[Date Prev][Date Next][Subject Prev][Subject Next][
Date Index][
Subject Index]
Re: Can v. 112 of Jumbo U2 Crash OS/2 ?
- Subject: Re: Can v. 112 of Jumbo U2 Crash OS/2 ?
- From: Robert Holmgren holmgren@xxxxxxxx
- Date: Tue, 23 Jul 2002 16:17:56 -0400
** Reply to message from "J. R. Fox" on Mon, 22 Jul 2002
23:04:51 -0800
Can U2 crash OS/2? I doubt it. Never has here! Carl's never mentioned it.
I'm running MCP v2 (Warp v4.52) rather than ECS, Carl's several fixpacks back I
think (around FP12). Look, all that happens when you LOAD U2 is that it gets
indexed. Nothing is "run". XyWrite could be crashing the VDM, though, in
numerous ways. It could be running out of memory, and that in turn is causing
an exception in the kernel. When you successfully load XyWrite including U2,
what does VA/NV $M+6 say? If it reports over 20[Kb], you are dangerously
close to OOM (out of mem) in the programming part of memory. Try excluding the
LOAD of U2 in STARTUP, and see what happens when you repeatedly launch XyWrite
-- does it crash? Make a note each time of the value of $M+6 -- then compare
with value when you *do* load U2. How much STUFF are you loading anyway? There
are limits! 28Kb is the absolute limit for programming memory!! U2 needs about
11Kb, which is nothing. But if you have a lot of programming Save/Gets attached
to specific keys, that will gobble up memory like candy. (It's a very stupid
and wasteful way to organize things too -- all code is loaded verbatim, whereas
with U2 all you load is a 15-20 byte pointer to each frame.) Don't load HLP
(and/or DLG) unless you really need it. Send me your STARTUP.INT as an
attachment. Humongous KBD files with a lot of keystroking around are also a
no-no -- put the keystrokes in a U2 frame, it's much more economical oir memory!
-----------------------------
Robert Holmgren
holmgren@xxxxxxxx
-----------------------------