Restricting string length

[khepin]

How would I put a constraint on a string's length? Right now I've only figured out how to do it via regexes:

#StrLen8: =~"^.{8}$"
{
    v: #StrLen8
}

{
    "v": "hello123"
}

This also requires 1 regex / length unless I'm missing something...

[mpvl]

import "strings"

v: strings.MaxRunes(8) & strings.MinRunes(8)

There is a PR pending to add NumRunes too.

But you could make regex's variable:

#StrLen: {
    =~"^.{\(#n)}$"
   #n: int
}
#StrLen8: { #StrLen, #n: 8 }

Scalar values may have definition fields associated with them. This is a fairly new feature and some syntax has been proposed to make this more natural.

[mpvl]

Note that the let builtin deliberately does not work on strings. String length is kind of a complicated matter, so we decided that one needs to be very explicit about the kind of string length that one is using.

[khepin]

Good point on string length not being straightforward. Easy to overlook. So from previous answers, I imagine MaxRunes counts UTF8 runes. And the regex version here would count bytes right?

[extemporalgenome]

I imagine MaxRunes counts UTF8 runes

There shouldn't be any distinction between "unicode runes" and "utf-8 runes" since rune is a semantic concept, regardless of the underlying encoding.

String length is kind of a complicated matter, so we decided that one needs to be very explicit about the kind of string length that one is using.

Could we just have a default meaning for bytes vs string, such that len('...') always returns the byte-length, while len("...") returns the number of runes, leaving the very-explicit functions for the job of counting bytes in a string or counting runes in a bytes?

[cueckoo]

This discussion has been migrated to cue-lang/cue#657.

For more details about CUE's migration to a new home, please see cue-lang/cue#1078.