[lkml]   [2012]   [Apr]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: Our failure on tracing tools unification
Em Wed, Apr 25, 2012 at 10:19:35AM -0400, Frank Ch. Eigler escreveu:
> mingo wrote:
> > [...] One thing that would suck is perf being bound by various
> > versions of this library floating out there.
> Could the event parser shared library be ABI-preserving? If so, what
> would perf have to fear?

Perhaps its because people working on perf grew in the kernel, where we
can change things as we see fit, i.e. change the networking API? Sure,
we can go over the drivers and change them, its all in the same source
tree, etc.

> > [...] We had nothing but pain from external libraries so far [...]
> If the elfutils, newt, gtk, bfd, iberty, zlib, python libraries didn't
> contribute useful and un-rewriteable functionality, perf would not be
> using them.

newt was a pain and will go away, 99% of things rely just on libslang,
but even that one could go, bfd we use just for C++ demangling, so
probably could go too.

- Arnaldo

 \ /
  Last update: 2012-04-25 18:41    [W:0.117 / U:13.048 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site