[Prev][Next][Index][Thread]
ANNOUNCE: 1.99o5
Hi Folks,
1.99o5 took a little longer than I had hoped (but isn't that always
the case). As usual, I was the weak link. This time a toothache got
in the way (I get my first root canal six hours from now). You can
expect my next excuse to be "my dog ate it".
The BleedingEdge/README file is included after my signature.
Because vorlon is so much faster than ftp.ardi.com, I'd recommend
picking up 1.99o5 from there:
DOS:
ftp://vorlon.mit.edu/pub/ardi/BleedingEdge/dos199o5.zip
Linux:
ftp://vorlon.mit.edu/pub/ardi/BleedingEdge/linux199o5.tar.gz
In other news, one of our customers from Microsoft has purchased a
copy of NT and *given* it to us (that's right, no strings attached --
we checked with the Microsoft company store to make sure this was
something he could legitimately do) and we scraped up the money to buy
our very own copy of Windows '95. So now we need to juggle disk space
a little bit and we'll have enough different DPMI providers for us to
get Executor/DOS up to Executor/Linux's level of reliability.
Executor/NEXTSTEP/Intel is running in house although it currently
can't print or access floppy drives. We hope to have those problems
resolved soon so that we can move on to performance enhancements and
re-porting to NEXTSTEP/motorola. We will probably make a BleedingEdge
version of E/NS available even before it's up to full speed.
Melissa Algeo will soon be joining us as a full-time ARDI employee.
This will take some of the pressure off Mat, Cotton and me, although
we won't slack off -- we still want to get 2.0 out ASAP.
--Cliff
ctm@ardi.com
This directory isn't even guaranteed to exist, but if it does,
it may contain binaries from ARDI that are even newer than the
latest 1.99<x> release. There are a variety of reasons why we
might want to replace one or two files in a 1.99<x> distribution
without making a new one, most often to help one of our customers.
If you use software from this directory PLEASE be sure you note
this on any bug reports you might submit. It is counterproductive
for us to look for a bug in one release when it really is a bug
in a different release.
1.99o5 fixes the problem in 1.99o3 and 1.99o4 that prevented Word 5 from
running. 1.99o5 also runs Apeiron, although running the Apeiron
installer under DOS is tricky (and Apeiron requires pretend sound and
refresh). If you want to try running the Apeiron installer under DOS,
you should make a temporary HFV with the Mac Volume name "tmp". It will
need to be about 4 MB long. After Apeiron is installed you can delete
it. Chiral should run, although swoop and Ultima III won't.
Starting with 1.99o5, it is now possible to tell Executor/DOS to ignore
certain disk drives. Here are the relevant switches:
-macdrives ARG
-dosdrives ARG
-skipdrives ARG
-skipaspi
ARG is a set of characters from the set a..z and A..Z. You can denote
a range by placing a '-' between two characters. Example:
executor -dosdrives ac-g -macdrives h-j -skipaspi
tells Executor to only consider drives A:, C:, D:, E:, F: and G: as
candidates for drives to find DOS filesystems on. Only H:, I: and J:
will be examined for a Macintosh filesystem, and there will be no attempt
to directly manipulate the SCSI bus via the ASPI drivers. It is hoped that
configurations that have had trouble with Executor in the past can be
made to work by judicious use of the above switches and the "-oldtimer"
switch mentioned below. Please send your experiences with these switches
to "bugs@ardi.com".
Starting with 1.99o4, Executor/DOS supports the "-oldtimer" switch,
which may solve the problem of Executor not running on some systems.
This switch causes Executor to take 18.2 interrupts a second instead of
1024 interrupts a second. Using -oldtimer will result in lower timer
accuracy, but may fix slow startup times on some machines. The main use
of this switch is to help us understand why some systems do not run
Executor. If you have a system that works when "-oldtimer" is specified
but doesn't when it is not, please send e-mail to bugs@ardi.com.
We've also changed how the "-debug" switch works. In general that switch
does not provide useful information ("no user serviceable parts inside").
We are aware that a number of regressions that were introduced with 1.99o3
are still there. These are being worked on.
Starting with 1.99o3, Executor supports "-system 7", which will tell
Executor to pretend that it is running System 7. However, right now
many filesystem calls that real System 7 apps may expect are not
working. One program that does appear to work with this switch is "NIH
Image 1.58". In fact, with 1.99o4 (at least in our artificial
environment at ARDI), plug-in filters work under NIH Image 1.58. Neato.
1.99o4 fixes DNA strider's copy/paste crashes. Ultima III gets
farther but still dies.
We hope to have our new filesystem code cleaned up, a rough port of
1.99<x> for NEXTSTEP/Intel, and 16bpp and 24bpp support in 1.99p, due
"soon".
Here are the current bleeding edge contents:
dos199o5.zip -- replacement executor.exe
linux199o5.tar.gz -- replacement usr/local/bin/executor
Follow-Ups: