Dependencies

[GoogleCodeExporter]

Hi Anders,
What are the details on the boost dependencies? Going to the boost page
that is linked from the main page of this project just takes one to the
boost page with no indication of what to download and compile. I run a
linux box and a Mac OS X machine here, so some more details would be good.
Thanks,
Dan

Original issue reported on code.google.com by ddfalck2...@yahoo.com on 27 Mar 2010 at 1:43

[GoogleCodeExporter]

On my Ubuntu 9.10 I have all of the "libboost-1.38.0" packages installed, but I 
think
the following packages should be enough:
libboost-python1.38-dev
libboost-python1.38.0
libboost-dev
libboost1.38-dev

If there's anyone interested in this project who is good at packaging and 
finding out
dependencies then a list of packages or build/install instructions for
Ubuntu/Windows/MacX would be a good addition to the wiki or the manual.

Original comment by anders.e...@gmail.com on 27 Mar 2010 at 3:38

[GoogleCodeExporter]

I started to document the build dependencies (for Karmic) in the debian 
"control"
file.  It currently says:

Build-Depends: debhelper (>= 7), libboost-dev, python-dev, python-support

debhelper and python-support are only needed if you want to build the debian 
package,
libboost-dev and python-dev are needed to compile.  I'm not sure if a 
finer-grained
boost dependency is possible. 

That list may be incomplete, i'll double-check it (using pbuilder) and update 
if needed.

Original comment by seb.kuzm...@gmail.com on 27 Mar 2010 at 7:33

[GoogleCodeExporter]

I was able to build it on a Ubuntu8.04 box here. One thing that I had to do to 
make
it work was install the python libs to site-packages instead of dist-packages. 
I'm
not up on how the dist-packages are supposed to work. My system in fairly old- 
maybe
1.5 years old since I set it up.
I found that the libboost1.36 packages that installed through synaptic did work 
and I
am able to use this lib through HeeksCNC.

Thanks,
Dan 

Original comment by ddfalck2...@yahoo.com on 27 Mar 2010 at 7:45

[GoogleCodeExporter]

I think the only part of boost that's needed to compile ocl is 
libboost-python-dev. 
This works on Hardy and Karmic.  It does not work on Lucid, because the name of 
the
library changed between Karmic and Lucid: on Karmic and earlier it's
"-lboost_python-mt", on Lucid (and later?) it's just "-lboost_python".

Sounds like a job for autoconf...  Anyone mind if i autoconf-ify ocl?

Original comment by seb.kuzm...@gmail.com on 28 Mar 2010 at 7:40

[GoogleCodeExporter]

seb, if autoconf makes it easier to build opencamlib on various platforms, then
please go ahead. If you have time, you could write down some build-instructions 
into
manual.h (documentation in Doxygen format)

Original comment by anders.e...@gmail.com on 28 Mar 2010 at 8:09

[GoogleCodeExporter]

I just added a cmake file (in r100).  cmake is an alternative to autotools 
that's a
bit easier on the eyes.  I've verified that is works on Ubuntu Hardy, Karmic, 
and
Lucid.  Lucid does not work with our old makefile, because the Ubuntu 
maintainers
changed the name of the boost-python library; the cmake stuff figures out which 
one
is available and uses that one.

If anyone cares about Windows or Mac OS X, it'd appreciate testing and feedback.

Original comment by seb.kuzm...@gmail.com on 30 Mar 2010 at 11:10

[GoogleCodeExporter]

make -j 4
is very useful on multi-core machines.
is there a way to have cmake produce a makefile which does that by default?

Original comment by anders.e...@gmail.com on 31 Mar 2010 at 12:21

[GoogleCodeExporter]

The Makefiles made by cmake do *allow* parallel compilation (ie, "make -j4" 
will work
and produce consistent results), but unfortunately they do not enable it by 
default.  :-(

Original comment by seb.kuzm...@gmail.com on 31 Mar 2010 at 3:24

[GoogleCodeExporter]

The debian package build script (debian/rules) runs "make -j4".

I'm going to call this issue closed now, since the original question has been 
answered.

Original comment by seb.kuzm...@gmail.com on 19 Apr 2010 at 5:43

• Changed state: Done