Re: [racket-users] Function that potentially can return more than one value
On 5/27/2020 1:50 PM, Jens Axel Søgaard wrote: Den ons. 27. maj 2020 kl. 19.27 skrev George Neuner mailto:gneun...@comcast.net>>: On 5/27/2020 6:34 AM, Jens Axel Søgaard wrote: > > In standard Racket an application doesn't communicate how many values > a function is expected to produce. You can ask: see procedure-result-arity https://docs.racket-lang.org/reference/procedures.html I think procedure-result-arity answers the question "how many values could this procedure return?" and not "how many values is the procedure expected to return in this context?". Oh, I get it ... it's not that you "don't know" what's expected, it's that you've got something that returns N values on some path and M values on some other path, etc. and you can't predict which path will be followed. Yeah, that is a problem: e.g., I don't think procedure-result-arity works for case-lambda where different invocations can return different numbers of results. In that case, better to return a container: a list, vector, hash, etc. rather than a bunch of individual values. Or, if the function is canned [e..g, in a library] to put them into a container for your code to work with. > I have written a small proof-of-concept of an assignment operator := > that communicates to a function how many extra values it is expected > to produce (and still works with standard functions). > > Are there alternative solutions that are better than the "use a > keyword" approach? You can use call-with-values which connects the function that produces with the one that consumes, but they have to agree on the number of values being passed or it won't work. My problem is that the producer still needs to be told how many values to produce. Sorry, I missed that - I thought you needed to know how many were produced by something. The only way to call attention - particularly compiler attention - to a particular argument is to make it a keyword with no default value. If your function has default behavior that the keyword(s) would only modify, then you can make keyword(s) optional using default values or by defining the function with make-keyword-procedure. George -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/racket-users/5bea165b-3357-7236-173c-934ca3a0ab58%40comcast.net.
Re: [racket-users] Function that potentially can return more than one value
Den ons. 27. maj 2020 kl. 19.27 skrev George Neuner : > > On 5/27/2020 6:34 AM, Jens Axel Søgaard wrote: > > > > In standard Racket an application doesn't communicate how many values > > a function is expected to produce. > > You can ask: see procedure-result-arity > https://docs.racket-lang.org/reference/procedures.html I think procedure-result-arity answers the question "how many values could this procedure return?" and not "how many values is the procedure expected to return in this context?". > > I have written a small proof-of-concept of an assignment operator := > > that communicates to a function how many extra values it is expected > > to produce (and still works with standard functions). > > > > Are there alternative solutions that are better than the "use a > > keyword" approach? > > You can use call-with-values which connects the function that produces > with the one that consumes, but they have to agree on the number of > values being passed or it won't work. > My problem is that the producer still needs to be told how many values to produce. /Jens Axel -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/racket-users/CABefVgwy4YEPcswLjLEf-FN2sZMa2xbPqOWxOJ%3DdOEnEUF6_Hg%40mail.gmail.com.
Re: [racket-users] Function that potentially can return more than one value
On 5/27/2020 6:34 AM, Jens Axel Søgaard wrote: In standard Racket an application doesn't communicate how many values a function is expected to produce. You can ask: see procedure-result-arity https://docs.racket-lang.org/reference/procedures.html I have written a small proof-of-concept of an assignment operator := that communicates to a function how many extra values it is expected to produce (and still works with standard functions). Are there alternative solutions that are better than the "use a keyword" approach? You can use call-with-values which connects the function that produces with the one that consumes, but they have to agree on the number of values being passed or it won't work. You can use a Rest argument in the consumer if the values from the producer are put into a list. Unfortunately, Scheme did not embrace the idea of using tuples for return values, nor did it allow to catch only the first N values as was permitted by Lisp. George -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/racket-users/fefd28d4-bab8-11bc-0a29-0983758cd454%40comcast.net.
