From: busybox <> on behalf of Tito 

Sent: October 13, 2016 3:01 AM
Subject: Re: Subject: [PATCH] strings: Correct output to be decimal per the 
help, instead of octal

On 10/12/2016 08:58 PM, Chris Renshaw wrote:
> Hello!
> They chose octal because they have -t (--radix) so other outputs may be
> chosen as needed. The usual default for implementations that do not have
> -t is decimal, reflected in busybox having this in the help for strings.
> If someone could write full -t support that would of course be ideal,
> but until then, decimal is the more common (and more useful) output.
> Cheers,
> Chris

can you list some of this implementations that have this
"usual default"?


PS.: please don't top post.

> ------------------------------------------------------------------------
> *From:* busybox <> on behalf of Tito
> <>
> *Sent:* October 11, 2016 3:38 PM
> *To:*
> *Subject:* Re: Subject: [PATCH] strings: Correct output to be decimal
> per the help, instead of octal
> On 10/11/2016 04:00 PM, Chris Renshaw wrote:
>>>From fe9333ce1d69e292e14144a61ff88a23ffed0053 Mon Sep 17 00:00:00 2001
>> From: Chris Renshaw <>
>> Date: Thu, 22 Sep 2016 16:17:05 -0300
>> Subject: [PATCH] strings: Correct output to be decimal per the help, instead 
>> of octal
>>  Most smaller/incomplete implementations of strings output decimal by 
>> default,
>>  so best to correct that behavior rather than just change the help to say 
>> octal
>>  (it would be great to get full POSIX --radix at some point, to give the
>>  output in any desired format, but this will do for now)
>> ---
>>  miscutils/strings.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>> diff --git a/miscutils/strings.c b/miscutils/strings.c
>> index 9f50182..cc4ad9a 100644
>> --- a/miscutils/strings.c
>> +++ b/miscutils/strings.c
>> @@ -68,7 +68,7 @@ int strings_main(int argc UNUSED_PARAM, char **argv)
>>                                                        printf(fmt, *argv);
>>                                                }
>>                                                if (option_mask32 & 
>> -                                                     printf("%7"OFF_FMT"o 
>> ", offset - n);
>> +                                                     printf("%7"OFF_FMT"d 
>> ", offset - n);
>>                                                }
>>                                                fputs(string, stdout);
>>                                        }
> Hi,
> man STRINGS(1) GNU Development Tools binutils-2.25 says:
> "-o  Like -t o.  Some other versions of strings have -o act like -t d
> instead.
> Since we can not be compatible with both ways, we simply chose one."
> As we mimick GNU strings for linux maybe correcting the help would be
> the better
> option.
> Ciao,
> Tito
busybox mailing list


Ah, sorry about the "top posting", it's the default in my mail client. :)

Most prominently "toybox" also has a similarly limited strings implementation 
and has decimal as the only offset output, also, on Android several OEMs have 
previously implemented strings as part of the old "toolbox" multi-call binary 
and, in my understanding of these, they also output decimal. So I do believe 
there's precedence for approving my patch to change the single character to 
have the output match the busybox help's advertised "decimal" (and it's been 
advertised as that for many years).

Further, I'd just like to reiterate that if it was an arbitrary by GNU to 
choose octal, maybe busybox could make a more informed decision; dd and a 
number of other important utilities use decimal only and it can be difficult to 
convert octal to decimal in a lot of shells, so decimal would be best for 
interoperability and usefulness.

Once again, proper -t support would of course be ideal so we could have our GNU 
cake and eat it too, but would need a hand more familiar than mine with the 
source. Perhaps you could help, Tito?

Thanks for your time and consideration everyone!
busybox mailing list

Reply via email to