#nuvie@irc.freenode.net logs for 16 Aug 2012 (GMT)

Archive Today Yesterday Tomorrow
Nuvie homepage

[00:13:24] --> Kirben has joined #nuvie
[00:19:54] <CIA-75> nuvie: 03efry * r1517 10/nuvie/trunk/ (5 files in 2 dirs):
[00:19:54] <CIA-75> nuvie: * Remove Gargoyle egg when freeing shrine. bug [3556852]
[00:19:54] <CIA-75> nuvie: * Play magic effect when removing force field on shrine.
[00:19:54] <CIA-75> nuvie: * Removed usecode for snake venom. It wasn't in the original
[00:27:55] <Malignant_Manor> The spell book gump no longer works. I wonder if it is from my mouse changes in map window.
[00:30:05] <Yuv422> it works in keyboard mode
[00:30:20] <Yuv422> oh you're talking about the fullmap mode gump
[00:30:29] <Malignant_Manor> Yeah
[00:30:52] <Yuv422> yeah the original doesn't work with the mouse either
[00:31:18] <Yuv422> it just says cast-nothing
[00:31:29] <Malignant_Manor> The original spellbook works for me with the mouse.
[00:31:56] <Yuv422> strange
[00:32:24] <Malignant_Manor> The scroll wheel, clicking on the map to cast, and clicking on the book work.
[00:32:51] <Malignant_Manor> Maybe configuration settings are different.
[00:38:28] <Malignant_Manor> I'll finish adding back the check for traps (since it is no longer get usecode) and check it out more.
[00:38:39] <Yuv422> cool
[01:02:48] <CIA-75> nuvie: 03malignantmanor * r1518 10/nuvie/trunk/ (9 files in 3 dirs):
[01:02:48] <CIA-75> nuvie: * Drag and drop from containers on the map to the map (container gump)
[01:02:48] <CIA-75> nuvie: * Fix 1501 regression where targeting a container with move moved to
[01:02:48] <CIA-75> nuvie: the actor holding the container instead
[01:02:48] <CIA-75> nuvie: * Search vortex cube
[01:02:48] <CIA-75> nuvie: * Fix check for getting traps now that they don't have get usecode
[01:06:25] <CIA-75> nuvie: 03malignantmanor * r1519 10/nuvie/trunk/ChangeLog: ChangeLog
[01:37:26] <Malignant_Manor> It was the keybinding changes.
[01:37:54] <Malignant_Manor> I'll have to look at it after I get done eating. I know the spell system is a bit of a hack.
[01:43:19] <Yuv422> yeah it's a big hack ;)
[01:43:36] <Malignant_Manor> I think it probably has to do with the simulated return.
[01:43:48] <Yuv422> yeah
[01:51:55] <Malignant_Manor> Svn access is being real slow.
[01:52:33] <Yuv422> that's always frustrating
[01:53:40] <Malignant_Manor> Yup, simulated return is the issue.
[01:56:38] <-- Dominus has left IRC (Read error: Connection reset by peer)
[01:57:24] --> Dominus has joined #nuvie
[02:02:55] <Malignant_Manor> Weird, the simulated return works when I put it as a function.
[02:03:37] <Malignant_Manor> But you can't get it to work at all so maybe it is some compiler magic.
[02:05:25] <Malignant_Manor> Peer doesn't cancel in fullscreen_map.
[02:17:42] <Yuv422> I'm just looking at the multi-tile actors leaving bits behind
[02:19:52] <Malignant_Manor> The gump button works for me even though it does the same thing.
[02:26:01] <Malignant_Manor> Damage tiles still damage when a character is blocked and tries to move.
[02:26:32] <Malignant_Manor> Party members still seem to be hit more than once.
[02:26:49] <Malignant_Manor> A lot of the time at least.
[02:28:42] <Malignant_Manor> Does clicking on the check mark still close the new spell gump on your build?
[02:30:03] <Yuv422> let me check
[02:31:08] <Yuv422> yes
[02:35:25] <Malignant_Manor> Does that patch work for you?
[02:38:01] <Yuv422> let me check
[02:41:27] <Yuv422> yeah that works for me
[02:41:40] <Malignant_Manor> Okay, thanks
[02:42:50] <Malignant_Manor> Maybe someday I will get around to dehardcoding things but stuff like this is definitely a deterrent.
