Re: Support bytesize comparison in sort

[Mart Somermaa]

Paul Eggert wrote:
> Mart Somermaa <address@hidden> writes:
>
>   
>> Why not use an upper-case letter? We have plenty of those left.
>>     
>
> There are a lot of different ways to sort, and we'll burn through
> the letters fairly quickly.  
Out of the 52 (2x26 for lower- and uppercase English alphabet letters)
option letters, only 18 are taken. There are 34 -- about twice as much
-- left.

$ sort --help | egrep '^\W+-\w,' | wc -l
18

> If we used a single letter, it should
> probably be -h for df -h and -H for df -H, for sanity's sake.
>   
As I already explained, the current implementation behaves correctly
with *both* SI and powers-of-1024 input, so there is no need for two
option letters.
> But I'd still prefer a scheme that doesn't require a single letter.
>
> How about something like "sort -k 3--ignore-case,--ignore-leading-blanks"?
> Then you could get the effect that you want with "sort -k 3--si".
>   
Pardon me, but it's ugly, counterintuitive and cumbersome.
You haven't convinced me in the inadequateness of the patch, but as I
respect the work you have done for coreutils, I'll just quit the
discussion :) .