Maven build file - generation of OSX, Windows, Linux files

[Chris H.]

Please do an svn update. I fixed the pom.xml yesterday. It "works for me" but let me know. The jyield.jar os the new dep for braid's code. I uploaded it into my Maven repo I use to support Forge.

On another front - I changed the Maven build so that it zips up the cardsfolder and includes only the zip file. I also am not including res/lib since those files are bundled in the jar-with-dependencies. This results in a build time that is *much* faster (~x4 for me) and *much* smaller release packages (50% of the previous size).

`
I updated to the current head rev 10272 and I am still getting the same error:

Code: Select all

[INFO] Scanning for projects...
[INFO]                                                                         
[INFO] ------------------------------------------------------------------------
[INFO] Building Forge 1.0.15-SNAPSHOT
[INFO] ------------------------------------------------------------------------
[WARNING] The POM for com.google.code.jyield:jyield:jar:1.0-SNAPSHOT is missing, no dependency information available
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 0.688s
[INFO] Finished at: Fri Jul 01 15:28:38 EDT 2011
[INFO] Final Memory: 3M/81M
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal on project forge: Could not resolve dependencies for project forge:forge:jar:1.0.15-SNAPSHOT: Could not find artifact com.google.code.jyield:jyield:jar:1.0-SNAPSHOT -> [Help 1]
[ERROR]

[jendave]

Please do an svn update. I fixed the pom.xml yesterday. It "works for me" but let me know. The jyield.jar os the new dep for braid's code. I uploaded it into my Maven repo I use to support Forge.

On another front - I changed the Maven build so that it zips up the cardsfolder and includes only the zip file. I also am not including res/lib since those files are bundled in the jar-with-dependencies. This results in a build time that is *much* faster (~x4 for me) and *much* smaller release packages (50% of the previous size).

`
I updated to the current head rev 10272 and I am still getting the same error:

Code: Select all

[INFO] Scanning for projects...
[INFO]                                                                         
[INFO] ------------------------------------------------------------------------
[INFO] Building Forge 1.0.15-SNAPSHOT
[INFO] ------------------------------------------------------------------------
[WARNING] The POM for com.google.code.jyield:jyield:jar:1.0-SNAPSHOT is missing, no dependency information available
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 0.688s
[INFO] Finished at: Fri Jul 01 15:28:38 EDT 2011
[INFO] Final Memory: 3M/81M
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal on project forge: Could not resolve dependencies for project forge:forge:jar:1.0.15-SNAPSHOT: Could not find artifact com.google.code.jyield:jyield:jar:1.0-SNAPSHOT -> [Help 1]
[ERROR]

please update one more time (r10274) and try it again. I enabled the use of SNAPSHOT jar dependencies (i.e. jyield)

[Chris H.]

please update one more time (r10274) and try it again. I enabled the use of SNAPSHOT jar dependencies (i.e. jyield)

`
I updated and tried again and it works. Thank you once again.

[Braids]

do deprecation warnings mess up the maven build? or any other compiler warnings, for that matter? it seems to be rather sensitive.

[Chris H.]

I am working on another build and release. In rev 10391 Dave moved main into the unit test. During the build and release there was a pause with a test window with the "Error" word visible.

We are now at a later pause and it is generating a checkstyle report, and there are 35686 checkstyle errors.

I see that we are generating a number of reports. We have 882 warnings. I think that this will finish and a working archive will be uploaded to googlecode cardforge.

It is interesting to watch the data in the terminal window.


EDIT:

It looks like the two archives are now on googlecode cardforge. Various reports are now being sent which I assume is related to the recent tests that were added to forge.


EDIT 2:

Rev 10396 [maven-svn-wagon] site

[jendave]

do deprecation warnings mess up the maven build? or any other compiler warnings, for that matter? it seems to be rather sensitive.

No. Neither deprecations nor warnings will mess up the maven build.

Dave

[friarsol]

Hey Dave, did you see this Issue posted? http://code.google.com/p/cardforge/issues/detail?id=150

I know you mentioned something about ant earlier but I can't seem to find it now.

[jendave]

Hey Dave, did you see this Issue posted? http://code.google.com/p/cardforge/issues/detail?id=150

I know you mentioned something about ant earlier but I can't seem to find it now.

Applied and closed. Thanks

[Chris H.]

OK, we have a new beta out. I noticed that the archives a somewhat bigger than before.

During the tests that are now a part of the Maven build process I get an error exception window. A choose dialog window comes up and asks me to select from two choices. One labeled "a" and the other "b".

I look at this in a puzzled state for a couple of seconds and these windows close and the Maven build process continues. I guess that the results of these tests will appear on the site that you mentioned a couple of weeks ago. Wish I had bookmarked that site.

In case anyone is interested, here is the console text from the Maven script run:

[jendave]

OK, we have a new beta out. I noticed that the archives a somewhat bigger than before.

During the tests that are now a part of the Maven build process I get an error exception window. A choose dialog window comes up and asks me to select from two choices. One labeled "a" and the other "b".

I look at this in a puzzled state for a couple of seconds and these windows close and the Maven build process continues. I guess that the results of these tests will appear on the site that you mentioned a couple of weeks ago. Wish I had bookmarked that site.

In case anyone is interested, here is the console text from the Maven script run:

I took all of the "tests" from the course, which were simply main() methods buried in the src, and copied them into "test/java". I know very quick and dirty. Since many were basic gui tests, I have them timeout so you do not have to interact with them. I also disabled tests that try to load the entire universe of cards. Hopefully, I can take a look at the tests soon to clean them up.

[Chris H.]

As long as the Maven script results in a successful build and release then I am not too concerned. This test process is another example of a work in progress. It is fun to watch and it occupies me while I wait for the process to complete.

Can you give us that URL for the test results again? At least I will bookmark it this time.

[Rob Cashwalker]

Chris, as far as archive size, there are a few items causing the inflation:

The foil overlays added 2.9 MB... but they look so good...

There's a file mtg-data.txt. I'm guessing it's used by one of Sol's python scripts, but we probably don't need to include it in the beta release.

I think we can eliminate the pics_link folder, since that functionality has been removed.

[jendave]

As long as the Maven script results in a successful build and release then I am not too concerned. This test process is another example of a work in progress. It is fun to watch and it occupies me while I wait for the process to complete.

Can you give us that URL for the test results again? At least I will bookmark it this time.

I added links to the wiki.

http://www.slightlymagic.net/wiki/MTG_Forge#Development_Reports

[Chris H.]

Chris, as far as archive size, there are a few items causing the inflation:

The foil overlays added 2.9 MB... but they look so good...

There's a file mtg-data.txt. I'm guessing it's used by one of Sol's python scripts, but we probably don't need to include it in the beta release.

I think we can eliminate the pics_link folder, since that functionality has been removed.

`
Besides the foil overlays I think that there may have been some increase in the lib folder. And I am not about to suggest that we remove the foil overlays, they are a nice addition.

There may be several old files located in the draft folder that are no longer of use.

[Chris H.]

I added links to the wiki.

http://www.slightlymagic.net/wiki/MTG_Forge#Development_Reports

`
Thank you Dave, those links on the first wiki page are helpful.