[Toybox] [PATCH] Fix find -perm.

[enh]

On Fri, Nov 13, 2015 at 6:53 PM, Rob Landley <rob at landley.net> wrote:
> On 11/13/2015 07:52 PM, enh wrote:
>> Yeah, we still have two different bug databases. This is the one you
>> can't see, but there's nothing in the bug that isn't in the commit message.
>>
>> Fun fact: you can append a commit message to random internal bugs by
>> adding a "Bug: " line with one of these numbers or URLs ☺
>>
>> I usually don't add the bugs because they're meaningless to you, but I
>> don't really have a way to communicate "this was found by some other
>> internal team, not by me".
>
> I assume you forward stuff from Google/Android the way Jose Bollo
> forwarded stuff from Tizen and Ashwini Sharma/Hyejin Kim from Company
> Carefully Remaining Nameless and so on.

not really. Gilad's working on Brillo, and they're at least as
pro-AOSP as my team. but PMs and release teams are all focused on the
internal databases, so the internal bug databases still tend to be the
canonical ones, sadly.

> That said, gotta attribute the commit to somebody. :)
>
>> (also, since Gilad fixed the bug in the same way at the same
>> time, but minus the tests, i didn't want to steal all the credit.)
>
> I prefer to credit the bug finder given a choice. (Implementation is
> easy, figuring out what to do is the hard part. Reproducible test
> sequences that demonstrate a bug are _gold_...)

bug report: Gilad. fix: identical fixes independently developed by
both of us, because i didn't expect someone to report a bug and fix
it. tests: me. Gilad has another patch (that he's already sent to the
list) that improves -perm support further.

> I still have a local diff against find left over from
> http://lists.landley.net/pipermail/toybox-landley.net/2015-September/004464.html
> and I don't remember if this was one of the two issues I was trying to
> fix then, or if it's a third.

i did notice that the test for an unterminated -exec fails; that was
one of the bugs you thought you'd fixed. i ought to run all the tests
more frequently, but -- in part because there are so many failures --
i tend to just run the tests for the command i'm working on.

> (Sorry, modifying a dozen commands in
> parallel, balls tend to get dropped... Currently packing for my trip to
> Santa Diego and Santa Clara next week, might not get to this until I'm
> on the plane.)

at least it's not Japan this time!

> Rob



-- 
Elliott Hughes - http://who/enh - http://jessies.org/~enh/
Android native code/tools questions? Mail me/drop by/add me as a reviewer.

 1447472453.0