Re: [coreutils] [PATCH 1/2] sort: avoid unnecessary use of sprintf

[Jim Meyering]

Pádraig Brady wrote:
> On 08/06/10 10:42, Jim Meyering wrote:
>> This first patch removes uses of sprintf in favor of
>> umaxtostr and stpcpy.
>
> A little less standard, and we don't really care
> about performance in --debug, but fair enough.

It was partly on principle: keep things tight/minimal,
since there were no other uses of sprintf.
Also, using sprintf at all raises warning flags,
since it's so easy to use incorrectly...

> Note I think you can remove the uintmax_t case from
> the umaxtostr parameters.

Thanks.  I've removed those two casts.

>> It also adjusts some uses of INT_BUFSIZE_BOUND to use
>> a variable name rather than a type.
>
> Good one.

>> Subject: [PATCH 2/2] maint: adjust INT_BUFSIZE_BOUND usage for 
>> maintainability
>>
>> * src/tail.c (xlseek): Give INT_BUFSIZE_BOUND a variable name,
>> not a type name.
>> * src/ls.c (gobble_file, format_user_or_group_width): Likewise.
>> * src/head.c (elide_tail_bytes_pipe): Likewise.
>> (elide_tail_lines_seekable, main): Likewise.
>
> There may be others?
> grep "BOUND ([ilus]" *.c

Yes, definitely.
I didn't finish the job.
Several uses in ls.c looked like changing would not benefit.
Same for these two in test.c, since there is no variable per se:

        {
          char lbuf[INT_BUFSIZE_BOUND (uintmax_t)];
          char rbuf[INT_BUFSIZE_BOUND (uintmax_t)];
          char const *l = (l_is_l
                           ? umaxtostr (strlen (argv[op - 1]), lbuf)
                           : find_int (argv[op - 1]));
          char const *r = (r_is_l
                           ? umaxtostr (strlen (argv[op + 2]), rbuf)
                           : find_int (argv[op + 1]));

But I haven't looked at many of the remaining cases.
If you feel like it, you're welcome to finish.