Yes I suppose it should. More manageable.
Be nice if the result could look more like first aid than a semester course.
On Tue, 5 Apr 2022 at 22:33, chris burke wrote:
> Shouldn't all debug pages go on or under the page
> https://code.jsoftware.com/wiki/Debug? For now this page is primarily for
Hi everyone,
I made a cheatsheet for all J primitives about a year ago. I thought somebody
might find it useful too. Here's the link:
https://sergeyqz.github.io/jcheatsheet/.
The main idea was to have a one-page plain HTML table of all primitives with
short descriptions, synonyms (e.g. Member
Your cheatsheet is very nice. However, in my browser I keep a tab
with file:///Applications/j64-807/addons/docs/help/dictionary/vocabul.htm open
always and at all times. And why it wins out over NuVoc etc is because it
fits on the screen. I never have to scroll to find the relevant
information. Yea
On Wed, Apr 6, 2022 at 4:20 AM Ian Clark wrote:
> Be nice if the result could look more like first aid than a semester course.
Something like this?
http://www.andyheld.com/wfr_info.htm
Or more like this?
https://winapps.umt.edu/winapps/media2/wilderness/toolboxes/documents/ranger/SAW_WFA%20Ref
On Wed, Apr 6, 2022 at 4:37 AM 'Sergey S' via Programming
wrote:
> I made a cheatsheet for all J primitives about a year ago. I thought somebody
> might find it useful too. Here's the link:
> https://sergeyqz.github.io/jcheatsheet/.
That looks like it would be nice as a wiki page, with
[[Catego
Present: Art Anger, Jon Hough, Devon McCormick,Raul Miller, Bob Therriault
1) After some discussion it was decided that we would look at setting up the J
playground as an independent project. It has accomplished much of what the J
wiki requires with every indication that it will be able to do mu
Ha-ha!
(…and I thought J online help was bad!)
What was I hoping for when I said "First Aid"? If I'd said "In Case Of
Fire" you'd find equally hideous examples.
I meant "First Aid for the J novice". I was hinting at something brief and
helpful.
But what's the briefest, most helpful thing one can
I think that's a little harsh on the debugger.
For my first 20 years of programming I single-stepped through every
program before trying anything bigger. That's an effective way to
develop IMO. If you keep your eyes open you can find most of your
errors. The J visual debugger has good suppo
On Wed, Apr 6, 2022 at 5:01 PM Henry Rich wrote:
> I think the debugger is robust enough now that I could single-step
> again. Maybe I'll try. I certainly wouldn't discourage a newcomer who
> wanted to do that.
In my experience, single stepping is fairly robust, with one exception
(which might
I too would like to see Sergey's cheatsheet on J Wiki.
There's no rule that says NuVoc should only have one portal.
Why does https://code.jsoftware.com/wiki/NuVoc take the form it does?
Because it's the best cheat-sheet? No – that's never been established.
If memory serves, it aimed to resemble Vo
It surprises me that the results are unpredictable. The code seems
clear: debug suspension happens after an error or before a stop, but
always between sentences (except for stops on tacit verbs). If anyone
finds anomalies let me know.
Henry Rich
On 4/6/2022 5:06 PM, Raul Miller wrote:
On W
The card that Ric Sherlock & I did,
https://code.jsoftware.com/wiki/File:J602_RefCard_color_letter_current.pdf,
tries to get it all on 2 pages. We are working on a J9 version, as a
very-low-priority task.
Henry Rich
On 4/6/2022 5:38 PM, Ian Clark wrote:
I too would like to see Sergey's chea
…now that's a card I did make use of!
As it should be used: printed out 2-sided, laminated and propped-up beside
the screen.
On Wed, 6 Apr 2022 at 22:52, Henry Rich wrote:
> The card that Ric Sherlock & I did,
> https://code.jsoftware.com/wiki/File:J602_RefCard_color_letter_current.pdf,
>
> tri
dbcut (13!:19) seems to go up one stack level and immediately reexecute
the line that failed. Huh?
Unless you have a stop on that line, you will rerun it and it will fail
again. If you do have a stop on the line, the reexecution is nugatory.
Can anyone suggest a reason why 'cut back' should
https://gist.githubusercontent.com/rdm/3ce3503ee63b51cd49c97ad185e1dd6f/raw/100e83ba7a87f10b844271180318c37fee1304dc/gph.txt
The text at that gist represents a "defective graph"
require'web/gethttp'
text=: gethttp url
'A B C'=: |:640 #.inv<.0".text rplc LF,' '
part1=: A=/B
part2=: B=/C
To get a
It works for me therefore I'm not sure if I understood what you said. Can
you give a simple example?
