Search
Friday, 26th of July 2019, 2:39:36 UTC
4:39:07
beach
Good morning everyone!
4:39:20
no-defun-allowed
Good morning beach!
5:06:07
beach
I fixed a few bugs in this figure metamodular.com/SICL/checking-every-keyword.pdf
5:21:38
beach
Fixed! Thanks! I am turning dyslexic. It is really annoying.
5:22:17
beach
I used to be able to detect typos automatically, just by the shape of the word. Not anymore.
5:23:09
beach
I am lucky that I can use a spell checker for everything I write.
5:26:00
beach
Sure, and I compensate by using tools. Still, it takes time.
5:37:54
ck_
Of course. What can I say, "it could always be worse"
5:39:12
ck_
on a different topic: I've never payed attention to this, but Common Lisp floats are not guaranteed or required to be IEEE-754 floats, right?
5:40:15
beach
I think that's true, yes.
5:50:22
ck_
sorry, don't know what happened there.
5:50:36
ck_
but sicl will use those ie3 floats?
5:52:05
beach
I had ideas about a float format that used 60 or so bits, but I don't think I will do it.
5:54:26
ck_
why? so you could use some bits for tagging?
5:56:39
beach
Yes, and immediate floating-point format with almost the same precision as double-precision IEEE.
6:03:45
beach
One more figure for someone to help me debug: metamodular.com/SICL/check-allow-other-keys.pdf
6:06:19
beach
I really like the capabilities of Xfig, but it is not very well maintained, and the UI feels old. I wish someone would write a CLIM version of it with a more modern UI. Plus, it would be Common Lisp, so it could be customized.
6:06:44
ck_
maybe that's not quite a weekend project though.
6:07:08
beach
It is good to know that I am not the only one who likes this program.
6:07:27
ck_
used to put it on my CV even :) nobody ever asked
6:52:48
beach
One more: metamodular.com/SICL/search-one-keyword-argument.pdf
6:53:04
beach
That should be the last one with a loop.
7:39:10
beach
The &REST parameter requires a loop as well, of course.
8:18:55
ck_
I never paid attention to the parameters of printing floats in the spec. Would you have known by heart the interval in which floats are printed without exponents?
8:39:36
ck_
That's reassuring. The interval is 10^-3 to 10^7 :)
8:40:58
ck_
file:///Users/ck/share/HyperSpec/Body/22_acac.htm
8:41:40
ck_
http://www.lispworks.com/documentation/HyperSpec/Body/22_acac.htm
8:47:50
beach
One more: metamodular.com/SCIL/creating-rest-parameter.pdf
8:49:08
ck_
metamodular.com/SICL/creating-rest-parameter.pdf ftfy
8:49:56
ck_
(also you have a 'ficnum' in there, I don't know whether you care about these typos)
9:30:30
beach
Brief power outage. I am back now.
9:49:41
beach
One more: metamodular.com/SICL/initializing-optional-parameters.pdf
10:31:16
beach
Two more: metamodular.com/SICL/no-more-arguments.pdf metamodular.com/SICL/check-even-keyword-arguments.pdf
13:22:39
beach
Chapter 25: metamodular.com/SICL/sicl-specification.pdf
13:23:04
beach
There is still one figure missing (initializing required parameters) but it is an easy one.
13:23:25
beach
And there are some inconsistent naming that I will fix.
13:23:54
beach
Also, I believe I imagined some fixnum comparison instructions that actually don't exist.
13:24:10
beach
I may have to rename and reverse the arguments of some of them.
13:25:54
beach
And I obviously need to add explanations, but I wanted to get the figures in there as soon as possible.
13:26:17
ck_
do you have a convention for which branch is which, when you use a comparison like fixnum-equal?
13:26:30
beach
Yes, true is to the left.
13:28:45
ck_
Got it. Why are there two arrows following from fixnum-sub?
13:30:34
beach
FIXNUM-ADD and FIXNUM-SUB have a "normal" successor and an "overflow" successor.
13:30:52
beach
But there can't be any overflow, so both branches go to the same successor.
13:31:57
ck_
Understood. Should I have known this from the context? I didn't find anything in the document.
13:32:12
beach
It's in the Cleavir documentation.
13:34:23
beach
metamodular.com/SICL/cleavir.pdf page 86.
13:34:57
beach
with a figure on page 88.
13:35:23
beach
Oops, the name of the instruction is wrong in the figure.
13:35:45
ck_
sorry, I haven't yet become familiar with all the repositories
13:36:19
ck_
the figure caption also says -add-
13:36:38
beach
Yeah, add is right, + is not.
13:46:50
beach
Oh, wait, that's not the name of the instruction, but how it is displayed in Graphviz, so it may be correct.
13:47:20
beach
We shortened the names in some cases, so that there would be more room.
13:49:01
ck_
At this point I can only smile and nod "of course". I probably shouldn't even be asking questions at this point.
13:49:52
beach
Heh. Not a problem for me.
13:50:27
ck_
Thank you. I have one on the float topic: it is okay to use integer-decode-float in the printer, right?
13:55:51
ck_
RYU-CL> (float-to-string 6.02214076d23) => "6.02214076d23"
13:57:41
beach
I don't think it would have to be the same though.
13:58:58
ck_
(float-to-string 23.42) => "2.342e1"
13:59:48
ck_
but I know you're talking about the read-write-consistency. I'm just posting intermediate progress to stay accountable. I'll finish it up this weekend hopefully, if temperatures don't get in the way
14:00:33
beach
We finally got some rain, and a 15 degree drop in temperature.
14:01:04
ck_
well that sounds like a lot, when you don't take into account that the drop started from >40 degrees
Friday, 26th of July 2019, 14:39:36 UTC