#gemrb@irc.freenode.net logs for 16 Apr 2013 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:05:52] <-- edheldil has left IRC (Ping timeout: 258 seconds)
[00:46:16] --> brada has joined #gemrb
[01:17:10] <-- brada has left IRC (Quit: brada)
[02:25:20] --> brada has joined #gemrb
[04:42:38] <-- brada has left IRC (Quit: brada)
[07:45:06] --> edheldil has joined #gemrb
[07:45:07] --- ChanServ gives channel operator status to edheldil
[08:14:13] --- ermo^ is now known as ermo
[08:47:50] <-- WingedHussar has left IRC (Quit: WingedHussar)
[12:57:26] --> brada has joined #gemrb
[13:17:20] <Seniorita> [commit] bradallred: iOS: remove _DEBUG flag from build. https://github.com/gemrb/gemrb/commit/6f958cd1566752b6b9a4d43678b303b15c26ee04
[14:06:30] --> WingedHussar has joined #gemrb
[14:59:38] <-- WingedHussar has left IRC (Quit: WingedHussar)
[16:16:52] --- ermo is now known as ermo^
[16:28:48] --> edheldil_ has joined #gemrb
[16:33:48] <-- edheldil_ has left IRC (Ping timeout: 264 seconds)
[16:35:15] --> fizzle has joined #gemrb
[18:17:47] --> edheldil_ has joined #gemrb
[18:33:39] <-- Cable_ has left IRC (Ping timeout: 258 seconds)
[18:46:58] --> Yoshimo has joined #gemrb
[18:47:09] --> Cable_ has joined #gemrb
[18:57:27] <Seniorita> [commit] fizzet: print warning when stream cannot be opened, too https://github.com/gemrb/gemrb/commit/6543dcdfe114150d9feb879c7d4af7854a30dfd8
[18:57:28] <Seniorita> [commit] fizzet: make stores auto-identify their items if lore is good enough https://github.com/gemrb/gemrb/commit/9dbe7358934623623588ddfe9fec9d6a116a0799
[18:57:29] <Seniorita> [commit] fizzet: fix unsetting the paperdoll flag https://github.com/gemrb/gemrb/commit/aae17ed68cf09ea2b27f01e9b54f7ffdc9c2a424
[18:57:30] <Seniorita> [commit] fizzet: change some attack stances for IE_ANI_CODE_MIRROR_2 https://github.com/gemrb/gemrb/commit/1cea90eecdb9203118bf01cbb702cd65f21ffc43
[19:12:41] --- ermo^ is now known as ermo
[19:15:58] <fizzle> and this is a possible fix for the item identify window error: http://paste.debian.net/hidden/2197e124/
[19:16:04] <Seniorita> Debian Pastezone
[19:16:31] <fizzle> not totally happy about this so happy to hear about alternative options
[19:19:51] <brada> that seems like it might break touch input in some way
[19:20:27] <brada> we should avoid posting fake events to SDL
[19:20:41] <brada> do it to the gemrb event manager if it must be done
[19:21:38] <brada> also you didnt even add that to SDL2 driver so it would totally break the compile anyway :p
[19:21:56] <brada> nevermind that last bit
[19:22:06] <brada> forget about the common inheritance :p
[19:22:30] --> lynxlynxlynx has joined #gemrb
[19:22:30] --- ChanServ gives channel operator status to lynxlynxlynx
[19:22:42] <fizzle> using the EventManager is insufficient in this case
[19:22:55] <brada> why?
[19:23:00] <brada> i guess i should read the note
[19:23:14] <fizzle> because it doesn't have a queue and processes the event *now*
[19:23:27] <fizzle> which is precisely what causes the breakage in the first place
[19:23:48] <fuzzie> don't we have a delayed window open system?
[19:24:16] <fizzle> how would that help?
[19:24:17] <brada> should be scriptable no?
[19:24:45] <fuzzie> oh, then I don't understand the bug
[19:25:07] <fizzle> let me try to explain
[19:25:22] <fizzle> inventory window
[19:25:24] <fuzzie> or rather, I don't see why you're trying to do some event hack
[19:26:06] <fizzle> identify window opens, choose way to identify
[19:26:28] <fizzle> now we close the identify window, then open the item description window
[19:26:33] <fizzle> however
[19:26:55] <fizzle> as soon as the identify window closes focus reverts to the inventory window
[19:27:20] <fizzle> and a fake mouse motion event will make ItemButton change if the cursor happens to be on an inventory button
[19:27:39] <fizzle> only after that do we actually open the item desc window
[19:27:55] <fizzle> possibly showing the wrong item (or maybe none at all)
[19:28:03] <fuzzie> right
[19:28:12] <fuzzie> so, bug: ItemButton is not preserved
[19:28:15] <fuzzie> fix: preserve ItemButton?
