Uppercasing a string with maplist/3 and char_type/2

[razetime]

I am not sure why i get a couple of variables in list D here. Only the last character is converted.

?-  maplist(\X^Y^char_type(X,to_upper(Y)),"AsdaSWQEf",D).
   D = [_A,_B,_C,_D,_E,_F,_G,_H,"F"].

I try to tell prolog that it is an alphabetical character beforehand:

?-  maplist(\X^Y^(char_type(X,alpha),char_type(X,to_upper(Y))),"AsdaSWQEf",D).
   D = [_A,_B,_C,_D,_E,_F,_G,_H,"F"].
?-  maplist(\X^Y^(char_type(X,alphabetic),char_type(X,to_upper(Y))),"AsdaSWQEf",D).
   D = [_A,_B,_C,_D,_E,_F,_G,_H,"F"].

How do I correct my code?

[UWN]

There are a couple of problems, none of which is related to maplist/3 or lambda.

?- char_type(a,to_upper(A)), char_type(b,to_upper(B)).
   B = "B",  unexpected.

This is unexpected for a couple of reasons. First, there is no answer substitution for A. This is clearly an internal bug somewhere in Scryer.

But then, the answer substitution itself is odd, it should be exactly the other way around, at least in SWI which itself has a problem SWI-Prolog/issues#130.

Also, I do not understand the point of producing here a string. The documentation mentions this, but no, I do not get it.

[UWN]

That is the German ß, which currently is often written as SS in uppercase letters. And then, sometimes it is written as SZ. In any case, there is now an uppercase sharp s too. But that is not my point, it's rather the question whether or not it makes sense to provide that functionality under a name which originally (in SWI) has different functionality. Adding this with another name would make sense indeed. Also the to_ logic has been reversed.

[razetime]

shall i transfer this to an issue then?

[aarroyoc]

Yes, it returns a string instead of a character because Unicode (not just German in fact). Rust API returns a string too, so we keep that semantic.

About the naming. I agree it is confusing. I found it confusing in SWI too ( I never understood where the to refers). I think we can keep the same semantics but renaming to_upper to just upper?

[triska]

I also found it confusing in SWI too! upper or upper_is seem better!

[triska]

Uppercase variants of length at least 2:

