1st
had to revise the syntax for my "type:" variable it is now type:blah:location, using the pipe line created nasty returns in some combinations of abilities as the parser confused the pipeline in type: with the one in TC.
2nd
more of a correction then a bug fix, i noticed that upcost had instances where it could still trigger its effect in the draw phase becuase the condiational for the ability resolve was in the update, i left the class mostly untouched but move the resolve out of update and into a receiveEvent for upcost, which triggers only once after you declared that you are leaving the upkeep before draw step begins.
3rd
reworked the construction of multiability to only produce a single multiability with all the abilities in the string instead of nested multiabilities, it is considerabily easier to debug and creates 1/3 less objects for the engine to deal with in comparison to the old method.
also noticed issues with transforms parsing which i was able to easily correct by making it parse transforms in double brackets, transforms((effects))...updated primitive in next revision.
test suite passes, and varified that the fix doesn't effect how "&&" ability works with transforms.
added newability[whatever ability] to transforms classes.
this allows for transforms to also add activated or mtgabilities to cards it targets.
example Prismatic omen:
auto=lord(land|myBattlefield) transforms(plains forest mountain swamp island,newability[{t}:add{w}],newability[{t}:add{g}],newability[{t}:add{r}],newability[{t}:add{b}],newability[{t}:add{u}])
Issue: 573
lets say in the case of Isleback Spawn you want to see if ANY player has less then 20 cards.
auto=aslongas(*|library) 4/8 <21 compare
this will check if either you or your opponent has less then 20 cards. rather then adding the total cards in both libraries.
over the last weeks ive seen this cards code change atleast 3 times. this should be the final version for it and any other cards like it.
added parsing to correct this bug,
cards which state "card from target players hand" should be coded "target(*|hand)" which allows targetting of any cards which are in a hand not just your hand.
|myhand)
|opponenthand)
|hand)
self explainitory.....
while i was at it reworked the aslongas class completely, i promised myself if another bug popped up with aslongas i would rewrite it. 20 hours later, we have an aslongas which does not care about when cards come or go.
aslongas should never have used the added/removed to determine the validity of its effect. i allow update to sort it out now in the update function of aslongas...also added a new function to it called sorterfunction tho it looks as it could be reduced in code i would appreciate that it remains as it is, simply because of the nightmare i went through fixing the previous 7 bugs or more reported for aslongas. the logic was far to jumbled up. this i hope is atleast cleaner and easier to understand for the future devs that might work on it.
note: all test passed, all manual testing also passed.
new while tag is used as follows.
[card]
name=Emperor Crocodile
auto=aslongas(other creature|myBattlefield) all(this) sacrifice while <1
text=When you control no other creatures, sacrifice Emperor Crocodile.
mana={3}{G}
type=Creature
subtype=Crocodile
power=5
toughness=5
[/card]
this will denote that the ability is NOT a oneshot ability and should remain effective for the entire time the card is in the field.
Issue: 613
- Updating version number for upcoming revision
- Note: test Thallid.txt fails, we need to investigate when this happened and fix asap, will open a ticket.
i was not 100% successful in getting the correct cost on altercost which reduced the amounts. however altercost which increased the amounts are now clean code. so the reducers are still using a dirty method of putting the card from the hand back into the hand to get the correct amounts.
the issue i ran into is how to get the correct amount in this situation.
target cost {5}
have 3 reducers that - {2}
thats {1} of overkill reduce...
heres the issue. now remove all the reducers from play. the cost returns to {6} instead of 5.
you would think, will have it track the over kill. which leads to another. have the 3rd card which would contain the overkill removed first, i factor the overkill and increase the amount by what was reduced, but that would not be correct. since the reduce was {2} the increase of removing the overkill card should increase the targets cost by {2} not the amount it actually reduced. as you can see, it is extremely complex. so for now im leaving it dirty, but removing the resetcost code.
