[Toybox] sort -V

Rob Landley rob at landley.net
Wed Dec 19 13:14:31 PST 2018



On 12/19/18 2:17 PM, enh wrote:
> On Wed, Dec 19, 2018 at 12:04 PM Rob Landley <rob at landley.net> wrote:
>>
>> On 12/19/18 1:41 PM, enh via Toybox wrote:
>>> it turns out that we have a few users of `sort -V`. i'd always assumed
>>> this meant "use the GNU strverscmp(3) function as your comparator" but
>>> no, it's slightly different...
>>
>> By the way, do any of the regex engines support utf8 yet?
> 
> libc you mean? i'm not sure anyone knows, because anyone serious is
> already using Java or PCRE or RE2 :-)

Regex support is 3 functions and 2 data types, I'm pretty happy to throw
wrappers in portability.h to support other libraries. I just dunno what my
options are. (Musl used TRE but that never attempted non-ascii support.)

>> (I'm aware that china despises utf8 because all their chars expand to 4 bytes
>> and they use a 2 byte encoding instead, but if utf8 gets supported we can do an
>> internal conversion, and it gets us the _rest_ of the planet...)
> 
> China's not your problem... https://www.unicode.org/faq/myanmar.html
> is your problem.

1.38 billion people vs 0.05 billion. I'm not worried about the hardest fixes,
I'm wondering what the low hanging fruit is. I'm also wondering why thunderbird
decided the this line does not need to be wordwrapped...
Rob


More information about the Toybox mailing list