See attached.
Here's a patch to improve the behavior of GNU APL on ANSI terminals.
David B. Lamkins
2014-03-02
The underlying problem is that color numbers for the CSI m sequence
are only standardized and widely implemented for values in the range
0-7.
Also, my reading of an available spec say
Juergen Sauermann
writes:
> I believe this relates to an earlier discussion on this list (around
> Feb. 24, 2014).
>
> You can put the left arg of /⍨ in parentheses which marks it as a
> value at ⎕FX time
> (the default interpretation is function) like this:
>
> (v)/⍨2|v←⍳4
>
> 1 3
Ah, tha
Elias Mårtenson
writes:
> Yay, an Emacs user. :-)
Yes, and one who is glad that there is an Emacs mode for GNU APL. So
thanks again. By the way, we already met on GitHub – I was the one
nagging about the standard keybindings.
> I've been thinking of a feature that would display the result of th
Hi Daniel,
thanks for reporting thisd. Fixed in SVN 149.
/// Jürgen
On 03/02/2014 11:37 AM, Daniel H. Leidisch wrote:
Hello!
Each first of a nested lists behaves weirdly.
↑¨(1 2) (3 4)
1 3
≡↑¨(1 2) (3 4)
1
So far so good, but:
8⎕cr ↑¨(1 2) (3 4)
┌→──┐
│┌─┐ ┌─┐│
││
Hi,
I believe this relates to an earlier discussion on this list (around
Feb. 24, 2014).
You can put the left arg of /⍨ in parentheses which marks it as a value
at ⎕FX time
(the default interpretation is function) like this:
(v)/⍨2|v←⍳4
1 3
/// Jürgen
On 03/02/2014 11:31 AM, Dani
Hi Kacper,
thanks for reporting this. I believe I have fixed this in SVN 149.
There is now a system limit of 65kByte for the size of shared
variables, caused by the UDP protocol used for the communication
between processors.
/// Jürgen
On 03/02/2014 07:42 AM, Kacper Gutowski wrote:
Moreover,
Yay, an Emacs user. :-)
I've been thinking of a feature that would display the result of the last
interactive command in a separate buffer. Similar to the variable tracing
that exists now, but for the return value of the last expression. One would
have the ability to choose display format for this
Hello!
When experimenting at the REPL, I often find myself typing n⎕CR
repeatedly at the beginning of lines. As an Emacs user, it wouldn't be a
problem for me to adapt Elias' mode to insert that for me automatically,
but wouldn't it be convenient to have it as a feature of the system
itself?
Mayb
Hello!
Multiply and add doesn't seem to work with axis specifications:
(3 5⍴1) ×[1] ⍳3
INDEX ERROR
(3 5⍴1)×[1]⍳3
^ ^
INDEX ERROR
(3 5⍴1)+[1]⍳3
^ ^
This works in NARS2000 and Dyalog:
(3 5⍴1) ×[1] ⍳3
1 1 1 1 1
2 2 2 2 2
3 3 3 3 3
(3 5⍴1)
Hello!
Bindings from inside d-fns/lambdas leak into the global environment:
x
VALUE ERROR
x
^
{x←⍵}1
1
x
1
I didn't read much about this, but as far as I understand, bindings
introduced in these functions are supposed to be lexical locals. Anyway,
this doesn't happe
Hello!
Nesting d-fns doesn't seem to work:
{⍵+{⍵}1}1
DEFN ERROR
This works in NARS2000, NGN and Daylog. I'm at r148.
Regards,
Daniel
Hello!
Commuting compress results in a syntax error:
v/⍨2|v←⍳4
This works in NARS2000, Dyalog, and NGN. I'm at r148.
Regards,
Daniel
Hello!
Each first of a nested lists behaves weirdly.
↑¨(1 2) (3 4)
1 3
≡↑¨(1 2) (3 4)
1
So far so good, but:
8⎕cr ↑¨(1 2) (3 4)
┌→──┐
│┌─┐ ┌─┐│
││1│ │3││
│└─┘ └─┘│
└───┘
(↑¨(1 2) (3 4))⍳3
3
This works (i.e. returns ⎕io+1) in NARS2000 and NGN.
Regards,
Daniel
13 matches
Mail list logo
