[Toybox] [PATCH] Fix netstat -p.

[enh]

On Tue, Sep 8, 2015 at 9:23 PM, Rob Landley <rob at landley.net> wrote:
> On 09/08/2015 09:19 PM, enh wrote:
>> On Tue, Sep 8, 2015 at 10:53 AM, Rob Landley <rob at landley.net> wrote:
>>>> yeah, the other patch looks sensible but is wrong. (that was what i
>>>> tried first too :-) )
>>>
>>> Basically what I told the other committer, but I'm trying to be more
>>> encouraging to other developers and less of a bottleneck.
>>> (I still totally should have named this project dorodango.)
>>
>> want to commit the other lsof patch then? :-)
>
> Grr. Too many balls in the air...
>
> Work sprung another pending trip to Japan on me, only for a week this
> time, but I fly out this coming Saturday and have to do a number of
> non-toybox things by then (mostly http://nommu.org/jcore related).
>
> I backed out the lsof changes I was working on and checked in yours, I
> can redo my stuff on top of it. (Possibly after Japan.)

/me mumbles something about ps :-)

i resubmitted the switch to toybox ls and it seems to have stuck this
time. i haven't been at my desk all week, so i'm very behind on mail,
but as far as i can tell there have been no problems reported. i'll
send you another roadmap update tomorrow when i've double-checked.

i want to check with the systems folks about df (to realpath or not to
realpath) because although i personally prefer the named links (both
on my "goobuntu" laptop which doesn't use uuids and -- more relevantly
-- on Android), i also mainly care about the "Mounted on" column, so i
should talk to those more likely to need the "Filesystem" column about
whether they want the named partition symlinks or the numbered mcc
devices those links point to.

but ps (and to a much lesser extent top) seem to be the larger things
standing between here and my goal of having switched to toybox in N.
(sure, vi and less would be very popular, but not having them wouldn't
be a regression.)

...in case you get bored on the long flights :-)

>>> Besides, "fixed it for them" is valuable, and the point of the pending
>>> directory is so I'm not going to forget to fix it properly if I ever
>>> make it that far down my todo list. :)
>>
>> fwiw, people definitely use netstat on Android and this was the first
>> bug report i've had. (and they were only trying to use netstat -p
>> because lsof couldn't decode sockets, which is now fixed too.)
>
> Yay.
>
>> that said, i did just notice that -e doesn't work right (it doesn't
>> add the extra user/inode columns to the tcp output), and non-wide mode
>> (i.e. without -W) doesn't truncate IP addresses. patch attached.
>
> And applied.
>
> Thanks,
>
> Rob



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

 1441938497.0