On Thu, 7 Apr 2022 at 7:20 AM Henry Rich wrote:
> dbcut (13!:19) seems to go up one stack level and immediately reexecute
> the line that failed. Huh?
>
> Unless you have a stop on that line, yo
Raul provides this example:
https://github.com/jsoftware/jsource/issues/131
Run
example2 10
It fails (valence error in problem). Suppose you want to look around in
the caller's frame. You try 13!:19'', but it restarts immediately,
calls problem again, and fails.
What does it look like wh
Haven't used the debugger extensively and haven't used dbcut. That said:
1. If you change some global state first, perhaps the code will start working
2. The dictionary says (emphasis mine):
Cut back one stack level, _stopping_ at the line at the next stack level
-E
On Wed, 6 Apr 2022, He
Even with emphasis, I can't say I'm 100% sure what that sentence says.
It means just what Roger chose it to mean, neither more nor less.
I am looking for comments on whether the current behavior is desirable.
Henry Rich
On 4/6/2022 7:33 PM, Elijah Stone wrote:
Haven't used the debugger exten
I don't know. However why did the sentence
1 2+3 4 5
give domain error instead of length error?
On Thu, 7 Apr 2022 at 7:29 AM Henry Rich wrote:
> Raul provides this example:
>
> https://github.com/jsoftware/jsource/issues/131
>
> Run
>
> example2 10
>
> It fails (valence error in problem). Sup
Suppose you want to look around in the caller's frame. You try
13!:19'', but it restarts immediately, calls problem again, and fails.
That seems like a valid use case. More generally, it seems like it would
be useful to poke around in _any_ stack frame, without destroying the ones
beneath (o
That sentence never executed, because (problem) was executed dyadically
and it has no dyadic definition.
Henry Rich
On 4/6/2022 7:41 PM, bill lam wrote:
I don't know. However why did the sentence
1 2+3 4 5
give domain error instead of length error?
On Thu, 7 Apr 2022 at 7:29 AM Henry Rich w
Quite right, but I don't want to go there yet. I want to see if I'm
misunderstanding 13!:19.
Henry Rich
On 4/6/2022 7:41 PM, Elijah Stone wrote:
Suppose you want to look around in the caller's frame. You try
13!:19'', but it restarts immediately, calls problem again, and fails.
That seems
I use 13!:19 often and have not seen the behavior you described.
On Wed, Apr 6, 2022 at 7:33 PM Elijah Stone wrote:
> Haven't used the debugger extensively and haven't used dbcut. That said:
>
> 1. If you change some global state first, perhaps the code will start
> working
>
> 2. The dictionar
On Wed, Apr 6, 2022 at 7:41 PM Elijah Stone wrote:
> > Suppose you want to look around in the caller's frame. You try
> > 13!:19'', but it restarts immediately, calls problem again, and fails.
>
> That seems like a valid use case. More generally, it seems like it would
> be useful to poke around
Just did this:
foo0=: 3 : 0
y=. >:y
foo1 y
)
foo1=: 3 : 0
y=. y,XX
y=. y*10
)
4!:55 <'XX'
1
foo0 10
|value error: XX
| y=.y*10
|foo1[1]
y NB. Currently undefined because XX is undefined
11 , XX
13!:19''
|stop: foo0
| foo1 y
|foo0[1]
y
(To be clear, I do not think continuations are a _bad_ idea; but they do
represent a unique set of challenges complications, as demonstrated e.g.
by dynamic-wind.)
On Wed, 6 Apr 2022, Elijah Stone wrote:
Suppose you want to look around in the caller's frame. You try
> 13!:19'', but it restar
Thanks much. That makes sense. The key point is that it stopped in
foo0 when you did the first 13!:19; in Raul's example it seemed to
continue. The stop is sensible behavior; I will look into what happened
in Raul's example.
Henry Rich
On 4/6/2022 7:51 PM, Devon McCormick wrote:
Just did
We are working on a J9 version
GREAT!
Sergey's sheet is nice and well designed.
But IMO it is for a somewhat experienced J programmer, while the old
J6.02 Reference Card had details and __examples__ useful to the newcomer
e.g. Format. It is dense, can be printed and posted. Thanks to Henry
29 matches
Mail list logo
