More fun with PowerLisp on 68k Mac (7.5.5)

PowerLisp hasn't frozen my Macintosh for the past several weeks, so
I haven't needed a cold restart since I last posted that I've
avoided that. As a result, the current uptime for my Macintosh is
longer than the average uptime for FreeBSD Unix on my ISP's shell
machine. (It needed a preventive re-start on Aug.28 and again on
Jul.05, while my Mac with PowerLisp sailed through all
that time *up* 24/7.)
So now I think I can say that PowerLisp on my Mac is my preferred
Common Lisp development environment
(among the very few currently available to me; MACL 1.2.2 on my
Mac Plus was even better before my Mac Plus fried its power+video
board again in mid-1999 and I couldn't afford to get it fixed
again has I had twice before, especially with Y2K coming at the
end of that year and the date+time control panel for System 6.0.3
didn't support years past 1999)
where I can keep stuff loaded for weeks/months unlike CMUCL
on FreeBSD Unix where after each system re-start I have to load my
Lisp stuff again.

So what have I been doing in PowerLisp on my Mac Performa? After
finishing (Aug.22) my IDE station for indent-prefix
(using indent prefix from first line, and using whitespace indent
from second line, two different IDE stations)
I have kept them around and have used them lots of times to format
newsgroup articles, including three times already in this article.
On Aug.24 I finished my IDE station to convert from my Caller-ID
log to the form contents to paste into the annoying-call complaint
form on whocalled.us, and have been using it on a regular basis.
And starting 2008.Sep.04 I've been developing a new s-expression
prettyprinter using an idea I tried to develop about 20 years ago
in PSL (similar to "dynamic programming" in that it generates a
bottom-up list of all possible abstract formattings for
each sub-expression and combines the best abstract
formattings at each level to produce all reasonable
abstract formattings at the next higher level)
but I think I have it right this time, keeping track of both the
longest line and also the length of the last line separately within
each abstract formatting, so that I can correctly put the extra
closing parens on the same last line if it fits else on a new line
by itself. The main combiner for using the best abstract
formattings at one level to generate one possible abstract
formatting at the next higher level per a given tentative line
width seems to be working as of a few minutes ago when I started
composing this article, at least for simple cases which exercise
all the code. Next I need to write the next higher level of code
that cranks down the tentative line width to generate all possible
formattings at that next higher level...

Note: My jargon "IDE station": The idea is that inside the IDE
there's a virtual (software) station, where "station" is used in
the sense of a place where you bring work to be processed, similar
to these standard definitions:
* a facility equipped with special equipment and personnel for a
particular purpose; "he started looking for a gas station"; "the
train pulled into ...
* A roller coaster's station is where the passengers board and
alight from the trains. ...
* ... A place where one stands or stays in order to perform a
task; ...
* ... a permanent or temporary location where scientific
observations and measurements are made. ...
The way an "IDE station" works is that there's a permanent
s-expression in the IDE, which is an executable form which when
executed reads lines of input until a blank line is found then
processes those non-blank lines and edits the result into the IDE
buffer. I paste the input data directly after the form, then mark
the form plus input with mouse, then submit it to PowerLisp by
Clover-Return, and the output appears inserted right after the
input. I then copy that output and paste wherever I need it, such
as replacing the original text in this message I'm composing.

If anybody knows a more standard jargon for my "IDE station",
please let me know (followup or e-mail through SpamGourmet).

At this stage in my use of PowerLisp, the most annoying things are:
- I need to repeatedly copy my working under-development code from
the IDE to an ordinary text edit file and save to disk, because
saving the IDE file doesn't actually happen!!
- I need to carefully tag all local variables with x- prefix while
doing line at a time development, then rename them without the
prefix when defining the function, because all variables are
SPECIAL and it totally breaks down if you already have global
bindings for something that is also local (lexical).
- Whenever I accidently click mouse or type immediately at the
front of an incomplete s-expression, it spends about ten seconds
scanning to the end of the IDE to try to find a matching close
parens. If I'm typing, it spends about ten seconds per character
I've typed, echoing each character only after that delay, if I'm
typing multiple characters just before an unmatched open-paren.
- Whenever an error is signalled, either because I goofed and
triggered an error within the library, or I am testing my own
ERROR call to make sure I have the format string and params
correct, it takes about a half minute before it finally prints
the error message to the IDE.
- And it doesn't implement GET-UNIVERSAL-TIME or related functions,
so there's no way to run anything that depends on date&time for
proper operation, such as my optimal-flashcard-drill algorithm.

So if anybody has an old Macintosh (68xxx CPU) and needs a halfway decent
Common Lisp for it, and is considering PowerLisp, please contact me
for info about what to **avoid** doing so that you won't freeze your Mac
and need a cold restart.