Add epics7 on rtems5 for beatnik #10
Conversation
| ======= | ||
| Tests: 15 | ||
| Passed: 15 = 100.00% | ||
|
|
There was a problem hiding this comment.
I do not see the value of adding the complete EPICS test output here, with no tests failing.
Any new version of EPICS Base is likely to produce slightly different output, anyway.
What about an ellipsis and cutting to the last handful of lines?
There was a problem hiding this comment.
The idea was to show that really all tests were successful and nothing remained open.
anjohnson
left a comment
anjohnson
left a comment
There was a problem hiding this comment.
Thanks for putting this together. Some suggestions inline...
| @@ -0,0 +1,4365 @@ | |||
| Getting EPICS 7 on RTEMS 5 (beatnik VMEbus board) | |||
There was a problem hiding this comment.
"beatnik" isn't a name that I've come across anywhere other than in relation to the RTEMS BSP of that name. I would suggest instead using a title like:
Building RTEMS 5 and EPICS 7 for MVME6100
There was a problem hiding this comment.
Maybe we should create a section "Latest developments" or "News"?
I can relate to all of Andrew's comments. My point was to support the current "hype" around RTEMS5. I wanted to give a documentation as accurate as possible. E.g. I used the IP addresses so that you can find the same server IP at beatnik-boot etc. I'm happy to modify and customize it as Andrew suggested and wait until we have an "official" github URL. I have already reached the currently active players ;-)
To the environments :
export RTEMS_MAKEFILE_PATH=${RTEMS_ROOT}/${RTEMS_ARCH}/${RTEMS_BSP}
export RTEMS_SHARE_PATH=${RTEMS_ROOT}/share/rtems${RTEMS_VERSION}
These are expected from the libbspext - Makefile.
There was a problem hiding this comment.
"latest" or "news" creates a maintenance task, as items have to be removed after they lose their novelty aspect.
We do have a "news" section on the main site - that's where announcements go.
How-Tos are instructions. As such they should be comprehensive and correct. A user following them doesn't care how old they are, as long as they work.
| :: | ||
|
|
||
| mkdir MVME6100 | ||
| cd MVME6100 |
There was a problem hiding this comment.
Shouldn't these paths be ~/MVME6100 since lower down you use $HOME/MVME6100? You're assuming the user starts off in their home directory, which is probably unwise.
| :: | ||
|
|
||
| RTEMS Source Builder - Set Builder, 5 (803d42cda7b3) | ||
| Command Line: ../source-builder/sb-set-builder --prefix=/home/junkes/MVME6100/rtems/5 5/rtems-powerpc |
There was a problem hiding this comment.
I know this is showing the logged output, but we might want to use something more generic than /home/junkes; maybe we should change all of those paths to something like /opt/MVME6100?
| export RTEMS_ROOT=${HOME}/MVME6100/rtems/${RTEMS_VERSION} | ||
| export PATH=${RTEMS_ROOT}/bin:${PATH} | ||
| export RTEMS_MAKEFILE_PATH=${RTEMS_ROOT}/${RTEMS_ARCH}/${RTEMS_BSP} | ||
| export RTEMS_SHARE_PATH=${RTEMS_ROOT}/share/rtems${RTEMS_VERSION} |
There was a problem hiding this comment.
I don't remember having to set any of those environment variables to build RTEMS-5, are these to avoid having to configure EPICS Base properly, or are they needed for BSP development? The only environment restriction I remember is that the source building refuses to run if . is in $PATH.
I hate advising people to add things to their ~/.bashrc unless they are really necessary, because very few people remove stuff later on and that file then fills up with crud that they're never sure whether they can delete or not.
|
|
||
| cd; cd MVME6100 | ||
| mkdir EPICS; cd EPICS | ||
| git clone --recursive https://github.com/hjunkes/epics-base.git |
There was a problem hiding this comment.
I'm reluctant to publish Heinz's GitHub URL on the official site like this. I know this is what people have to do to get the code right now, but hopefully we'll be able to merge this work into the 7.0 branch soon and I would prefer that our documentation be written to help us come up with the final solution. Publishing this document now is valuable for the 3-4 people working on it, but it should be temporary IMHO.
| RTEMS_VERSION = 5 | ||
| RTEMS_BASE = /home/junkes/MVME6100/rtems/$(RTEMS_VERSION) | ||
|
|
||
| make -j4 |
There was a problem hiding this comment.
The make line isn't obviously separated from the edit instruction, users might add it to the file; this also assumes the user has 4 CPUs. When doing parallel builds I would add the flag -O (or -Otarget which is the default for -O) to avoid the build output getting mixed up.
| :: | ||
|
|
||
| cd bin/RTEMS-beatnik | ||
| scp -v libComTestHarness.boot [email protected]:/srv/tftp/ |
There was a problem hiding this comment.
You might not want to publish your internal IP addresses. Oh, and a few lines above s/th/the/.
| ========================== | ||
|
|
||
| All tests successful. | ||
| Programs=37, Tests=3062, 352 wallclock secs |
There was a problem hiding this comment.
Note that the section just above here ("EPICS Test Harness Results") is showing that all the above libCom test programs passed. Unfortunately in the libCom tests we have to run epicsExitTest() afterwards so it isn't obvious that there is a final summary of the results. I agree with Ralph that you could cut out all of the results above and just show this bit (this document is a bit intimidating due to its length). Maybe place the full results on a separate page, and similarly have pages with the results from all the other test harnesses in the other test areas (this was just testing libCom)?
|
Hi, |
4 participants
No description provided.