> But the other side of the coin is, why hold up a release for a relatively
> minor
> bug like this? The number of people who use bitfields is very small, and
for
> those that do, there are simple workarounds. Why hold up a major release
for
> something this minor?
I guess everything is relative isn't it? For me it renders the debugger
completely useless in the R5 upgrade. R6 will be useless to me too until
they fix it. It will be far less effort for me to stay with the initial
release of R5 even though it crashes frequently from another
subsequently-fixed bug, rather than sitting there with paper and pencil
trying to unpack complex bit-fields in structures trying to figure out what
values are present.
I think one of the major purposes of a debugger is to examine data at
run-time. And where it is MOST needed is in presenting data in complex
structures - after all it's easy enough to look at a word in hex and see
what the value is - but it's much harder to look at bits 9-13 and see what
the value is right away, so the fact that functionality is completely broken
is in my mind quite significant.
The other flaw, as someone else alluded to, is that by releasing R6, it is
now unlikely that I will see this fix in R5, so I will be forced to upgrade
to get something fixed. Too many software vendors resort to this
well-you-have-to-buy-the-upgrade-to-get-the-fix approach to supporting their
software and this is part of the reason that we are constantly saddled with
unreliable software since, as you well point out, any new release is always
going to introduce new bugs.
As for CW, I would merely point out that this type of problem here should
have been caught by automated test routines. In a large application, it is
ESSENTIAL to have automated test routines that throw a series of tests at a
compiler/debugger and verify that the correct values are returned. I guess I
am just assuming that would be the case with CW and am puzzled as to why
their test routines would not have picked up on something as obvious as
this.
I may be in a minority, but I would far rather have a solid, reliable R5
than a buggy R6 release. My concern about upgrading to R6 now which still
has this bug is that I may have to wait until R7 to finally see it fixed -
especially since you and others may view this as "something this minor"
(that's not a criticism of your statement by the way as I am quite willing
to accept the assertion that most people may view this as minor). I'm merely
pointing out that completely breaking some existing functionality is, in my
mind, a more serious problem than having random flaws in some new
functionality).
____ ____
/ ___) / ___)
( (___ ( (___
\____)heers! \____)ESD
Pimlico Software, Inc.
Home of DateBk3 and WeekView: http://www.gorilla-haven.org/pimlico
The Dewar Wildlife Trust, Inc.
Home of Gorilla Haven: http://www.gorilla-haven.org
. . . . . .