?- length(_, L), char_code(C, L), char_type(C, to_upper(Us)), Us = [_,_|_].
   L = 223, C = ß, Us = "SS"
;  L = 329, C = ʼn, Us = "ʼN"
;  L = 496, C = ǰ, Us = "J̌"
;  L = 912, C = ΐ, Us = "Ϊ́"
;  L = 944, C = ΰ, Us = "Ϋ́"
;  L = 1415, C = և, Us = "ԵՒ"
;  L = 7830, C = ẖ, Us = "H̱"
;  L = 7831, C = ẗ, Us = "T̈"
;  L = 7832, C = ẘ, Us = "W̊"
;  L = 7833, C = ẙ, Us = "Y̊"
;  L = 7834, C = ẚ, Us = "Aʾ"
;  L = 8016, C = ὐ, Us = "Υ̓"
;  L = 8018, C = ὒ, Us = "Υ̓̀"
;  L = 8020, C = ὔ, Us = "Υ̓́"
;  L = 8022, C = ὖ, Us = "Υ̓͂"
;  L = 8064, C = ᾀ, Us = "ἈΙ"
;  L = 8065, C = ᾁ, Us = "ἉΙ"
;  L = 8066, C = ᾂ, Us = "ἊΙ"
;  L = 8067, C = ᾃ, Us = "ἋΙ"
;  L = 8068, C = ᾄ, Us = "ἌΙ"
;  L = 8069, C = ᾅ, Us = "ἍΙ"
;  L = 8070, C = ᾆ, Us = "ἎΙ"
;  L = 8071, C = ᾇ, Us = "ἏΙ"
;  L = 8072, C = ᾈ, Us = "ἈΙ"
;  L = 8073, C = ᾉ, Us = "ἉΙ"
;  L = 8074, C = ᾊ, Us = "ἊΙ"
;  L = 8075, C = ᾋ, Us = "ἋΙ"
;  L = 8076, C = ᾌ, Us = "ἌΙ"
;  L = 8077, C = ᾍ, Us = "ἍΙ"
;  L = 8078, C = ᾎ, Us = "ἎΙ"
;  L = 8079, C = ᾏ, Us = "ἏΙ"
;  L = 8080, C = ᾐ, Us = "ἨΙ"
;  L = 8081, C = ᾑ, Us = "ἩΙ"
;  L = 8082, C = ᾒ, Us = "ἪΙ"
;  L = 8083, C = ᾓ, Us = "ἫΙ"
;  L = 8084, C = ᾔ, Us = "ἬΙ"
;  L = 8085, C = ᾕ, Us = "ἭΙ"
;  L = 8086, C = ᾖ, Us = "ἮΙ"
;  L = 8087, C = ᾗ, Us = "ἯΙ"
;  L = 8088, C = ᾘ, Us = "ἨΙ"
;  L = 8089, C = ᾙ, Us = "ἩΙ"
;  L = 8090, C = ᾚ, Us = "ἪΙ"
;  L = 8091, C = ᾛ, Us = "ἫΙ"
;  L = 8092, C = ᾜ, Us = "ἬΙ"
;  L = 8093, C = ᾝ, Us = "ἭΙ"
;  L = 8094, C = ᾞ, Us = "ἮΙ"
;  L = 8095, C = ᾟ, Us = "ἯΙ"
;  L = 8096, C = ᾠ, Us = "ὨΙ"
;  L = 8097, C = ᾡ, Us = "ὩΙ"
;  L = 8098, C = ᾢ, Us = "ὪΙ"
;  L = 8099, C = ᾣ, Us = "ὫΙ"
;  L = 8100, C = ᾤ, Us = "ὬΙ"
;  L = 8101, C = ᾥ, Us = "ὭΙ"
;  L = 8102, C = ᾦ, Us = "ὮΙ"
;  L = 8103, C = ᾧ, Us = "ὯΙ"
;  L = 8104, C = ᾨ, Us = "ὨΙ"
;  L = 8105, C = ᾩ, Us = "ὩΙ"
;  L = 8106, C = ᾪ, Us = "ὪΙ"
;  L = 8107, C = ᾫ, Us = "ὫΙ"
;  L = 8108, C = ᾬ, Us = "ὬΙ"
;  L = 8109, C = ᾭ, Us = "ὭΙ"
;  L = 8110, C = ᾮ, Us = "ὮΙ"
;  L = 8111, C = ᾯ, Us = "ὯΙ"
;  L = 8114, C = ᾲ, Us = "ᾺΙ"
;  L = 8115, C = ᾳ, Us = "ΑΙ"
;  L = 8116, C = ᾴ, Us = "ΆΙ"
;  L = 8118, C = ᾶ, Us = "Α͂"
;  L = 8119, C = ᾷ, Us = "Α͂Ι"
;  L = 8124, C = ᾼ, Us = "ΑΙ"
;  L = 8130, C = ῂ, Us = "ῊΙ"
;  L = 8131, C = ῃ, Us = "ΗΙ"
;  L = 8132, C = ῄ, Us = "ΉΙ"
;  L = 8134, C = ῆ, Us = "Η͂"
;  L = 8135, C = ῇ, Us = "Η͂Ι"
;  L = 8140, C = ῌ, Us = "ΗΙ"
;  L = 8146, C = ῒ, Us = "Ϊ̀"
;  L = 8147, C = ΐ, Us = "Ϊ́"
;  L = 8150, C = ῖ, Us = "Ι͂"
;  L = 8151, C = ῗ, Us = "Ϊ͂"
;  L = 8162, C = ῢ, Us = "Ϋ̀"
;  L = 8163, C = ΰ, Us = "Ϋ́"
;  L = 8164, C = ῤ, Us = "Ρ̓"
;  L = 8166, C = ῦ, Us = "Υ͂"
;  L = 8167, C = ῧ, Us = "Ϋ͂"
;  L = 8178, C = ῲ, Us = "ῺΙ"
;  L = 8179, C = ῳ, Us = "ΩΙ"
;  L = 8180, C = ῴ, Us = "ΏΙ"
;  L = 8182, C = ῶ, Us = "Ω͂"
;  L = 8183, C = ῷ, Us = "Ω͂Ι"
;  L = 8188, C = ῼ, Us = "ΩΙ"
;     error(representation_error(character_code),char_code/2).

[razetime]

corrected in f10d7c0

[UWN]

Still, calling this to_upper/1 is highly problematic.

[aarroyoc]

I opened a PR to change the name of to_upper/1 and to_lower/1: #2129