[02:47:25] <CIA-75> nuvie: 03malignantmanor * r1520 10/nuvie/trunk/ (6 files in 2 dirs): *Fix problems with the mouse selecting spells with a fake keyboard event
[03:20:43] <Yuv422> I think I've found the problem with the multi-tile actors losing object tiles
[03:21:23] <Malignant_Manor> Is it from your recent change?
[03:21:37] <Yuv422> no I think it's been there for a long time
[03:21:50] <Yuv422> obj = obj_manager->get_obj(x, y, z);
[03:21:50] <Yuv422> if(obj == NULL || actor_obj_n != obj->obj_n || (obj->quality != 0 && obj->quality != id_n))
[03:22:03] <Yuv422> this code only checks the top object
[03:22:31] <Yuv422> but if two or more actors have object tiles at the same location then it might not return the correct object
[03:22:41] <Yuv422> which would cause it to create another object
[03:23:52] <Yuv422> I should use get_obj_of_type_from_location() instead
[03:28:59] <Yuv422> yeah the code was from commit 403
[03:32:39] <Malignant_Manor> Is there a way to have skiffs be drawn in the same order as regular objects when it is on non-water tiles?
[03:33:11] <Malignant_Manor> Right now it is easy to lose it under things.
[03:33:17] <Yuv422> hmm
[03:33:58] <Yuv422> is that because it is being put at the wrong end of the object list
[03:34:53] <Malignant_Manor> I have no idea what the issue is. It just renders underneath everything.
[03:37:36] <Malignant_Manor> Maybe it is from the movement type.
[03:38:23] <Malignant_Manor> But MOVETYPE_U6_WATER_LOW doesn't bring up any relevant results.
[03:46:37] --> Dominus1 has joined #nuvie
[03:46:37] <-- Dominus has left IRC (Read error: Connection reset by peer)
[04:14:48] <Malignant_Manor> A really high cur_y also causes the roof crash.
[04:19:56] <Yuv422> ah k
[04:21:15] <Malignant_Manor> At least there is the void and no roof areas.
[04:21:46] <Yuv422> yeah I imagine the original devs put that void in to save effort in programming. ;-)
[04:21:57] <Malignant_Manor> I altered the skiff to fly high.
[04:22:14] <Yuv422> did that fix your problem?
[04:22:29] <Malignant_Manor> Mouse movement screws up in places but it seems handy that the balloon doesn't have any mouse controls.
[04:22:35] <Malignant_Manor> No.
[04:22:45] <Malignant_Manor> The render order is still the bottom.
[04:27:55] <Yuv422> I'm just going to commit my fix for the body parts bug
[04:28:36] <Malignant_Manor> There seems to be a crash when double clicking near the wrap location that causes a crash.
[04:29:15] <CIA-75> nuvie: 03efry * r1521 10/nuvie/trunk/ (ChangeLog actors/U6Actor.cpp): * Fixed bug "Dragon body parts left behind" [3556227]
[04:30:02] <Malignant_Manor> Yeah, being in a vehicle screws up the Avatar portrait in the new converse gump.
[04:31:57] <Malignant_Manor> I think the doubleclick crash is only in fullscreen_map mode since the cur_x and cur_y can be higher negatives.
[05:30:54] <Malignant_Manor> The double click crash is caused by tile_is_black.
[05:32:20] <Malignant_Manor> (y - cur_y +1) * tmp_map_width + (x - cur_x +1) produced lovely numbers like -24298 and 1638455
[05:59:17] <-- Malignant_Manor has left IRC (Ping timeout: 268 seconds)
[06:20:30] --> Malignant_Manor has joined #nuvie
[06:56:44] <Yuv422> hmm the original game makes lots of rabbits at sutek's castle too
[06:56:57] <Yuv422> 14
[06:57:02] <Malignant_Manor> Geez.
[06:57:25] <Malignant_Manor> But the other stuff that is supposed to spawn does so right?
[06:59:47] <Malignant_Manor> I seemed to have fixed the crash. Using the inline bool tmpBufTileIsBlack seems to work.
[07:01:26] <Yuv422> I don't know if that is actually fixing the underlying issue though
[07:01:43] <Malignant_Manor> Neither do I but it is better than the checks.
