On Sun, Apr 12, 2015 at 11:56:30AM +0200, Janne Grunau wrote:
> On 2015-04-09 18:44:59 +0200, Diego Biurrun wrote:
> > On Thu, Apr 09, 2015 at 12:39:33AM +0200, Janne Grunau wrote:
> > > On 2015-04-07 14:38:59 +0200, Diego Biurrun wrote:
> > > > On Tue, Apr 07, 2015 at 10:22:16AM +0200, Luca Barbato wrote:
> > > > > On 06/04/15 23:28, Vittorio Giovara wrote:
> > > > > >On Sat, Apr 4, 2015 at 6:43 PM, Diego Biurrun <[email protected]> 
> > > > > >wrote:
> > > > > >>On Thu, Apr 02, 2015 at 01:55:26PM +0200, Vittorio Giovara wrote:
> > > > > >>>On Thu, Apr 2, 2015 at 1:20 PM, Diego Biurrun <[email protected]> 
> > > > > >>>wrote:
> > > > > >>>>>--- a/configure
> > > > > >>>>>+++ b/configure
> > > > > >>>>>@@ -315,6 +315,8 @@ Developer options (useful when working on 
> > > > > >>>>>Libav itself):
> > > > > >>>>>                             (group) and PROB the probability 
> > > > > >>>>> associated with
> > > > > >>>>>                             NAME (default 0.5).
> > > > > >>>>>    --random-seed=VALUE      seed value for 
> > > > > >>>>> --enable/disable-random
> > > > > >>>>>+  --disable-valgrind-backtrace do not print a backtrace under 
> > > > > >>>>>Valgrind
> > > > > >>>>>+                           (needs --disable-optimizations)
> > > > > >>>>
> > > > > >>>>I don't like this.  Why not automatically enable it when 
> > > > > >>>>optimizations
> > > > > >>>>are disabled instead?
> > > > > >>>
> > > > > >>>It is automatically enabled when optimizations are disabled, I just
> > > > > >>>added some help text because users could wonder why it's not 
> > > > > >>>enabled
> > > > > >>>(eg. they do need to disable optimizations).
> > > > > >>
> > > > > >>The text is still confusing.  But before fixing that I think we 
> > > > > >>should
> > > > > >>consider enabling this unconditionally if optimizations are 
> > > > > >>disabled.
> > > > > >>What's the point of having a disable switch for this?  (We already 
> > > > > >>have
> > > > > >>far too many configure parameters.)
> > > > > >
> > > > > >I think it was to stifle any concern for packagers having to deal (or
> > > > > >not) with this feature.
> > > > > >I am completely fine with removing the additional switch.
> > > > > 
> > > > > As one of the concerned packager I'm against removing it, builds must 
> > > > > be
> > > > > predictable.
> > > > 
> > > > How is a build w/o this new parameter unpredictable?
> > > 
> > > it enables/disables the feature depending whether the build enviroment 
> > > has a proper valgrind/valgrind.h header.
> > 
> > Yes of course, but I see nothing unpredictable there.
> 
> unpredictable is maybe the wrong word if you assume knowledge of 
> configure / the build system. It creates build ebviroment dependent 
> binaries without explicit control from configure. That's enough reason 
> to avoid it for me. If you're concerned about the length/details of the 
> $(configure --help) output, we can think about adding --help=full and 
> move more obscure options there.

I still don't think it's a problem worth worrying about in practice.
Maybe somebody else can weigh in to convince me?

Diego
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to