On 2014-02-04 12:17:16, Elias Mårtenson wrote:
When I look at the result from ,/1
2 3 it looks like an array that contains a single element: another array with
the values 1 2 3 in it. But it isn't.
Technically you aren't wrong at all, it IS an array that contains a
single element. But
Hello,
There are multiple problems with shared variables facility.
1. Shared variable protocol is not safe in multi user environment.
When user starts apl with shared variables enabled, it removes all
other users' sessions from shared database as “stale” so currently
even honest user without
On 2014-02-12 12:58:05, Elias Mårtenson wrote:
The key feature that is needed is a way to be informed when a variable is
changed.
Shared variables have such feature. Though, it's probably not the right
place for you too look in this case, as shared variables are whole separate
name class and
On 2014-02-14 16:15:33, Juergen Sauermann wrote:
fixed in SVN 120.
In r122, it doesn't give domain error anymore, but I think that
results are not correct:
⎕RL
1
8⎕CR Y←⊂[1]?2 10⍴2
┌→──┐
│┌→──┐ ┌→──┐ ┌→──┐ ┌→──┐ ┌→──┐ ┌→──┐
On 2014-02-20 01:05:16, Elias Mårtenson wrote:
∇N disploop S
→(N=0)/0
next:
S←next S
disp ← '.#'[1+S]
⎕DL ÷4
N←N-1
→next
∇
In this function, next is 2 (a label) rather than function you
defined earlier. So you literally have [3] S←2 S.
By the way,
On 2014-02-16 18:09:06, Juergen Sauermann wrote:
Ad 1) I changed the assertions Symbol.cc to short warnings visible in )MORE.
I wouldn't guess to check )MORE upon getting VALUE ERROR on shared
variable, but I guess it's better than failed assertion.
Ad 3) hopefully fixed.
I've mistakenly put
On 2014-03-11 17:29:08, Frederick H. Pitts wrote:
If one executes
i ← 0
63 1 ⍴ { ⍵ , i ← ¯1 + 2 × i + 1 }¨ 1 + ⍳ 63
one should get a 2-column table of bit count and corresponding maximum
positive representable integer (assuming high order bit is a sign bit).
The
On 2014-03-30 20:50:31, baruc...@gmx.com wrote:
At first glance, it seems to work: http://baruchel.hd.free.fr/apps/apl/i/
You can know use it as an online interpreter. I will add some colors/themes
in the days to come, but the most difficult part is done.
Great work!
It's really nice to have
On 2014-03-31 15:37:45, Elias Mårtenson wrote:
Are you sure it's a bug? You seem to be taking the value of r when it has no
value.
I think the problem is that you can't edit a function that is currently
on SI stack, not that Parse erred.
-k
On 2014-03-31 15:40:06, Elias Mårtenson wrote:
The Nable editor should (in my opinion) open a separate editor just like the
Emacs mode does. :-)
That's one way to do things.
But wouldn't it be nicer to just use something like )EDIT rather than
hijacking ∇ for that purpose? I actually hated the
On 2014-04-26 12:08:14, Juergen Sauermann wrote:
I could make the ∇-editor aware that a )COPY is in progress and that
functions shall be deleted automatically by the ∇-editor.
I like a lot that there is now a command that sources another file
raw, much like . in shell or #include in cpp. It's
On 2014-04-26 21:15:43, Elias Mårtenson wrote:
On 26 April 2014 20:52, Kacper Gutowski mwgam...@gmail.com wrote:
∇mean[0]
R←mean B
R←(+/B)÷⍴B
∇
This can be )COPY-ied many times without any problems as it enters the
editor with a name only and then explicitly rewrites header line at [0
Apparently, when array contains a big number that is to be displayed
in exponential notation, the whole column containing it is printed
with exponent. If array of mixed type has char and such big number
in the same column, PrintBuffer segfaults on it:
+X←?2 3 3⍴100
40 18 54
19 66 55
55 60
Outer product of any function with empty left or right argument,
results in segmentation fault:
'abc' ∘.= ''
SEGMENTATION FAULT
-- Stack trace at main.cc:122
Thanks for previous two. Is it expected, by the way, that assert
ends in a segfault? Isn't this also a bug?
