On Thu, May 25, 2006 at 09:15:55PM -0500, Jeff Epler wrote:
Paul,
You have a lot of guts to show up and demand that we make our debian rules build on your OS. To condescendingly talk like I don't know how to call uname(2). To make accusations that there are GPL problems in emc2, but grow mysteriously silent when you're asked what they are. [...] I apologize to everyone, particularly Paul, for the accusation I made. It's now reasonably clear that in the case of the python-axis packages, the problem is one of communication and documentation, rather than a violation of the GPL.
After I posted this message, Matt Timmermans contacted me personally. Thanks to his help, I've been able to document how to build the python-axis package on bdi4. The process is unobvious and different to the way I'm familiar with building debian packages, but it does appear to work. The biggest surprise is that it's necessary to first copy and build emc in /usr/local/emc before building axis. These are the instructions that I followed (excuse the markup, it's a copy of what I placed on the bdi4emc wiki): = Recompiling the 'python-axis' package = ''The following instructions are based on ones given in a private communication with Matt Timmermans'' On a vanilla bdi-4.46 system with the install cdrom in the drive, the following script successfully creates the 'python-axis' and 'python2.3-axis' debian packages: {{{ sudo apt-cdrom add sudo apt-get update; sudo apt-get upgrade sudo apt-get install xlibmesa-gl-dev xlibmesa-glu-dev fakeroot build-essential rtai-dev-2.6.16-rtai sudo apt-get build-dep emc python-axis apt-get source emc python-axis sudo -s << EOF cp -R emc-1.0/src /usr/local/emc/src cd /usr/local/emc/src ./configure && make EOF cd python-axis-1.2.1 dpkg-buildpackage -rfakeroot -uc -us }}} I would be tickled pink if someone familiar with the process to build the emc and emc-modules debian packages for bdi4emc could also document that on the wiki. Jeff
Jeff On Sunday 28 May 2006 00:23, Jeff Epler wrote:
demand that we make our debian rules build on your OS.
First: Not "my OS", but Debian - The same base that Knoppix, Morphix, Ubuntu, Xandros, and countless others use. Second: There was no "demand to", simply an observation that hard coded dependencies was wrong (in my opinion) - Anyone wanting to compile, install, and run on any flavour of GNU/Linux (Debian, Red Hat, SuSE, Gentoo, et al) will run in to the same problem.
To condescendingly talk like I don't know how to call uname(2).
It was a suggestion for a viable kernel independent solution and in no way intended as a condescending insult.
To make accusations that there are GPL problems in emc2
Copyright infringements (which may be construed as GPL violations). Alex has a list of files that are suspect - But to repeat. The genhexkins files are written by Brian Register, (to the best of my knowledge) he was not in the employ of NIST at the time, so the two files are not public domain. Therefor to slap a GPL notice on his work without express permission is questionable at best. Without express permission from Mesa Electronics, any/all of the 5i20 drivers and firmware should be regarded as non-free - I believe Peter Wallace gave permission to release the source code as GPL, but as far as I am aware, copyrights were not transfered. Both points need to be publically clarified.
I apologize to everyone, particularly Paul, for the accusation I made.
Accepted.
It's now reasonably clear that in the case of the python-axis packages, the problem is one of communication
Not just axis, and just one of the reasons I don't subscribe to the emc-dev list and rarely read emc-users. Regards, Paul.
I was just playing around Saturday night, updating and such, and noticed that there was a new set of packages on the bdi4emc repository. Not seeing anything on the mailing list, I thought it may be just packages for testing, but I decided to try them out. I updated emc, python-axis, and installed linux-image-2.6.16.18-rtai and modules. EMC failed as follows: starting EMC MOTION PROGRAM -- quickstep... Can not find -sec EMCMOT -var STEPPING_TYPE -num 1 Can not find -sec TRAJ -var BLEND_TOL_REL -num 1 IO_BASE_ADDRESS 0x378 modprobe quickstep PERIOD=10000 IO_BASE_ADDRESS=0x378 SHMEM_KEY=100 QS_PINOUT=0x76543210 done starting EMC IO PROGRAM -- minimillio...done simDioInit(garten.ini) iniFind is depreciated starting EMC TASK PROGRAM -- minimilltask...done iniFind is depreciated rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed running EMC DISPLAY PROGRAM -- axis... rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed garten.ini was not changed. generic.var was not changed. Removing kernel module rtai_math and anything that depends on it. Including quickstep Removing kernel module rtai_shm Removing kernel module rtai_hal and anything that depends on it. Including rtai_lxrt The above is an incomplete copy/paste. Axis will come up, and stay until shut down, but nothing works.
Hi Thomas On Sunday 04 June 2006 22:05, Thomas Fritz wrote:
I was just playing around Saturday night, updating and such, and noticed that there was a new set of packages on the bdi4emc repository. Not seeing anything on the mailing list, I thought it may be just packages for testing, but I decided to try them out.
Yup, a couple of new packages that are destined for the next BDI-4 CD.
I updated emc, python-axis, and installed linux-image-2.6.16.18-rtai and modules.
EMC failed as follows:
<snip>
rtai_alloc(100 (0x64), 688592 (0xA81D0) ) failed
The only time this error message should crop up is when the kernel modules failed to load or udev is running but failed to create /dev/rtai_shm - The last can be discounted because the standard BDI-4 install does not use udev.
The above is an incomplete copy/paste.
Could you send me the complete copy of the output of dmesg, along with the ini file please. Regards, Paul.
participants (3)
-
Jeff Epler -
Paul -
Thomas Fritz