lkml.org 
[lkml]   [2014]   [Apr]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [GIT PULL 0/4] perf/urgent fixes

* Jiri Olsa <jolsa@redhat.com> wrote:

> SNIP
>
> >
> > Okay, so the problem is that we don't have a simple binary-state
> > feature in this case, but three possible states: 'libunwind', or
> > 'libdw-dwarf-unwind', or 'OFF', right?
> >
> > If so then the solution would be to replace those 3 last lines with
> > just this line:
> >
> > ... DWARF unwind library: [ libunwind ]
> >
> > Where 'libunwind' is printed in green (like the 'on' lines are
> > printed). If there's no suitable library available then output:
> >
> > ... DWARF unwind library: [ OFF ]
> >
> > Because the user looking at the output is really only interested in
> > 'is an unwind library available', and maybe in 'which one'.
> >
> > Is there preference between library choices? I.e. is 'libunwind'
> > preferred over 'libdw-dwarf-unwind', or the other way around? If yes
> > then if we pick an inferior library we could print it in yellow color
> > - and only use green if it's the 'best' choice.
> >
> > That way the color codes also still keep working: red means problem,
> > green means OK, yellow something inbetween.
>
> sounds good.. TODO list updated ;-)
>
> >
> > But in any case we should try to keep the 'one feature, one line'
> > fundamental output concept.
> >
> > ( Under V=1 we can output whatever details might be useful to
> > developers, there's no restriction on what to output there. )
>
> thats what we put VF for.. maybe we should for verbose
> features code detection output for V=1 as well

Yeah, I think it's only rarely needed, so might make sense to merge it
into V=1.

Thanks,

Ingo


\
 
 \ /
  Last update: 2014-04-25 10:41    [W:0.100 / U:0.232 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site