Anyway, any scalar function given empty right argument and non-conforming
non-empty left argument misses check for shape equality:
⍬ = 1 2
LENGTH ERROR
⍬=1 2
^^
0↑0
SEGMENTATION FAULT
-- Stack trace at main.cc:122
0x7f81f5e43b45 __libc_start_main
0x4359b5 main
0x52a66d Workspace::immediate_execution(bool)
On 2014-05-27 18:06:06, Juergen Sauermann wrote:
Note that there are some subtle differences between IBM APL2 and GNU APL
when localizing ⎕-vars. In IBM APL2 they are undefined after localizing
them.
In GNU APL they are pre-initialized with their respective default values.
This
gives simpler
Hi,
Using index with axis on a vector results in assertion failure.
It works correctly with arrays of higher dimension or without axis.
1⌷[⎕IO]⍳1
==
Assertion failed: IX.value_count() != 1
in Function: index
On 2014-06-05 22:39:52, Elias Mårtenson wrote:
May I also ask that after reading the main config file, the interpreter also
reads $HOME/.gnu-apl.d for load user-level configuration. And finally, it
should also check the commandline so that the paths can be overridden on a
session-basis.
On 2014-06-10 14:08:47, Blake McBride wrote:
It is funny because I call this function a lot. Sometimes it works, other
times it gives me the domain error. Here are some more facts:
DOMAIN ERROR
Tmfmt[1] z←(2 0⍕⍉d[,1;]),':',2 0⍕⍉(d←(2⍴100)⊤,d)[,2;]
^ ^
On 2014-07-02 14:36:37, Juergen Sauermann wrote:
I have added monadic ⊣ and ⊢ see SVN 355.
Instead of making them identical I though it might be better to do
different things. What I came up with is this:
⊢ B returns (a copy of) B as a normal APL value.
⊣ B also returns a copy of B, but
On 2014-07-08 18:03:21, Juergen Sauermann wrote:
SOL_SOCKET is 1 on linux but according to the man page of TCP(7) SOL_TCP (=6)
should be used.
I have changed the code to use he number 6 instead of SOL_TCP. SVN 368.
Acctually, it should be IPPROTO_TCP as both mandated by POSIX and
described in
On 2014-07-14 00:01:13, Elias Mårtenson wrote:
Most modern (and some not-so-modern) languages have a standard way of
attaching
documentation to functions and other symbols.
(...)
I think the most natural way to describe a function is simply by
starting function *body* with a comment:
∇
Whoops, this thread was about the message rather than return value.
About that, standard states that event-message is local to a context and
that ⎕ES should create an exception in caller's context, but I don't
think it's properly specified what does it exactly mean.
But the point is, that ⎕ES
When either of arguments of = has imaginary part of magnitude greater
than comparison tolerance, a domain error is thrown.
⎕CT
1E¯13
0J1 = 0J1
DOMAIN ERROR
0J1=0J1
^ ^
There are no conditions under which = is allowed to throw a domain
error at all so this is serious.
On 2014-07-15 20:09:40, Juergen Sauermann wrote:
Except that (not so obviously) :
1E¯15 = 1E¯14
0
Sorry, my mistake. You're, of course, right.
Current (r376) behaviour seems to be correct with one exception:
1E¯15 = 1E¯14
0
1E¯15J0 = 1E¯14J0
1
Both should be zero,
On 2014-07-31 15:54:44, Peter Teeson wrote:
Looked in the APL2 IBM manual but do not understand how to determine the data
type of a variable.
Neither the primitives nor the Quads sparked the answer in my brain.
It must be something pretty obvious but not to me right now.
So if I have a
See below.
As a side note, you might want to look at warnings that lintian
gives for debian package even if most of them can be safely ingored.
-k
Index: workspaces/APL_CGI.apl
===
--- workspaces/APL_CGI.apl (revision 413)
+++
On 2014-08-13 19:42:25, Juergen Sauermann wrote:
I believe the VALENCE ERROR comes because = is dyadic and
the lambda is monadic?
In GNU APL (and in the ISO standard) there is no monadic =, is it?
