grails upgrade and app debugging issues

I upgraded to rc4 and groovy 1.5.1 and had some problems developing on my HP laptop. New apps did not run and the class files were all scattered in the top level of the project.

Comcast laptop had racetrack at version from grails rc1 to rc3. I did the grails upgrade. The app still ran np. The Problems view showed the same Unbound claspath variable : GRAILS_HOME problem. I imported the groovy-all-1.0.jar file by "import groovy libs into project". No change.

I added the GRAILS_HOME variable in the project properties Java Build Path dialog (libraries) and it errored out with the same "Variable points to a folder: Use 'Extend...'to select an archive inside the folder." WTF? I can hit cancel, the variable is set, and the app no longer has the "Problems" view issues. I can then do an eclipse clean and the app builds, and all the class files are deposited in the project top level. again WTF?

I played with turning off the java builder and it works and makes changes to the .project file. I changed it back.

When running grails test-app in RC-3 all the tests fail.

I reverted the environ to rc1 and reverted the project to what's in trunk. I ran grails test-app and all but 4 of the tests passed. Not sure what it was before (all passed or what)

Did an experiment: Upgraded the project to rc3 and ran test-app in an rc-1 environment: Won't run because the app and environ grails version's don't match!! Did another grails upgrade from rc1 (app was rc3) and it downgraded the app! Tests (except 4) ran successfully.

Upgraded back to rc3 (app) and ran it in rc3 (environ) and the tests (cept 4) passed. Wtf.

Will experiment with TestGrails app.

trying to fix grails groovy tests in grails apps

Most of the grails test-app test fail with some odd problems about not being able to find some of the domain classes. The problems view in eclipse says stuff like

Unbound classpath variable: 'GRAILS_HOME/ant/lib/ant.jar' in project 'racetrack-TRUNK' racetrack-TRUNK Build path 1200916693560 24138
...
Unbound classpath variable: 'GRAILS_HOME/dist/grails-core-1.0-RC1.jar' in project 'racetrack-TRUNK' racetrack-TRUNK Build path 1200916693576 24197

When upgrading the groovy plugin I found that I needed testNG so I have added the testNG to my eclipse installation and I could then upgrade the groovy plugin. There's a maven plugin too, and I'll have a look at that too.

So next to fix the Unbound classpath variable problems...

Need to create a variable for GRAILS_HOME and point it to the latest grails version.

Setting the GRAILS_HOME only worked if I had also upgraded he project to RC-4. Have to test this with another app to be sure.

One weird thing is that dozens of .class files get build and put into the root folder. Messy. These are definately created by the eclipse build. If you delete them and then do a clean you get them regenerated.

Also seem to need to configure the .classpath so that all the class files go somewhere other than the root folder...

service initialization gui application

/usr/bin/system-config-services (may not work) or
usr/sbin/serviceconf

To use the maven grails plugin
First create a project (or see how to generate a pom for an existing grails project)
mvn -DgrailsHome=d:\install\groovy\grails-1.0-RC1\ grails:create-app
Then cd to the directory containing the new project and create a pom
mvn -DgroupId=com.summitbid grails:create-pom
There's no longer a run-app target for the mvn plugin but you can test it with:
mvn grails:test-app

LIRC notes

Trying to get LIRC to work on RH Fedora 6 at home.

Compiled 0.8.3 using 0x3f8 irq4 which is com1

[root@localhost lirc-0.8.2]# ls -la /dev/lirc*
srw-rw-rw- 1 root root 0 Nov 18 14:58 /dev/lircd

/bin/setserial /dev/ttyS0 uart none
[root@localhost lirc-0.8.2]# /sbin/modprobe lirc_serial

/sbin/lsmod | grep lirc
[root@localhost lirc-0.8.2]# ls -la /dev/lirc*
crw------- 1 root root 61, 0 Nov 19 06:42 /dev/lirc0
srw-rw-rw- 1 root root 0 Nov 18 14:58 /dev/lircd

start another xterm and watch log
tail -f /var/log/lirc.log

Wed evening ride report

Wednesday Evening Deer Creek Ride Report

---

I managed to get tires on trail by 5:20 last night but due to trailhead confusion I started the ride without my expected riding buddies. They started at the lower lot off of Deer Creek and the Lockheed Martin turn of (I guess). I was on my way down after my second lap when I met them coming up and decided to go back up and ride another lap with them. Never did see Sandan or GotDirt.

I rode up the main trail and did three laps (clockwise) around Red Mesa loop before blasting back down. There were very few people when I started but by the end of the third lap there were quite a few people on the trail heading both up and down. The temperature was perfectly cool up top in the shade, and I was having so much fun ripping down the Red Mesa backstretch, I wanted to keep doing laps. The trail is awesome traction wise but "The Wall" is wicked as usual. Didn't ride Plymouth Mt. Trail on the way down since I spent so much time up top.

Unlike on Tuesday's 6:AM Dakota Ridge ride, it was nice to be able to tailgate party for a bit afterwards. I met some guys with plenty of Blue Moon and patio chairs to ralax in.

The temperature is supposed to push 100 by the weekend. Anyone up for a nice early morning ride (at altitude and/or in the shade) somewhere?

Ride Report

Sunday Morning Three Sisters Ride Report

---

Andy and I made it to the lower lot by about 7:30AM or so and were riding just before 8. The weather was perfect and this early, it wasn't hot yet.

We started across the road up Evergreen Mtn trail east, did the loop up top around Evergreen Mt. and blasted down Evergreen Mt. trail west, dodging bikers, hikers and dogs about every 100 feet.

We crossed the road again at the upper lot and climbed up to the top of the Sisters trail. On the climb up to the Sisters I stalled out on the first relatively easy switchback but then cleaned every remaining tricky rocky obstacle the rest of the way.

After catching our breath at the top we talked about all the injuries from the next section (down the Sisters). Andy told me about Seth Walen's broken collar bone and I reminded him of my nose wheelie off the cliff/lacerated leg incident and his helmet/rock impact crash a couple years back. Despite those lingering memories we rode down pretty fast (Andy did anyway) and then at the bottom we decided to take the Dedisse trail down to the river and back.

The climb up to the the top was a fast zig zaggy obstacle course that was a ton of fun. The descent to the river was a fast obstacle course of other bikers and what I swear was a swarm of Africanized Killer bees. I nearly crashed at high speed while swatting the meanest nastiest one as he stung me on the chest. The guy behind me (who happened to be a pharmacist) advised me to get down to the lake "before you go into anaphalactic shock" because he said he was fresh out of whatever the stuff is that you use to counteract bee stings.

Since I am apparenly not in mortal danger of bee sting poison, I survived the trip down to the river and then the (painful, and now pretty hot) climb back up to the top and to the car.

After talking Andy into having a post ride beer, "Geez, it's not even 10:AM!", I was surprised and kinda disappointed to see that my bee sting was hardly noticible. As other cars streamed into the parking lot we chilled with our beers, counted how many people asked us if we were leaving and talked about Andy's sweet Ace Hardware emblazened Jeep Rubicon that we rode up in.

It was a great ride. Check out the attached jpeg of the route.

__________________
Ride dangerously. Take chances.