[Toybox] all argument parsing broken right now?

[Rob Landley]

On 07/16/2016 11:38 PM, Rob Landley wrote:
> On 07/16/2016 12:06 PM, enh wrote:
>> specifically #define OPTSTR_toybox 0 versus #define OPTSTR_toybox " "...
>>
>> On Sat, Jul 16, 2016 at 10:01 AM, enh <enh at google.com> wrote:
>>> seems to be the mkflags.c change in
>>> https://github.com/landley/toybox/commit/adef5dcb1857f524e2a24dd0223f31db7cd7a8b9
> 
> Oops. Sorry, I tested standalone but didn't test the multiplexer.
> 
> The point of the change is to make OLDTOY() argument parsing work so
> sha256sum and friends could share help text with sha1sum. And it did!
> 
> The problem I was trying to solve is that if the base command is
> disabled but an OLDTOY() for it is enabled, the base command's option
> string was a null pointer. The NEW problem is that the multiplexer
> itself has an option string that SHOULD always be a null pointer, and
> now isn't.
> 
> Gotta get the conditionals right. Hmmm...

Ok, simple fix checked in and pushed, but now I've got to think through
the design stuff some more.

I'm writing up a big long blog entry explaining the issues (basically a
sequel to http://landley.net/notes-2016.html#23-06-2016) but probably
won't have it up tonight.

(I could post it here, but the darn web archive's gone down enough times
I dislike having this sort of thing vanish if it does. And no it's not
just Dreamhost, I vaguely recall I did some long elaborate writeups back
in 2008-2010 that got erased when
http://landley.net/notes-2010.html#03-11-2010 happened. I still have the
blog entry complaining about it, but not the content of the old posts.)

Rob