Maybe in Dyalog APL there is?
No, it was supposed to be dyadic. I'm not sure how GNU APL
(⍬/X)←⍬ ⊣ X←⍬
SEGMENTATION FAULT
-- Stack trace at main.cc:140
0x7f4f289fbb45 __libc_start_main
0x43b725 main
0x556c75
{1=⍵} ,1
1
{1=⍵[1]} ,1
1
{⍵=1} ,1
1
{⍵[1]=1} ,1
==
Assertion failed: Avec::is_quad(idname[0])
in Function: get_nc
in file: NamedObject.cc:42
Call stack:
Currently GNU APL uses LCG with modulus 2⋆64 and then reduces values modulo
desired range. This, beside being slightly biased for ranges not dividing
the modulus, yields reduced periods when range is power of two.
?10 16⍴16
11 2 13 12 15 6 1 16 3 10 5 4 7 14 9 8
11 2 13 12 15 6 1 16 3 10
I don't think there is anything to gain by changing the main LCG part
given the constraints. I don't know how good are the parameters used by
GNU APL but they should be okay since those are the ones proposed by Knuth.
LCGs have a number of well known weaknesses, one of which is that their
less
Oh, I see you already fixed the bias exatly the way I wanted you to
in r440 even before I posted that :)
I still recommend using xorshift premutation instead of bit reversal
and iterating ⎕RL only once per output value.
-k
When used with negative shape as left argument and non-literal array
on the right, reshape dies with segmentation fault instead of raising
a DOMAIN ERROR.
¯1⍴0,0
SEGMENTATION FAULT
-- Stack
1↑⍤0 ⍬
SEGMENTATION FAULT
thread: 0x7fb6c6ec2740
thread_contexts_count: 1
busy_worker_count: 0
active_core_count: 1
thread # 0: 0x7fb6c6ec2740 pool sema: 0 RUN job: 0 no-name
--
On Sat, Apr 25, 2015 at 4:38 AM, Christian Robert
christian.rob...@polymtl.ca wrote:
a←1 0 0 0 0 1 0 0 1 1 0 0 1 1 1 1 0 1 1 0 0 0 0 0 1 0 0 0 0 0 0 0
b← 26 27 28
a[b]
0 0 0
b⊃a
DOMAIN ERROR
b⊃a
^ ^
I think this should be rank error rather than domain
On Fri, Apr 24, 2015 at 7:23 PM, Juergen Sauermann
juergen.sauerm...@t-online.de wrote:
Now, ATF files are what )IN and )OUT read resp, produce. Chances are that
those
ATF files are the same as for IBM APL2. In that case all you need is to
write them in Dyalog and read them back into GNU APL
I just checked out 592 and it failed to build because of missing buildtag.hh.
Apparently this is just a typo in configure: cd src source ./buildtag ; cd ..
Missing .
-k
On Sat, Jun 20, 2015 at 12:32 PM, Louis Chretien lchret...@mac.com wrote:
I agree the error message should not be as dire as a segfault, but the basic
intent seems dubious…
There's nothing inherently wrong with this. Argument of ⎕EX should
be a character array with rows containing identifiers
There is a number of problems with circle functions.
Inverse trigonometric and hyperbolic functions ¯1 ¯2 ¯6 ¯7
as well as function 0 are not properly extended to complex
values when given argument is real; they work correctly with
real-valued complex cells:
0 ¯1 ¯2 ¯6 ¯7 ∘.○ ¯2 ¯2J0
I have no idea how that happens.
¯1J0⋆2
DOMAIN ERROR
¯1⋆2
^ ^
-k
On Thu, Oct 22, 2015 at 3:29 PM, Michael Potter wrote:
> APL2000 does not appear to have an )OUT command.
>
> We get:
> "INCORRECT COMMAND" in response to:
> )OUT
> )OUT SOMEFILE
> )OUT C:SOMEFILE
I think in APL2000 it should be ]OUT starting with square bracket.
At least
On Fri, Oct 23, 2015 at 5:45 AM, wrote:
> Is there documentation on how to use ⎕TRACE ? I ran info apl | grep ⎕TRACE
> and found nothing. It does not look like it is listed in the IBM APL2 manual
> either.
