« first day (941 days earlier)      last day (1700 days later) » 

dzaima
dzaima
8:32 AM
currently is parses (with an unknown level of correctness) 165/207 of the dfns functions (~10 of which i modified because dfns likes to define variables in reverse order, understanding which statically is annoying), and gives out this list of 25 places where `⍥` could be used:
https://tio.run/##pVZNb9tGEL3rV2xPIW1KilTHARjUQD@QtEAbIEiA9lAIWJFLiRXFpXeXtGzXOTSFIqtW0bQo2kvRtOhBtx4SI4ABX@x/wj/izixJl5YlWWkJATu7/Hhv37yZlUPDhMqLCyql3wlJOvyRSMWiTR3thWyg7AopLj1Phz/sMcHlunH@i8OlkmYVArxz@Vw6@Y3QPgtd5hL9LOlT0WOiRspXhSy7WpUZShslIk8b6WjMIxbaRno0rhtWn8qeWbcyPvkGJsc5lZCTOHR4wkTBxyKaLz5I7M3aUiqtStsXquvS3ZxAo…
also because why not, here's 56 places where each-right or each-left could be used:
https://tio.run/##5VnPb@PGFb7rr5ibJFiWJWcRBEKcHoqiLZCemp4Wa2BIjixG/BXOSJbheA9u4XqdVdBNG2wuBdpc4gQF9rA1XAjoRfuf8B/Zfm@GpEhZoumkeygqLLwcUeL3vTfvfe@9kc2DKZdv33IvCiOWXHzFPFeqQYPRiy6Tiz@1kqtzN3BmybO/JfPXyfyfY3Ei23vL62S@YCyZ/5XFwpnYwtFfYFtejW032GGDKxW7loaXs6Gf4dPrbHndY322z/rJ1e8BeXWezG@2PIioyJNA8RkLh2w4CWzlhkG3FpXDhsWleP9RCm1zJZvN5PKbYTiJl9f7j1jEY8U@YJarJNtOobb1hw2be9pi2CtTVLqEw/dZD8jJ/Ovldcv3D3wf3nYcudVme8Tj3WEY@/AjdsHhJyyY@JaIZXcDNJD9aKahhRR2wde0BHxy8eJ9tny1n3EIhFJvbu1wEqg3L/u9Xm8dX6TgDN8PA0d279/3AokoFsMCCVqCxCl…
 
 
12 hours later…
Adám
Adám
8:44 PM
@dzaima Cool, we should keep this in mind to update the collection upon release of 18.0.
@J.Sallé We are planning to use ⍶ and ⍹ in the syntax specification of magic arrays.
@user113586 Hello Jimmy. Welcome to The APL Orchard!
 
Adám
Adám
9:17 PM
@dzaima Actually, Marshall is speaking about ⍥ (and ⍤ and ⍨) at D'19. Can I point him to this?
 
dzaima
dzaima
@Adám him pointing to it was the reason I made that, so definitely
sadly searching for use-cases of ⍤ is way more difficult, as using it in place of f A g B is stupid and trains are very complex, and it's even worse with ⍢ as that needs actually understanding the language, which that parser definitely doesn't
 
Adám
Adám
@dzaima OK, sent, but we value your opinion, so feel free to submit your input to the email thread :-)
 
dzaima
dzaima
speaking of ⍨, how much sense does it make to space (⊢*2⍨) as (⊢*2 ⍨) to indicate that the ⍨ is more a part of the train definition?
 
Adám
Adám
@dzaima You might be able to identify (f g).
 
dzaima
dzaima
@Adám yeah, that makes sense, but again that's too easy (but actually probably still worth checking out)
 
Adám
Adám
9:25 PM
@dzaima Interesting. I guess it makes (…⍨) kind of a meta syntax for trains, even though it syntactically is just a simple derived function.
@dzaima How about searching for constant f⍨ g?
 
dzaima
dzaima
@dzaima hmm this being the only result definitely doesn't feel right..
@Adám that doesn't seem to return anything at all. have i completely messed up my train parsing somehow (i did check that it does work on simple examples) or does dfns just not use these things at all?
 
Adám
Adám
@dzaima Possibly, they are not dfns, after all, and being that there is no good way to store trains, it would only be inside dfns, at which point the explicit form is the obvious choice.
 
dzaima
dzaima
oh, f←++ is interpreted as (f←+)+, i recall that exact issue being in dzaima/APL :|
 
Adám
Adám
9:42 PM
@dzaima These are the only parenthesised trains of simple primitives I find: 
(+⊣)
(-⊢)
(⊃⌽)
(⊃⌽)
(+?)
(⍎?)
(??)
(∪⍳⊢)
(+⊣)
(-⊢)
And some are them are obviously not true hits.
Afaict, name←train never occurs.
 
dzaima
dzaima
@Adám it does, for a lazy definition of a train
 
Adám
Adám
@dzaima Right, I was only searching pure primitives. Did you notice I←⊃¨∘⊂? i.e. ⊇
 
dzaima
dzaima
9:59 PM
yeah, a dfns workspace doesn't seem to be the best place to search for trains..
 
Adám
Adám
@dzaima Aaron Hsu's compiler source…
 
dzaima
dzaima
@Adám good idea, but i'm afraid of the long long list of problems that might arise from it
 
dzaima
dzaima
10:19 PM
oh, i completely messed up updating ← - A f← B made As assumed type be function
 
Adám
Adám
oops
 
dzaima
dzaima
first couple files of the compiler use namespaces, which my parser doesn't handle at all, due to their dynamic nature
and d.cd has out-of-order function creation, which i could fix by rearranging them manually, so e.cd is the only file that parses (but nothing of interest is found)
 

« first day (941 days earlier)      last day (1700 days later) »