[PD] helpbrowser and the paths

[Hans-Christoph Steiner]

On Thu, 05 May 2011 10:20 +0200, "IOhannes m zmölnig" <zmoelnig at iem.at>
wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 05/05/2011 03:38 AM, Hans-Christoph Steiner wrote:
> > 
> > On May 4, 2011, at 5:47 PM, Paul Brossier wrote:
> > 
> >> On Wed, May 04, 2011 at 12:18:04PM -0400, Hans-Christoph Steiner wrote:
> >>>
> >>> Using Pd-extended 0.43.1-20110430 on Mac OS X I don't see this
> >>> behvaior with Gem or gridflow.  The helpbrowser.tcl code is the same
> >>> between pure-data and pd-extended.
> >>
> >> How do you start pd to load Gem?
> >>
> >> Removing '-path /usr/lib/pd/extra/Gem', the help browser behaves as
> >> expected.
> 
> yes, but then Pd doesn't find Gem's abstractions.
> that's the reason why the "-path" is there in the first place.
> 
> > Check Gem_addownpath() in src/Base/GemSetup.cpp.  If Gem's ./configure
> > finds s_stuff.h, it should enable this.  This is something like what
> > gridflow does too.  Then pd -lib Gem also puts path/to/Gem/abstractions
> > in the path also.
> 
> which i have enabled now in my latest push.
> (and disabled the -path in pd-gem)
> 
> i still think that the separating libraries and files in the
> help-browser is a good idea (just like static and dynamic content are
> separated, and patches and non-patches are separated)

If this is related to the patch you submitted, as long as its clean, I
am fine with including it.  But sorting folders before files is not the
solution to the problem of pd -lib Gem not loading everything into the
global namespace.  Luckily there is already Gem_addownpath() because I
think that is the way it should be handled.

.hc