Well, I managed to get this all up and running on Arch Linux. I had to
edit the ganv-svn PKGBUILD, as it was stuck using an old revision from
svn (from a couple of years ago I guess). Getting it to build from the
newest version resolved the issues with it not building against
graphviz. I also had to edit the 'provides' field in several PKGBUILDS
and rebuild them, as it seems there is some strict version checking
going on somewhere in the drobilla config files, which does not allow
for building one svn against another svn package (it's checking for
the presence of release versions rather than svn revisions).
And after all that, I was able to (for the first time ever) get ingen
up and running. Unfortunately a couple of minutes playing around seems
to be all I can get before hitting a segfault (which has always been
my experience with AMS anyway, on various different machines and
distros). Now, I don't have much experience of bugfinding, so perhaps
before I file issue # 1 on the ams-lv2 bug tracker, somebody could
suggest a sensible route to determining whether it's ams-lv2 or ingen
that's to blame.
Learning, as ever....
On 10 February 2014 12:26, mark hadman <markhadman(a)googlemail.com> wrote:
On 9 February 2014 19:52, michael noble
<looplog(a)gmail.com> wrote:
"The easiest way to build Ingen from SVN is to build the entire
http://svn.drobilla.net/lad repository (since there are dependencies between
the various projects)."
I guess Mark's point is that it has been impossible to build ingen from the
AUR scripts for some time, which is sadly true. Trying to mix svn of some of
drobilla libraries with AUR or Arch packages of those libraries, or even the
full svn tree with apps that depend on the official packages, quickly
results in a mess.
Ganv doesn't build at all; I've already tried the ganv-svn PKGBUILD in
Arch. Canvas.cpp makes a call to graphviz (which is 'a mess' to
paraphrase someone involved in maintaining ganv) with the wrong number
of arguments. I guess I could also try graphviz-git (instead of the
Arch binary) in the vague hope that the authors of graphviz have
reverted the change that broke ganv, but graphviz-git requests over
100MB of dependencies to build, for which I don't have the bandwidth
to burn unless someone out there can assure me that it's worthwhile.