[Toybox] [PATCH] Added wc -L longest line option

[Rob Landley]

On 11/4/23 11:51, Oliver Webb via Toybox wrote:
> Heya, I added a -L[ongest line] option to wc
> 
> I thought that this option was considered and discarded since
> it isn't that hard to implement and is a GNU extension. But grep-ing through
> the list archives shows no mention of a -L option in relation to wc.

Because it's not in posix, not in the default output, and nobody ever asked for
it. Did you spot it in use in the wild?

> I added the -L option in the below patch, along with a couple of tests for it.
> GNU wc -L doesn't count newlines, so it makes sense that toybox's one shouldn't.
> the "(signed)" cast is so the output is never "-1"
> 
> Also, when I originally tried to implement this option, I appended it
> to the end of the option string. Then -l started acting in the same way as -L.
> Which made me take a look at generated/flags.h to check if options were being parsed
> case-sensitively.
> 
> Then I realized that option strings are parsed into "FLAG_" macros backwards.
> (End -> Start, First options get higher value flags)
> Which from my searching isn't mentioned in lib/args.c, nor scripts/make.sh.

Lines 20-22 of lib/args.c:

 * Each option in the get_opt string corresponds to a bit position in the
 * return value. The rightmost argument is (1<<0), the next to last is (1<<1)
 * and so on. If the option isn't seen in argv[], its bit remains 0.

https://landley.net/toybox/code.html#lib_args has a more elaborate description.

The bits are in the same order that binary digits go in. If option string abcd
has bits 1010 (I.E. 0x0a) then -a and -c were set and -b and -d weren't.

Rob