⎕STOP and ⎕TRACE are defined in ISO 13751. They are part of
--- src/QuadFunction.cc (revision 686)
+++ src/QuadFunction.cc (working copy)
@@ -1604,7 +1604,7 @@
}
Function * fun = fun_symbol->get_function();
- if (fun_symbol == 0)
+ if (fun == 0)
{
CERR << "symbol " << fun_name_ucs << " is not a function" << endl;
On Fri, Oct 30, 2015 at 11:01 AM, Elias Mårtenson wrote:
>
> 4×+/{(¯1*⍵+1)÷¯1+2×⍵}¨⍳100
>
> The above expression performs 1 million iterations, but is also allocates a
> 1 million element array to do so.
I don't know GNU APL internals well enough to speak about when
On Fri, Oct 30, 2015 at 2:39 PM, Kacper Gutowski <mwgam...@gmail.com> wrote:
> i←0 ◊ 4×{⍵+(¯1*i+1)÷¯1+2×i←i+1}⍣100⊢0
This is slightly faster and doesn't use global variables:
↑{(+/⍵[1],4÷1↓⍵),4 ¯4+1↓⍵}⍣50⊢0 1 ¯3
Nowhere near “simple and natural,” though.
-k
On Mon, Oct 12, 2015 at 5:48 AM, wrote:
> Why do you choose to write it like this?:
> ContentsOfPOST←{⍵, FIO∆fread 0}⍣{⍺⊢FIO∆feof 0}''
>
> This method (in my limited knowledge) seems equivalent:
> ContentsOfPOST←(FIO∆feof 0) FIO∆fread 0
I don't think it is.
FIO∆fread
On Sat, Oct 10, 2015 at 3:29 AM, wrote:
> I am struggling to follow our previous conversation. Pretty much, I am
> confused about the source of the POST contents. I believe you mentioned that
> it should come from stdin, and if I understood, it is not possible to use
On Sun, Aug 30, 2015 at 12:02 AM, Alexey Veretennikov
alexey.veretenni...@gmail.com wrote:
Thanks, it looks indeed like I have to specify all sizes in GNU APL.
In Dyalog you can use Dyalog's extensions to the standard but in GNU
APL, you can use GNU APL's extensions. You don't need to specify
On Sun, Sep 13, 2015 at 11:53 PM, Louis de Forcrand wrote:
> Is it possible to terminate a function while it's running
> (eg. when you accidentally set a very large number
> as an operand)?
Pressing Control-C should signal ATTENTION which should suspend
execution of current
I came up with this silly variant of APL-in-shell embedding:
#!/bin/sh
true ←⎕inp'#'
exec apl --script "$0" -- ${1+"$@"} #
)erase true
But I think there's something wrong with ⎕INP, actually.
In the example above, true is ⍬ (sic) just before erasing.
]BOXING 8
⎕INP "E"
AE
On Fri, Jan 15, 2016 at 4:29 AM, wrote:
> I was wondering if anyone is familiar enough with xmodmap to know why it
> would disable the existing Alt-Tab functionality in xorg (switching between
> windows) when the key combination has no apparent use in the interpreter.
On Mon, Feb 15, 2016 at 12:08 AM, wrote:
> the system's word count mechanism says length four. APL says length three, I
> thought this was going to be length 1 and the symbol '⍝'. Does anyone know
> what I am doing incorrectly:
>
> a@a:~/aplstuff$ echo "⍝" > txt
>
On Sun, Feb 21, 2016 at 10:38 PM, wrote:
> it seems that format adds fuzz after a certain point.
It's not a “fuzz,” but an artifact of using IEEE 754 double precision
floating point numbers. ISO APL leaves a lot of freedom when it comes to
representation of numbers,
On Sun, Mar 13, 2016 at 6:20 PM, Juergen Sauermann
wrote:
> it actually does create conflicts.
>
> In IBM APL2 and in GNU APL, the expression
>
> ⍺ (f g h) ⍵
>
> gives a 3 item vector with the items being ⍺, (f g h), and ⍵.
