Quote:
>So the *only* issues addressed with Service Pack 2 are those that cause
>a buggy application to crash that previously didn't crash. In other
>words, Microsoft is admitting that it's acceptable to have bugs in your
>application and that they will put out a Service Pack just to accomodate
>your buggy application. This indeed sounds very fishy.
>--Matt Brunk
Correct me if I'm wrong, but:
All applications have bugs.
Most applications are tested on a target platform before release.
The public buy some applications from companies other than Microsoft.
Then, when companies that buy microsoft products write a program that
*appears* to work, and release that to the public, and that same product
works for 2 or 3 years without an issue, and then Microsoft issues new DLL's
which replace functionality at the C run-time library level. Magically
programs stop working.
The semantics might be slightly different, but not however, the published
semantics.
Is it wrong for Microsoft to make improvements that ostensibly benefit
everyone? Maybe we should go back to the bad old days, and statically link
everything into our own code. Then we have bugs forever, and must
continually reissue code to customers. We also use up more space on the
customers hard disk.
Alternatively, we can take the rough with the smooth, and rely on Microsoft
to improve matters.
FWIW: I believe that Microsoft did this heap bodge to prevent them having
to sort out the problems with Hyperterminal, etc., which was bundled with
windows itself, and is broken by the problem. Given two solutions
(re-releasing all affected programs, or bodging the DLL), they chose the
same solution I would have, to maximise the benefit to everyone.
In addition, the fix seems to have been present on the Critical Updates
site. Unfortunately for us developers, it was described as being associated
with Microsoft Library, not MSVCRT.DLL.
Tony Wieser
Wieser Software Ltd.