STEVEH%mit-mc@sri-unix.UUCP (01/28/84)
From: Stephen C. Hill <STEVEH@mit-mc>
I have significantly increased the speed with which
WordStar 3.3 comes up.
First of all, I had to get rid of the silly MicroPro
advertisements. This was accomplished by using a combination
of DDT and ZDASM. I traced the beginning of the program,
breakpointing just after each CALL instruction until I
discovered the guilty module. Eventually, I narrowed it down
to where I replaced the instruction located at 3CEB with a jump
relative of 3A, which branched around all of the craziness that
MicroPro put up.
That just left getting rid of the long delay loops that
they had left in there. I could have used the same trick of
branching around some code, but the faster fix was to change
the three delay-count bytes located at 2B0 through 2B2. I have
changed them all to 0, and have noticed no problems of leaving
things on the screen with the Molecular MT-100 terminals. Give
it a try with larger numbers, if your terminals have any
problems, although they probably won't.
Someone will probably tell me that there is an option int
the WINSTALL menu, but I sure couldn't find it. WANCHO%stl-host1@sri-unix.UUCP (01/29/84)
Steve, Several years ago someone else also came up with the idea of changing the delay bytes to zero, as well as a couple of other items. He since discovered that zero was NOT correct, although it appears to work. The correct value for each of those delay bytes should be one. About a year ago, I received verbal permission to publish his patch file, stripped of identification, but I forgot to actually do so until you brought the subject up. I will upload the file to the SIMTEL20 tomorrow. Look for PATCHWS.ASM in MICRO:<CPM.WSTAR>. It applies to WS 3.0, and may have to be slightly modified for 3.3. The remainder of the patch concerns itself with intercepting the console status calls, and making the actual calls every 16th. This significantly improves the screen updates! --Frank
STEVEH%mit-mc@sri-unix.UUCP (02/16/84)
From: Stephen C. Hill <STEVEH@mit-mc> Upon re-reading my preceeding message, I discovered a typo (Murphy at work again!) The actual instructions that should have been placed at 3CEB are JR 38 and NOP (18 38 00). The last NOP is just placed there for neatness, since I am replacing a three byte instruction with a two byte. Please note that the relative jump is a 38 NOT a 3A, as previously reported. Please correct the WS files at SIMTEL. Thx.