[19:28:43] <fizzle> also possible, but not very nice either
[19:29:00] <fuzzie> well, nicer would be just passing the right item to the new window
[19:29:16] <fizzle> because OpenItemWindow is used as an EventHandler and thus called without parameters
[19:29:30] <fuzzie> there's nothing stopping you from passing an optional parameter to it, though
[19:29:41] <fizzle> yes
[19:30:21] <fizzle> but using params some times and not other times is what I'd call "not very nice either"
[19:30:26] <fizzle> "icky", too
[19:30:41] <fuzzie> well
[19:30:43] <brada> ouldnt you make a diffrent eventhandler that will call OpenItemWindow with a parameter for this one case?
[19:30:50] <fuzzie> that sounds horrible.
[19:31:26] <fuzzie> this sounds like exactly what default parameters are useful for.
[19:32:01] <fuzzie> you do 'def OpenItemInfoWindow(ItemSlot = None):' and fish it out of a global variable if there wasn't one passed.
[19:32:37] <fizzle> sure, can do
[19:33:00] <fuzzie> i mean, i'm just giving a viewpoint here :-p
[19:33:05] <fizzle> there's just too much state keeping code around as is
[19:33:12] <fizzle> but maybe that's just me
[19:33:25] <fuzzie> well, yes
[19:33:40] <fuzzie> the gui params are meant for gui state
[19:33:43] <fizzle> and, obviously, I haven't been able to come up with a nice solution yet
[19:33:46] <lynxlynxlynx> or add another var that is only set in the identify window and always reset in the description one
[19:33:58] <fuzzie> there's no point in using vars for things which aren't live gui state
[19:34:16] <fuzzie> that just adds complexity
[19:34:19] <fizzle> lynx: I hope that wasn't serious...
[19:34:31] <fuzzie> well, I was tempted to make that suggestion too
[19:34:40] <brada> better than phony events :p
[19:34:43] <fuzzie> i mean, if you're prepared to insert *fake events* into SDL, you are much much crazier than that :)
[19:35:01] <fizzle> heh, maybe
[19:35:18] <fizzle> I mean, the real fix would be not requiring fake events for cursor updates
[19:35:33] <fizzle> but that's also much more intrusive and more work, obviously
[19:35:40] <brada> not worth it imo
[19:36:11] <brada> ive come to realize plenty of other software does the same type of thing for updating cursor too
[19:36:45] <fizzle> that's never a good reason
[19:36:54] <brada> what isnt?
[19:37:04] <brada> the fact that its a lot of time and little payoff?
[19:37:09] <brada> i think thats a fine reason
[19:37:13] <fizzle> "other software also does it"
[19:37:15] <brada> oh
[19:37:22] <brada> thats not a reason at all :p
[19:38:04] <brada> just pointing out that its often the route that is taken because others involve too much work
[19:38:36] <fuzzie> in any case this is the correct behaviour
[19:38:53] <fuzzie> if you close a window then the gui state should change to whatever's below the window
[19:39:14] <fuzzie> so you'd still have this bug if you correctly updated the focus-change code
[19:41:01] <fizzle> depends, but I'm not going to try that anyway
[19:42:05] <fizzle> so the easiest solution without injecting fake events at the SDL level is certainly the default param
[19:58:42] <Seniorita> [commit] fizzet: directly pass the slot number to OpenItemWindow where possible https://github.com/gemrb/gemrb/commit/596bcde85950cd7ece2eb09782b3bcf57e00afd7
[20:15:38] <-- fizzle has left #gemrb
[20:33:26] <-- Yoshimo has left IRC (Quit: Yoshimo)
[21:11:55] <brada> fuzzie: did you ever get around to poking SDL people about 16bpp textures?
[21:14:55] <fuzzie> not at all
[21:29:53] <DrMcCoy> fuzzie: While you're at it, also tell them that they're damn stupid for removing AudioCD support from SDL 2
[21:31:16] <DrMcCoy> Preferably with an appropriate amount of yelling
[21:41:49] <brada> whats an audio CD?
[21:41:58] <brada> :p
[21:43:48] <DrMcCoy> Right, SDL should also support cassette decks and vinyl players
[21:58:00] <fuzzie> i'd settle for it not being broken in a variety of stupid ways
[22:11:45] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:34:07] <-- edheldil_ has left IRC (Ping timeout: 264 seconds)
[22:35:23] <-- brada has left IRC (Quit: brada)
[23:49:57] --- ermo is now known as ermo^