- added 2 menus to wait for connection and wait for deck selection
- tested compilation on Qt Linux, Qt Windows and PSP
- deactivated everywhere (NETWORK_SUPPORT to activate).
exsample
[INIT]
mode=mtg
[PLAYERS]
life:2000
poisoncount:4
customphasering:untap,firstmain,firstmain,draw,firstmain,upkeep,draw
auto=shuffle
auto=draw:7
auto=@each my draw:draw:1
auto=maxPlay(land)1
by default every turn will have normal order we've always been use to
however
before begins --required added automatically
anything you want in any order
anything you want in any order
anything you want in any order
anything you want in any order
anything you want in any order
anything you want in any order
anything you want in any order
anything you want in any order
cleanup---required added automatically
after end of turn---required added automatically
svntax
customphasering:blah,blah,blah
listed in the order you want it to go. you can repeat a phase as often as you want, you can omit phases.
please note, this was not designed with MTG in mind..some of the rules require certain phases to work...this is for CUSTOM game building...i repeat NOT DESIGNED FOR MTG...tho it can be used with most(almost all, with the exception of those which require certain phases to work, exsample, ninjitsu will only work in a phasering which contains blockers and attackers)
enjoy.....
-fix for issue 604 (Land play limitation should not apply in all cases)
-- this adds a "castMehod" variable to MTGCardInstance. IF this variable is 0, the card was not "cast" (or for lands, "put into play" as part of the lands rule), but "added" to the battlefield with some other effect. On the other hand, if this variable is set, it means the card was cast
-- as we discussed, I did not touch the "alternateCostPaid" variable, as I'm still not really sure these two concepts are actually the same
case changes, "Time" != "time" aparently our parser now strings all counters as lower case...so "Charge" is no longer string "Charge" and can not be compared to "Charge"..not sure if that makes any sense :) anyways removel of time counters on exiled cards by triggers or effects now happen correctly.
added some extra checks to how Ai determines if it should cast a "all(" or other lord...it was only taking the amount of cards it had into account and not looking at what the suggestion was for the nestedability...meaning it treated "destroy all" the same way it treated "all creature +10/+10" it would only check if it had more cards effected...completely ignore the effects....
Thus, Construct thought the string length was 1 greater than it should have been causing an out of bounds runtime exception. In VS2008, the same thing
happened, however no exception was thrown since VS2008 allows access into the undefined memory space and returned a 0 (null) value when it tried to access
outside the string array. This fix targets specifically the color filters since they seem to be the ones with the error. " color:g;" and not "color:g;" was being
passed.
- fixed issue 595 (MaxCast not working). Root cause was the stupidest typo ever in the code (maxCost instead of maxCast...)
- fixed an issue with Storm mentioned by zethfox (was counting only spells played by current player)
- Moved "max lands per turn" rule outside of the code (yay!) Please be sure to update your Rules folder!
but only by chance that the previous step set the current deck file. Used victory percentage
from the MetaData object rather than forcing what has already been computed.
TODO: getVictoryPercentage() needs to represent Overall VictoryPercentage for all decks, not just
player decks. AI Decks use getVictoryPercentage() to represent the VictoryPercentage against a
specific deck. There should be an additional method that says getVictoryPercentageVsDeck( deckObject )
- submit a fix for TestSuiteAI, I don't understand why it wasn't submitted with my other changes yesterday. This fixes a bug in the test suite (One storm test would fail without this)
- minor refactor to AI code
* getCardByName seemed to have a initialization error if you tried to use it before the entire game loaded. The cache would
throw an exception if you tried to use find and it was empty. I put a guard around it to avoid this issue.
* refactored Zeth's "toggledifficuly" feature to be stored in meta data. http://code.google.com/p/wagic/source/detail?r=3106
-- This is slightly modified as it forces a 1 for 1 swap of cards that are specified.
from the example given this is how it seemed to be used anyways.
-- since all the information is stored in the meta data, there's no need to alter the deck's
definition.
For some reason, the VS2008 compiler allows this illegal access and VS2010 does not. In all actuality it should never be permitted.
Please review the thread posted on the SVN thread for details as to why this needs a fix.
http://code.google.com/p/wagic/issues/detail?id=593&sort=-id
Issue: 593
wagic.the.homebrew@gmail.com
omegablast2002@yahoo.com
Xawotihs
techdragon.nguyen@gmail.com