- and the solution turned out to be 'your hardware math chip is
defective'.