Forge version 1.3.4

[Chris H.]

Tentative target release date: Friday December 28.

I thought that a 2 week period of time would be advisable this time around. This will give us time to fix a few things and to further play test the new sideboard feature.
 
Friday January 4 is a possibility.

[Agetian]

I think there are several issues currently up that deserve a solution (or at least looking into) before the 1.3.4 beta goes live, at least as long as they're still issues (I can't even compile the latest SVN version anymore, see below, so I can't check the absolute latest svn state):

1. Apparently there's an issue with Draft mode hanging after the draft procedure is over and after the draft is named. It was an issue around 28th-29th Dec and I'm not sure if it's been fixed since but I haven't seen any commits related to that. This is a critical issue, as it makes the Draft mode unplayable. I reported it in the SVN bug reports thread but there has been no response thus far.

2. There's something wrong with getMana - there's a regular NPE on line 239 of ManaPool.java inside getMana in the latest build that I could get to compile (late December, around 28th-29th) when the AI plays cards such as e.g. Coal Stoker. Not sure if it got fixed since (once again, didn't see any related commits go in), I reported it a while ago but also didn't get a response thus far, but anyhow, there's also number 3:

3. There's a test failure related to mana facilities and draft rankings which makes it impossible for me to compile Forge from inside NetBeans (it just says 2 tests failed and stops compiling). I thought it was an issue similar to the one I've had a while ago which required me to clean and rebuild Forge from scratch, but doing that doesn't solve the problem for me this time, it still errors out. Also, it *might* be the thing that broke the auto nightly builder (there hasn't been a build since Dec 26, and unless it's a planned thing with the auto-build being disabled for the time being, I picture it's not impossible that it's just trying to compile but receiving the same test failure that I do). Here's the exact report:

Code: Select all

-------------------------------------------------------
 T E S T S
-------------------------------------------------------
Running TestSuite
Configuring TestNG with: org.apache.maven.surefire.testng.conf.TestNGMapConfigurator@f221f6
[white, red, blue]
failed : 0.9
failed : 0.91
failed : 38.0
failed : 41.0
failed : 42.0
failed : 43.0
ManaCost : addMana() error, mana not needed - 1
failed : 45.1
4
Tests run: 9, Failures: 2, Errors: 0, Skipped: 0, Time elapsed: 7.702 sec <<< FAILURE!

Results :

Failed tests:   testPayManaCost(forge.card.mana.ManaPartTest)
  test(forge.ReadDraftRankingsTest)

Tests run: 9, Failures: 2, Errors: 0, Skipped: 0


(this is after a full refresh of the local SVN copy with full clean+rebuild).

Part 3 is probably not critical for the end user (as long as it's not trying to point out some insidious bug, of course - not sure what this test is pointing at, exactly; and anyhow, I'd love to know how to get this fixed so that I can get back to developing again), part 1, if it's still the case, is a critical bug (Draft mode is unplayable), and part 2, while not super critical, makes me forfeit about 10% of the games in the latest SVN I could get to compile because in them, at least some card NPEs in getMana. :\

- Agetian

[ArsenalNut]

2. There's something wrong with getMana - there's a regular NPE on line 239 of ManaPool.java inside getMana in the latest build that I could get to compile (late December, around 28th-29th) when the AI plays cards such as e.g. Coal Stoker. Not sure if it got fixed since (once again, didn't see any related commits go in), I reported it a while ago but also didn't get a response thus far, but anyhow, there's also number 3:
- Agetian

In r18856, Max mtg checked in some changes to ManaPool.java. The log message was "fixing a crash when AI had spare mana after playing a spell." Have you seen the NPE error since this update? Line 239 in the current ManaPool.java is now a comment line.

[Chris H.]

3. There's a test failure related to mana facilities and draft rankings which makes it impossible for me to compile Forge from inside NetBeans (it just says 2 tests failed and stops compiling). I thought it was an issue similar to the one I've had a while ago which required me to clean and rebuild Forge from scratch, but doing that doesn't solve the problem for me this time, it still errors out. Also, it *might* be the thing that broke the auto nightly builder (there hasn't been a build since Dec 26, and unless it's a planned thing with the auto-build being disabled for the time being, I picture it's not impossible that it's just trying to compile but receiving the same test failure that I do). Here's the exact report:

Code: Select all

-------------------------------------------------------
 T E S T S
-------------------------------------------------------
Running TestSuite
Configuring TestNG with: org.apache.maven.surefire.testng.conf.TestNGMapConfigurator@f221f6
[white, red, blue]
failed : 0.9
failed : 0.91
failed : 38.0
failed : 41.0
failed : 42.0
failed : 43.0
ManaCost : addMana() error, mana not needed - 1
failed : 45.1
4
Tests run: 9, Failures: 2, Errors: 0, Skipped: 0, Time elapsed: 7.702 sec <<< FAILURE!

Results :

Failed tests:   testPayManaCost(forge.card.mana.ManaPartTest)
  test(forge.ReadDraftRankingsTest)

Tests run: 9, Failures: 2, Errors: 0, Skipped: 0


(this is after a full refresh of the local SVN copy with full clean+rebuild).

Part 3 is probably not critical for the end user (as long as it's not trying to point out some insidious bug, of course - not sure what this test is pointing at, exactly; and anyhow, I'd love to know how to get this fixed so that I can get back to developing again), part 1, if it's still the case, is a critical bug (Draft mode is unplayable), and part 2, while not super critical, makes me forfeit about 10% of the games in the latest SVN I could get to compile because in them, at least some card NPEs in getMana. :\

 
This one appears to be now fixed. Try updating to the head version and see if the project will now compile under netbeans.

The Maven commands that I use to do a build and deploy to cardforge ends prematurely when the tests fail.

So I guess that only numbers 1 and 2 need to be addressed before the beta release.

[Agetian]

Yeah, compiles fine now! Thanks, Chris!

UPDATE: The Draft mode hanging is still an issue in the latest SVN, just confirmed after compiling the latest head revision.

In r18856, Max mtg checked in some changes to ManaPool.java. The log message was "fixing a crash when AI had spare mana after playing a spell." Have you seen the NPE error since this update? Line 239 in the current ManaPool.java is now a comment line.

I'll keep an eye out on that, like I said, I haven't recompiled in a while due to technical reasons which are now fixed, now I managed to compile the head revision, so if it happens again, I'll either fix it myself or report it if I can't.

- Agetian

[Sloth]

I've tried to fix the problem with draft mode, but i have to give up for today.

The error is happening in the buildDeck() function of the LimitedDeck class. The list of on-color cards "colorList" is always empty, which causes some infinite loops in later functions, because the decklist will always stay empty.

[Sloth]

r18920: Draft mode is fixed.