> In Dyalog APL it gives (quote):
>
> (⍺ f
On Mon, Mar 14, 2016 at 2:03 PM, Juergen Sauermann wrote:
> Consider this:
>
> 1 (+//) 1 2 3 4
> 1 2 3 4
FWIW, Dyalog returns the same, but it certainly does something weird
I don't understand here. It treats it as an operator even when it
says it is a function:
f←/
⎕NC'f'
3
Logarithm of negative real number fails to coerce it to complex and
results in a malformed value:
2⍟¯1
0.0
2⍟¯1J0
0J4.532360142
-k
Dyadic ⎕CR with 5 or 6 as its left argument and character scalar
instead of vector as right causes segmentation fault:
5⎕CR '1'
SEGMENTATION FAULT
-- Stack trace at main.cc:63
If I understand correctly, Elias wants to have something that would
ensure that every file opened is eventually closed (only more general so
it can be applied to any external resources not only files). That is in
a code similar to the following, to ensure the line 3 is always run even
if there's
Hi,
Currently GNU APL doesn't know what ⎕SVQ is because it
became ⎕SQQ due to typo back in r625.
⎕SVQ''
VALUE ERROR
⎕ SVQ ''
Amusingly:
⎕SQQ
SYNTAX ERROR
⎕SVQ
^
-k
On Mon, Apr 11, 2016 at 5:27 AM, Elias Mårtenson wrote:
> This seems similar to what GNU APL provides with the ⎕EA feature. The
> problem with this (and, as far as I can tell, also with this Dyalog feature)
> is that in an error handler, the original error has disappeared.
If you call some
I recall there being some similar attempts presented on this list but
I couldn't find working example so I hacked up the following. It has
few rough edges and wasn't tested thoroughly but appears to work.
]USERCMD ]EDIT EDIT⍙ 1
∇S EDIT⍙ Q;FN;FD
⍝ Edit a function or operator in external
Giving malformed string to 19⎕CR results in it printing an error
message and a stack trace, and then returning an empty vector. As
it's generally very useful to have generic routine parsing UTF-8
data at hand, I think it shouldn't do that. If it's an error, it
should be possible to handle it
⎕EC¨ returns the expected 3-element vector only for the first element
of ravel list of its argument and plain values for others. If they
return no values, it fails; if they err, the interpreter hangs.
]boxing 7
{⎕EC⍵}¨'12'
┌→──┐ ┌→──┐
│1 ┌→──┐ ┌─┐│ │1 ┌→──┐ ┌─┐│
│ │0
Looks like “Bad number” is a different kind of error than other
syntax errors. If, say, lone ¯ is evaluated in ⎕EC¨ last then
it seems to abort the whole line ⎕EC¨ appears in.
I'm at r712.
⎕EC¨'1¯' ⍝ no result at all
⎕FX'f x' '⊣ ⎕EC¨x'
f
f'1¯' ⍝ unexpected output
1 0 0 1
>
> On 03/20/2016 01:29 AM, Kacper Gutowski wrote:
>>
>> ⎕EC¨ returns the expected 3-element vector only for the first element
>> of ravel list of its argument and plain values for others. If they
>> return no values, it fails; if they err, the inter
On Thu, Mar 3, 2016 at 12:40 PM, Juergen Sauermann
wrote:
> the IBM APL2 language reference says (page 111):
>
> If R is a simple scalar, ⊂ R is R. If R is not a simple scalar, the depth of
> R is 1+ ≡R.
>
> And the ISO standard says the same (page 169):
>
> Z ← ⊂B
On Sat, Mar 5, 2016 at 6:45 AM, Christian Robert
wrote:
> [0] z←Execute cmd;⎕io;fh
> [7] Fini: ⎕FIO[25] fh
> [8] ⊃⊃z
(...)
> I do not understand why it repeat the result twice. one via stdout and one
> via boxing as I can guess.
The function returns z as
On Wed, Apr 27, 2016 at 5:02 AM, Elias Mårtenson wrote:
> So, I wanted to simply take a string of characters, and create a list of
> each character and the number of occurrences of each character in that
> string.
