Solved [g]sed gone mad

[graudeejs]

I'd like to share a bug that I noticed after upgrading to FreeBSD 11.

$ export LANG=en_US.UTF-8
$ echo "Test" | sed 's/[A-Z]/X/g'
XXXX

More info: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=216868

 

[graudeejs]

Interestingly textproc/gsed is also affected. I really didn't expect that.

 

[graudeejs]

graudeejs
Have you contacted devels upstream? They probably don't read FreeBSD-PRs.

Nope. How would you suggest doing that (other than PR)?

 

[chrbr]

I see no issue on 11-STABLE, same as comments #1 and #2. My locale is

# locale
LANG=de_DE.UTF-8
LC_CTYPE="de_DE.UTF-8"
LC_COLLATE="de_DE.UTF-8"
LC_TIME="de_DE.UTF-8"
LC_NUMERIC="de_DE.UTF-8"
LC_MONETARY="de_DE.UTF-8"
LC_MESSAGES="de_DE.UTF-8"
LC_ALL=

A test with a temporary export LANG=en_US.UTF-8 shows the same result.

 

[graudeejs]

http://pasteboard.co/D6PQ2eeuC.png (img tag seams to not work correctly)

 

[tobik@]

http://pasteboard.co/D6PQ2eeuC.png (img tag seams to not work correctly)

So what FreeBSD version are you on? The output of freebsd-version and uname -a is missing from your screenshot.

I see no issue on 11-STABLE, same as comments #1 and #2.

Same on 12.0-CURRENT.

 

[Eric A. Borisch]

I think you're seeing this:

http://unix.stackexchange.com/questions/227070/why-does-a-z-match-lowercase-letters-in-bash

Try LC_COLLATE=C

 

[graudeejs]

I think you're seeing this:

http://unix.stackexchange.com/questions/227070/why-does-a-z-match-lowercase-letters-in-bash

Try LC_COLLATE=C

You're right. This solves my problem.
However I want to point out, that this problem wasn't present on FreeBSD 10.


For the record: I'm running 11.0-RELEASE-p8

 

[graudeejs]

(comment removed)

 

[chrbr]

echo "Test" | sed 's/[:upper:]/x/g'

According to re_format(7) you need more brackets as

echo "Test" | sed 's/[[:upper:]]/x/g'
xest

I was surprised, too .