i seriously think we devs should consider making all ability coding required lowercase. not card names and text, but coding such as targetchoosers "thisForEach(creature[Black;White]|myBattlefield) moveto(myGraveYard) ...<---this is not uniform...and leads to people using such coding in the source ....i think i suggested this once and was instantly shot down...however, this is the 2nd instance where a bug has arose from the comparing of capitalized vs lower case..becuase we all assume that every line is going to be dropped to lower case as the game reads them.....
-- Zeth, please verify that I didn't break your mods if you have time
-- The test suite does NOT pass in this revision (see issue 711), but I verified that the problem does not come from my change, so we need to dig
-- Campaigns creators will have to update their campaigns to make them compatible with Wagic 0.16. It consists in adding "include mtg.txt" at the top of each one of their rules.txt file.
This helps a lot in the deck selection menus (duel and deck editor).
That, added to my previous fix, fixes issue 693 for me.
In general, we should try to avoid 512x512 32bits textures on the PSP. That happens with a 480x272 png image. One solution is to convert the file to a jpeg, and load it in 551 mode whenever possible. Another solution is what I did: use a 450x255 image instead. The best is of course to cumulate both, which divides the size in Ram by 4 on the PSP.
I'm ok with distributing lower res graphics on the psp, and higher res on other platforms, if that helps. Main culprits are 480x272 images (when there's too many of them)
The BIG CARD SAFARI 2011 is over: After 4 months of intensive testing, i can proudly present you this huge card package. All these cards are working 100%.
Card list --> First comment.
- Added 20 tests to save the cards whose code is not covered by already existing tests.
- Added Magic2012 (M12) with 75 completely new cards.
-- Res is the default folder in Res.txt, in case it exists, to avoid confusion when debugging
-- added sign_prx in the makefile, harmless if you have the default build.mak, but signs the app for you if you have the correct files
-- uninitialized variable warning fixed in Rules.cpp
-- The problem was that the previous fix for keybindings actually fixed dead code for the PSP version... the real call to Defaultbindings was still happening in JGE's code...so the fix made the call twice.
-- Note: shouldn't we use ResetBindings instead of JGECreateDefaultBindings? The former resets the bindings before assigning the default ones...
http://code.google.com/p/wagic/issues/detail?id=706
these types were converted into part of the "types" enum and are no longer considered a subtype in the engine.
as they appear before Subtypes::LAST_TYPE they are to return "" as their subtype search result.
i figured there was no harm in including them in the types list filtering as they were moved from subtypes becuase they were some of the most frequently used subtypes.
otherrestriction=type(creature|mybattlefield)~lessthan~type(creature|opponentbattlefield)-3
which means, if the statement above is true you can play the card....if you subtract 3 from the amount the opponent has in creatures, and you are still less then the opponent...then that means he has atleast 4 more creatures than you do.
so right after the type(blah) you can add +number or -number...and it will modify the amount to compare.
so i converted it from string comparisons and counting the battlefield..into targetchoosers with declarable operator.
simplified the code from the mess it was before into something alot easier to card code with.
example before:
mytypemin:less type(land),opponenttypemin:* type(land)
becomes
type(land|mybattlefield)~lessthan~type(land|opponentbattlefield)
the new syntax is:
type(targetchooser)~operator~number
type(targetchooser)~operator~type(targetchooser)
the operator are:
morethan
lessthan
equalto
1) prevent the textureQuad map from containing references to cards when we're returning a temp (card backdrop) image in the threaded version.
2) we had two copies of the pspicon image in the cache (ie the regular & managed). Changed the reference in the DeckMenu code to use the managed instance.
3) Fixed a problem with a default constructor param that would set RESOURCE_LOCKED on certain images. I noticed that the 'NEW' image was locking itself in the cache because no resource modes was being passed in, so the default constructor param was setting it to locked. This seems like a bad idea - if you're not explicitly locking a resource in the call, that shouldn't be the default behaviour.
Also added an assert safeguard to catch the possibility of 'overlocking' an image resource, as this probably means the image fetching routine is being abused somehow.
The problem was, the game would initialize first, reading the
keybindings from the configuration, and THEN set up the default key
bindings.
I think this affects only PSP, X, SDL and Qt. A quick scan of the
windows version code looks to me that the order is correct and the
problem should not exist, but I may be mistaken. The iPhone version
looks to me like it has no keybindings at all, though the call is
there ; I put the call in the right order so that the problem does
not happen in the future if we implement some key bindings for it.
If someone can check for windows...