Indeed it could be written in Dyalog very succinctly as
⊣,∘≢⌸ but I would write it
On Thu, May 19, 2016 at 12:37 AM, Xiao-Yong Jin wrote:
> Now,
> (22⍴8)⊤2⊥63⍴1
> 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
>
> Dyalog does the same thing.
> Can someone please explain to me why?
Because the result of ⊥ doesn't fit into integer and it gets converted to a
floating point
On Tue, May 17, 2016 at 4:49 PM, Juergen Sauermann wrote:
> is seems like IBM APL2 computes
> 1 2 3 /¨ 4 5 6
> as:
> (1 2 3/4) (1 2 3/5) (1 2 3/6)
> However, the ISO standard says
> Z ← A f¨ B
(...)
The difference isn't in each operator, but in parsing / always as a monadic
operator
On Wed, May 18, 2016 at 5:09 PM, Xiao-Yong Jin wrote:
(...)
> relying on ¯1=2⊥64⍴1 (always a 64-bit signed integer), for which Dyalog
> does differently.
I think this is a bug.
-k
This assertion can be still made false in r725 with:
+⍤1 ⍬
==
Assertion failed: 0
in Function: init
in file: Cell.cc:48
Call stack:
-- Stack trace at
On Thu, Apr 14, 2016 at 1:54 AM, Xiao-Yong Jin wrote:
> Exactly so. ‘-Ofast -fno-finite-math-only’ gives correct results.
> What’s going on here? I don’t expect this has anything to do with NaN or Inf.
pow(-8., 1./3) is nan.
And -ffinite-math-only eliminates isfinite check at FloatCell.cc:422.
⊂⍤¯1¨0
result: 'SI_PUSHED' at Bif_OPER1_EACH.cc:257
-- Stack trace at Bif_OPER1_EACH.cc:257
0x7fa8af92d5f0 __libc_start_main
0x444ae5 main
0x58069d Workspace::immediate_execution(bool)
0x48a852
According to ISO, 0⍟0 should be one.
GNU APL gives:
0⍟0
DOMAIN ERROR
0⍟0
^^
-k
On 29 July 2016 at 16:08, Juergen Sauermann wrote:
> unfortunately it is not that easy. make does not pass its -j option to
> sub-makes.
I don't know if it works as documented, I never checked it, but according
to GNU make manual, it should coordinate between sub-makes to run the
number of tasks
I like the second of the proposed solutions. Since ambivalent function
can always be called monadically, I think that making the header of
d-fns always λ←⍺ λ1 ⍵ regardless of whether ⍺ occurs in the body or not,
shouldn't break anything and would only make a language a bit simpler.
Although maybe
I just got this at r775:
{'⍝'}
copying: '1' at Executable.cc:655
tidx:'1' at Executable.cc:655
-- Stack trace at Executable.cc:655
0x7f055ac095f0 __libc_start_main
0x444ae5 main
On 13 July 2016 at 13:21, Juergen Sauermann wrote:
> I see. Which leaves the question if "equal" shall be strict or within ⎕IO.
>
> Since we are dealing with real numbers (and therefore often rounding errors)
> within ⎕IO
> makes more sense to me but the standard does not mention ⎕IO for ⍟.
I
On 13 July 2016 at 12:04, Juergen Sauermann wrote:
> my ISO (June 9, 2000) says DOMAIN ERROR.
I'm using the PDF that was linked from GNU APL documentation,
that is, exactly the same Jun 2000 version.
> If A and B are equal, return one.
As Jay just have written, the evaluation sequence in section
I commented out a line of code and suddenly GNU APL exited when FX-ing it.
Looks like it hit some not-implemented-yet part, but could you please make
FIXME exit with non-zero exit code?
⎕FX 'F' '⍝}' '{{}}'
copying: '0' at Executable.cc:655
tidx:'2' at
On Wed, Aug 03, 2016 at 07:09:38PM -0500, Blake McBride wrote:
> APL with AKT works great. But now I need to use it in an xterm. Doesn't
> seem to work there. Any suggestions would be greatly appreciated.
You need to set metaSendsEscape. It can be configured in X resource
or switched in run
On 3 February 2017 at 20:31, Alexey Veretennikov wrote:
> How can I silence this output when I execute the )copy command from the
> script?
