Forge version 1.5.28

[Chris H.]

Tentative target release date: Friday September 26 2014.

[elcnesh]

As discussed in the Network Multiplayer thread, I'm planning to merge the GUI Refactoring branch into trunk soon (probably tomorrow). There may very well still be some display-related bugs, although all basic functionality is in place and a lot has been tested. I'll fix anything that comes up, before the next release.

If you commit to forge-gui* before tomorrow morning (GMT+2) I'll merge it, otherwise the code may be outdated. I'll try and make a brief manual in the wiki so everyone understands how the new code works.

I think it'd also be good to create a snapshot build for the non-developer users to test, to find bugs even quicker.

[Hanmac]

http://cardforge.org/releases/snapshots/forge/forge-gui-desktop/ is gone, where can i access the snapshots now?

[Agetian]

@ elcnesh: sounds good and i agree about a public test snapshot. Good luck!

To whomever may be able to get to the bottom of this: there are some important-looking regressions in 1.5.26 (see my report in the relevant thread), they may also need solving before 1.5.27 by way of a snapshot).
UPDATE: I tried to fix the Sphinx's Revelation bug (not sure if maybe I introduced an even bigger bug with my fix, please review), but not sure what to do with Extort.

- Agetian

[KrazyTheFox]

http://cardforge.org/releases/snapshots/forge/forge-gui-desktop/ is gone, where can i access the snapshots now?

You can find daily (and whenever I feel like generating them) snapshots here: http://www.krazyweb.net/forge

[friarsol]

UPDATE: I tried to fix the Sphinx's Revelation bug (not sure if maybe I introduced an even bigger bug with my fix, please review), but not sure what to do with Extort.

- Agetian

Extort is weird. Sometimes it works and sometimes it doesn't. I just tried casting a spell twice with Extort on the battlefield, one time I gained a life one time i didn't. People have reported issues with it in the past, but each time I could never repro it.

[Agetian]

UPDATE: I tried to fix the Sphinx's Revelation bug (not sure if maybe I introduced an even bigger bug with my fix, please review), but not sure what to do with Extort.

- Agetian

Extort is weird. Sometimes it works and sometimes it doesn't. I just tried casting a spell twice with Extort on the battlefield, one time I gained a life one time i didn't. People have reported issues with it in the past, but each time I could never repro it.

I'm pretty much certain that the Extort problem is related to the "Number$0" expression in the Extort keyword setup (CardFactoryUtil.java:2454). It always seems to return 0 (which seems logical) - is it supposed to be somehow magically specially handled by the game somewhere (but sometimes isn't)? As it works right now, it makes the AI lose the correct amount of life (1), but then it passes "Number$0" to the amount part of the life gain sub ability, which returns 0 and makes the human not gain any life. :\

An absolutely definite way to consistently reproduce this issue is given by me here: viewtopic.php?f=26&t=15333#p160762

- Agetian

[Chris H.]

Forge version 1.5.27 was a test build and deploy that Rob requested last night.

So the next beta and upcoming snapshots should be 1.5.28

[drdev]

Is there a reason we just suddenly skipped 1.5.27? I had already updated the mobile app to use 1.5.27, but then it got bumped to 1.5.28.

At the moment, I just committed changes to make it go back to 1.5.27 so the mobile app works. If there's a reason the version was skipped, let me know and I'll change it back and update the mobile app.

EDIT: Off by one minute, lol. At any rate, do we really have to skip 1.5.27? I think that's confusing for our end users.

[Chris H.]

The build and deploy Maven command makes changes to both slightlymagic and to cardforge.

If we change the poms back to 1.5.27 we might cause a problem when I run the next beta build and deploy.


It might be safest to revert the poms back to 1.5.28 and then to add a brief expanation about 1.5.27 to CHANGES.txt. We had to do something similar on a couple of past occassions.

[drdev]

The build and deploy Maven command makes changes to both slightlymagic and to cardforge.

If we change the poms back to 1.5.27 we might cause a problem when I run the next beta build and deploy.


It might be safest to revert the poms back to 1.5.28 and then to add a brief expanation about 1.5.27 to CHANGES.txt. We had to do something similar on a couple of past occassions.

Ok, I'll bump it back to 1.5.28. I'll let you add the explanation to CHANGES.txt.

[Agetian]

Ok I think I got Extort fixed up in r27338, the amount of life lost was not propagated in time to the AF LifeGain subability. Looks fine to me now (unless I just broke it some place else ).

- Agetian

[drdev]

Ok I think I got Extort fixed up in r27338, the amount of life lost was not propagated in time to the AF LifeGain subability. Looks fine to me now (unless I just broke it some place else ).

- Agetian

Looks like you didn't need my help with Sphinx's Revelation or Extort. Nice work.

[Agetian]

Ok I think I got Extort fixed up in r27338, the amount of life lost was not propagated in time to the AF LifeGain subability. Looks fine to me now (unless I just broke it some place else ).

- Agetian

Looks like you didn't need my help with Sphinx's Revelation or Extort. Nice work.

Hehe thanks, yep I got to the bottom of it. Thanks for all your hard work on everything, too!

- Agetian

[drdev]

So I just committed a decent-sized refactoring of the logic used for AI mana payment (and the Auto button) in r27348. People were reporting issues with cards being sacrificed for mana ahead of lands being used, which wasn't right.

The new logic adjusts the scoring logic added previously for cards to increase the score based on the number of non-mana activated abilities available to the card, whether the card can attack or block, additional costs associated with the mana ability (such as paying life for painlands), and most importantly, a large score increase if the mana ability is not undoable (such as a sacrifice effect). That way, mana abilities that are non-undoable should be used as an absolute last resort.

Please let me know if this logic or the specific score adjustments I'm using need to be tweaked.

Thanks.
-Dan