From: Kaz Kylheku <773-297-7223@kylheku.com>

On 2020-04-08, Ben Bacarisse <ben.usenet@bsb.me.uk> wrote:

Kaz Kylheku <773-297-7223@kylheku.com> writes:

On 2020-04-08, stbalbach2@gmail.com <stbalbach2@gmail.com> wrote:

Given a unicode string:

/usr/bin/printf "\u041c\u043e\u0442\u0438\u043d"

Result: "D£D1ÑéD,D½"

Is there a native gawk way other than invoking /usr/bin/printf ?

Awk doesn't have \u escapes.

The following works for me with GNU Awk and Mawk on
Ubuntu 18:

$ awk 'BEGIN { print "D£D1ÑéD,D½" }'
"D£D1ÑéD,D½"

That is to say, the implementations appear to be 8 bit clean in the
handling of string literals, so you can write source code in UTF-8,
embedding the extended characters directly.

The gawk I use (4.2.1) appears to be UTF-8 aware, not just 8-bit clean:

$ awk '/caf[ÄCe]/ {print length($1)}'
cafe
4
cafÄC
4

To yank your leg a bit, that could plausibly just be 8 bit ISO-Latin. ;)

But if we try it on the original example, it does report 5 for length("D£D1ÑéD,D½").

OTOH, mawk reports 10.

I also tried with an older 4.1.60 that I built from sources a few years ago; that already had it working.

Indexing works too:

0:sun-go:~/gawk$ ./gawk 'BEGIN { s = "D£D1ÑéD,D½"; print substr(s, 2) }' D1ÑéD,D½
0:sun-go:~/gawk$ ./gawk 'BEGIN { s = "D£D1ÑéD,D½"; print substr(s, 2, 2) }' D1Ñé

--- BBBS/Li6 v4.10 Toy-4
* Origin: Prism bbs (1:261/38)