On 03/03/2014 10:54 AM, Pádraig Brady wrote:
> On 03/03/2014 09:36 AM, Pádraig Brady wrote:
>> On 03/03/2014 04:55 AM, Paul Eggert wrote:
>>> Pádraig Brady wrote:
>>>> +  -h, --human-readable  print abbreviated sizes in powers of 1024 (e.g. %s)\n\
>>>> +  -H, --si              print abbreviated sizes in powers of 1000 (e.g. %s)\n\
>>>
>>> Thanks for looking into improving this.  Please put commas after "e.g.".  Also, I suggest removing "abbreviated", and simplifying the translator's job by simply interpolating the human.  E.g.:
>>>
>>> -h, --human-readable  print sizes in powers of 1024 (e.g., 1023M)\n\
>>> -H, --si              print sizes in powers of 1000 (e.g., 1.1G)\n\
>>
>> Better thanks.
>> I've rolled that into my local patch.
> 
> Actually I left the %s in the string to be replaced at runtime,
> so that 1.1G or 1,1G is output as appropriate.

I suppose it's better to err on the side of less code,
and let the translator adjust, assuming LC_MESSAGES == LC_NUMERIC.
The alternative edge case is probably not worth the extra code.

I'll push the attached soon.

thanks,
Pádraig.