Subject: From: Date:

Re: [ECS] Dreaming? Ingo Pakleppa Wed, 03 Mar 1999 22:39:37 -0800

---

That's nothing new at all! The same "problem" occurred ever since the old
Windows 1.0 days - the cause is simply that 49.7 days is 4 billion
milliseconds, the limit of a 32-bit variable - and it has been *documented*
as early as at least Windows 3.1. Thus, it's very much like the Y2K problem
in that respect. And, yes, I have had Win95 run that long, only once, and
with only a single program running. Win95 eventually crashed around that time.

Actually, the problem is inherently unfixable because it's not only Windows
itself that relies on this millisecond timer, but thousands of application
programs use the same one, and many will also not handle the rollover
correctly (for those technical people out there, the problem is related to
the GetTickCount() API).

There may be a problem within 95 and 98 that causes an actual crash, but
again, even if Microsoft fixes that, application programs are still quite
likely to crash in the same situation.

Oh, and before I forget it: NT is also affected by such application program
problems.

The other Ingo

At 08:47 PM 3/3/99 -0500, Ingo Dean wrote:
>Has anybody actually had their Win95 machine stay up anywhere near 49.7
>days?
>
>http://www.news.com/News/Item/0,4,33117,00.html?st.ne.ni.rel
>

---