You can do
⊣⍎')COPY 5 FILE_IO FIO∆errno'
-k
On 25 January 2017 at 05:54, Elias Mårtenson wrote:
> Is there anything inherently wrong with having more quad-commands? What is
> the reason we have to do ⎕FIO[49] instead of simply ⎕ReadLines or something
> like that?
>
> It's like we're going out of way to make code unreadable. I know APL is
The sqlite provider seems to silently truncate integer values in results:
)COPY 5 SQL
DUMPED 2017-02-12 10:13:53 (GMT+1)
db←'sqlite' SQL∆Connect ':memory:'
'create table a(b)' SQL∆Exec[db] ⍬
'insert into a values(?),(?)' SQL∆Exec[db] 2⋆31 32
'select b from a'
On 14 February 2017 at 14:31, Elias Mårtenson wrote:
> I've taken a look at this now, and I believe this patch should fix your
> problem. Please try it, and I'd like to ask Jürgen to merge it.
Yes, this fixes it.
Thanks.
-k
On 15 February 2017 at 03:59, Frederick Pitts wrote:
> But the problem does seem specific to my platform.
I can reproduce it under debian testing. It appears that strings
longer than 12 are somehow misinterpreted by AP100 and garbage is
sent to popen. It doesn't seem to have anything to do with
On 9 September 2016 at 23:39, Ala'a Mohammad wrote:
> the errors happened inside 'hist' function, and I presume mostly due
> to the jot dot find (if understand correctly, operating on a matrix of
> length equal to : unique-length * words-length)
Try (∪⍵)∘.≡⍵ instead of ∨/¨(∪⍵)∘.⍷⍵.
-k
On 28 August 2016 at 04:12, enztec wrote:
> works here
Perhaps it would fail if you tried bigger array?
9 seemed enough to reproduce it reliably in my environment.
And no, this is a complete transcript, nothing else was defined.
-k
Hi,
Matrix division returns no value for large-shaped arguments.
There doesn't seem to be any fixed threshold above which it fails,
it differs from case to case.
A←?9 3⍴256
X←A⌹A
X
VALUE ERROR
X
^
-k
On 28 August 2016 at 14:02, Juergen Sauermann wrote:
> I believe I managed to reduce the amount of memory needed for A⌹B
> considerably. SVN 791.
Great!
Thanks.
-k
Actually, disregard that.
It's just:
'0006-06-06@06:06:06.000'⍕⎕TS
2017-01-12@10:42:39.088
Formatting dates is literally what was used as an example
of using formatting code 6 in apl2lrm.
-k
On Tue, Dec 20, 2016 at 10:30:14AM +0100, Alexey Veretennikov wrote:
> Using )HOST Vim complains "Vim: Warning: Output is not to a terminal"
> and not really useful (can't see what I type for instance).
> Same if I use popen() via ⎕FIO.
)HOST and popen are essentially the same thing and they both
On Sat, Dec 24, 2016 at 12:14:41AM +0100, Alexey Veretennikov wrote:
> I'm running GNU APL in the X terminal with geometry 48x15 (it is a small
> screen). How could I set this width(48) to the GNU APL interpreter so it
> will behave correctly on deletion of characters in long lines etc?
Setting
On 26 December 2016 at 20:09, Juergen Sauermann wrote:
> Kacper: you can now enable change of ⎕PW via the WINCH signal. This may not
> work on all platforms, therefore is has to be enabled in one of your
> preferences files.
Works great, thanks.
It's okay that it requires enabling, it's not
Hello,
There is another case of not handling near-real numbers properly:
⍋,1J0
DOMAIN ERROR
⍋,1
^^
It should of course return ,1.
-k
Hello,
Binomial returns malformed result when called with some non-integer arguments:
0!200.1
0.0
-k
On Fri, Dec 16, 2016 at 11:43:19AM +0100, Nick Lobachevsky wrote:
> Wouldn't this really be a domain error, the maximum argument value for
> ! being around 170?
>
> !170
> 7.257415615307994E306
> !171
> DOMAIN ERROR
I'm using dyadic ! here so it should return 1 nevertheless; it does
1 - 100 of 159 matches
Mail list logo