« first day (1392 days earlier)      last day (100 days later) » 

3:38 AM
@Adám the jpg curse strikes again
I guess i should just copy the do-while over
 
 
2 hours later…
5:27 AM
I may have found a bug in 18.0.38756 ... Try it online!
This code gives the expected result on TIO and tryapl, but on my local interpreter with the above version the top row of the result matrix is all zeroes
Weirdly enough if I modify the leftmost function to print the results like {⎕←⊖⍉⍵,+\{+/⍵}⌺3⊢/⍵} then I actually do get the expected answer
Maybe I'm missing something?
 
 
1 hour later…
6:35 AM
@voidhawk Wow, that's a nasty one. Thank you so much for reporting it. I can reproduce it, even with a much cut-down expression. Logged as issue 18649.
 
No worries! Very curious what's going on there
 
It seems fails on some unsqueezed values.
 
7:34 AM
More stencil weirdness: ⊂⌺3⍳10 gives me a very strange error on 18.0.38756, wheras TIO and tryapl just give me LENGTH ERROR
LIMIT ERROR
{⍺⍺{⍺ ⍺⍺ ⍵}{⎕IO←0 ⋄ ⎕ML←1 ⋄ (s m)←↓(⍉1,⍨⍪)⍣(2>≢⍴⍵⍵)⊢⍵⍵ ⋄ ws←(≢s)↑⍴⍵ ⋄ n←⌈m÷⍨ws-~2|s ⋄ j←⊃∘.,/⊂⍤0¨↓¨(m×⍳¨n)∘.+¨⍳¨s ⋄
c←⌊2÷⍨s-1 ⋄ y←(ws+2×c)↑(-ws+c)↑⍵ ⋄ d←0⌈((m×n-1)+s-1)-ws+c-1 ⋄ p←0~⍨¨0⌈c-m×⍳¨c ⋄ q←0~⍨¨0⌊d-⍨⌽¨m×⍳¨d ⋄ i←,¨⍣(1=≢⍴⍵)⊃∘.

,/p,¨((n-(≢¨p)+≢¨q)⍴¨0),¨q ⋄ i ⍺⍺{(⊃⍺)⍺⍺(⊃⍵)⌷y}⍤0⊢j ⋄ }⍵⍵⊢⍵}
 
<moon-child> @voidhawk I think dyalog should hire you to do QA ;)
 
@voidhawk Two issues here. Magic operator exposing its internal code. That has been fixed. The other is the LIMIT ERROR which should be a DOMAIN ERROR. That was logged by me as issue 18487 two months ago.
 
8:09 AM
@voidhawk "Explanation" (I don't understand) from Roger Hui:
> The bug is in assuming that the argument type (APLINTG) is always "larger" than the result type (APLSINT).
 
8:29 AM
<moon-child> Maybe APLINTG is a (g)eneric-sized int (so, might be any size; we don't know yet) and SINT is just a regular 32-bit int? Just a guess
 
8:57 AM
I'm sure I've had that error before as well, I remember getting an error with a bunch of code when I first started
 
9:30 AM
It used to manifest itself for a lot of operators.
 
@Adám this effectively summarizes what I feel most of the time lol
 
Do we know everyone / have links to all blog posts of this type:mathspp.com/blog/2020-apl-competition ?
 
 
1 hour later…
RGS
Where is Jared Semerad from? (company)
I didn't want to click the link to his linkedin profile because it then let's them know I was there, but nvm. I can find out for myself, I just didn't want them thinking I was interested. I'm just curious :P
 
Can people see when you view their linkedin profiles?
 
RGS
@rak1507 In general, yes.
 
Oh well, I guess several people will think I'm a stalker now
 
RGS
If someone views my LinkedIn profile, the least information I get is that someone viewed it, and usually it tells me exactly who looked at it. It depends on if the person that viewed it is already a connection of yours or not, and if they have LinkedIn Premium or not (smth like that, I think)
 
 
2 hours later…
12:28 PM
@Adám When applying
Experience: 1 month
 
You're applying? cool
 
12:48 PM
no obviously not
 
Oh
 
need to be a lot more qualified lol
 
Yeah fair enough
 
1:21 PM
@DyalogAPL moon-child: INTG is 16 bits and SINT is 8 (LONG is 32).
 
1:33 PM
@RikedyP Are you still working on the APL topanswers? Just wondering where it got to.
 
RGS
2:12 PM
@Razetime was that on the job description?
 
2:28 PM
actually, not sure
but I wouldn't cosider myself fit for an APL job yet
 
RGS
ofc it is all very relative
 
@Razetime Though you just did one!
 
RGS
But usually I'd leave those decisions to the ones hiring
I also have very limited APL experience and I got an APL-related internship :P
 
2:47 PM
@Adám nuh-uh the jpeg part is still not done
 
@Marshall Yeah it's coming, just a few final touches - on that note got any must-have post tags?
 
@RikedyP BOOLEAN
 
@Marshall As in a post about booleans?
@Marshall Or for yes/no polls?
 
I guess it would be nice to have a way to separate posts about how to do something in a specific language from what an array approach to a problem would be. Language-specific tags work for the first, but maybe something like "array thinking" or "array techniques" for the second (I'm sure there's a better name).
@RikedyP The boolean thing is just a joke, sadly the rest of the world is probably not interested enough in booleans to make a whole tag out of it.
Actually everything in Category:Paradigms is a reasonable thing to consider for a tag.
 
 
2 hours later…
4:36 PM
 
4:49 PM
@voidhawk Correction:
> The bug is in assuming that the argument type (APLINTG) is always "smaller" than the result type (APLSINT).
\○/
 
it's patched?
 
Not yet. Roger is working on it.
 
Ah RojerGS
 
 
Roger that
 
5:06 PM
Exactly.
 
RGS
5:17 PM
XD
 
 
3 hours later…
8:03 PM
@voidhawk Fixed in 19.0 for now. I expect the fix backported to 18.0 shortly.
 
@Adám Awesome, thanks!
 

« first day (1392 days earlier)      last day (100 days later) »