what -I- consider a bug in *cabe
Fairlight
fairlite at fairlite.com
Thu May 31 23:12:43 PDT 2007
I written my subject line at least in a way that admits up front that it's
my -opinion-, and this may not actually be a bug. However, I consider it
one subjectively. We've had a few things that I have brought up lately
that were apparently design decisions and some have felt that my calling
them flat-out bugs was incorrect, so I'm just doing this up front--it's my
subjective opinion here. That said...
Working on Windows 2003 Server. The uhm...console windows aren't really
friendly about copy&paste of selected areas in terms of speed of use, since
you have to throw it into mark mode and it acts a bit flaky, and do it one
line at a time, etc. So I basically just had two windows back to back and
was transcribing some of my older, known-good code from one to the other
far faster than I could copy and paste line by line. Apparently slightly
too fast...
I just had an issue with a time discrepancy. Basically there's a
calculation I do that determines if 30min or more have passed between two
events, and it was fine if I changed the hour, but it was -not- catching
the difference right at the 31min mark.
I was looking over the code wondering where my miscalculation was, except
this code ran the session expirations for the file sharing system that
was up in the fP Room for several years before I took it down. I knew it
should work.
Looking over the logic, I noticed something that obviously threw the
calculation:
TMin=mid(TDiff,"$","2")
That was the segment that caused problems. Obviously I was going for "4"
in the starting position, and I did that whole, "I'm holding shift for more
than one keystroke by accident," thing. I know -how- it happened given the
character that was there...a "$" instead of a "4". How it happened is not
in dispute, and that's obviously my own fault.
However, could someone please explain to me why having a string literal
that is non-numeric in a mid positioning argument should not be a syntax
error worthy of failing syntax check at save time? Unless that "$" was
being cast from char to a byte numeric, I'm just not seeing it. Further, I
can't see that being intentionally something anyone would actually use in
fP at all, at least in this context.
IMHO, this could use a syntax flagging, flogging, and possibly some harsh
language from the parser. YMMV, but that's my take on it, having been
bitten by it. Mildly of course, and it took all of 5min to find and fix
once I realised there was a calculating error and looked at the code. It's
just my contention that cabe never should have accepted that braindead,
typo-riddled code that made -no- semantic sense whatsoever.
That was an interesting one. I was quite surprised to see that, actually.
mark->
--
Fairlight-> ||| "The very powerful and the very | Fairlight Consulting
__/\__ ||| stupid have one thing in common. |
<__<>__> ||| Instead of altering their views to | http://www.fairlite.com
\/ ||| fit the facts, they alter the | info at fairlite.com
||| facts to fit their views....which |
||| can be very uncomfortable if you |
||| happen to be one of the facts that |
||| needs altering." --The Doctor |
More information about the Filepro-list
mailing list