[07:02:04] <Yuv422> also the monsters in around suteks castle aren't mutated in nuvie
[07:02:07] <Yuv422> like the original
[07:02:28] <Malignant_Manor> The mutation is minor compared to missing spawns.
[07:02:41] <Yuv422> true
[07:02:54] <Malignant_Manor> I wasn't worried about it this release.
[07:03:07] <Malignant_Manor> It is something know that hasn't been done.
[07:04:09] <Malignant_Manor> I can try testing in MD.
[07:04:17] <Malignant_Manor> The house of the dead area
[07:06:00] --- Dominus1 is now known as Dominus
[07:09:10] <Malignant_Manor> I'm not sure if the original engine solved the issue.
[07:11:23] <Malignant_Manor> There's the area of darkness around on the left and back side of that 0,0,0 location.
[07:12:31] <Malignant_Manor> You can't go there without cheating either.
[07:18:36] <Malignant_Manor> MapWindow::drawActors() if(tmp_map_buf[(actor->y - cur_y + 1) * tmp_map_width + (actor->x - cur_x + 1)] != 0)
[07:19:47] <Malignant_Manor> get_objAtCursor() get_objAtCursor() if(tmp_map_buf[(cursor_y+1) * tmp_map_width + (cursor_x+1)] == 0)
[07:20:07] <Malignant_Manor> Not sure if those will have any issues
[07:31:07] <Yuv422> it looks like berserk gargs get double stats
[07:31:42] <Malignant_Manor> That's missing from Nuvie right?
[07:31:48] <Malignant_Manor> I'm going to commit the change for tile_is_black. After a bit of testing it still seems okay.
[07:34:46] <CIA-75> nuvie: 03malignantmanor * r1522 10/nuvie/trunk/ (ChangeLog Event.cpp MapWindow.cpp):
[07:34:46] <CIA-75> nuvie: *Don't accept drops onto map or 414 teleport when in a vehicle
[07:34:46] <CIA-75> nuvie: *Call tmpBufTileIsBlack from tile_is_black to prevent a crash
[07:39:30] <Malignant_Manor> The mint worker Terri (30) doesn't seem to ever show up if you keep passing and walking around in the mint.
[07:40:35] <Malignant_Manor> Her house is a decent distance away. I wonder if it is off the map buffer.
[07:41:36] <Yuv422> maybe use an alt cheat to see where she's at
[07:45:36] <Malignant_Manor> I've done that before and she was at home.
[07:47:26] <Malignant_Manor> I added it to the npc schedules don't update properly -teleport (orb of moons) tracker.
[07:49:00] <Malignant_Manor> Nuvie needs to have actors who aren't on the map go to their scheduled location on the hour change. If they are on the map buffer where the player could see, they should do like Exult and appear from just offscreen.
[07:49:27] <Yuv422> yeah I thought it was doing that
[07:57:30] <Malignant_Manor> I tested it again and she was in bed. When you teleport there, she gets up and starts walking to work.
[07:58:55] <Malignant_Manor> I've had two crashes when using the peer cheat but I wasn't running the GDB either time.
[08:12:46] <Malignant_Manor> There's a segmentation fault when closing and you have none as the lighting style.
[08:13:11] <Malignant_Manor> if(shading_globe[i])
[08:13:13] <Malignant_Manor> free(shading_globe[i]);
[08:13:27] <Malignant_Manor> Is where it happens.
[08:13:48] <Malignant_Manor> line 90
[08:21:55] <wjp> looks like the Screen constructor is zeroing the wrong number of values
[08:25:24] <Malignant_Manor> memset( shading_globe, 0, sizeof(uint8*) * 3 );
[08:25:33] <Malignant_Manor> changed to memset( shading_globe, 0, sizeof(uint8*) * 6 );
[08:25:43] <Malignant_Manor> and no more seg fault
[08:26:16] <Malignant_Manor> memset( shading_globe, 0, sizeof(uint8*) * 5 ); does have a segmentation fault so 6 seems to be the correct number.
[08:29:38] <wjp> that's the size of the array, yes
[08:30:47] <wjp> could also just use sizeof(shading_globe), by the way
[08:30:49] <Malignant_Manor> yes, if it starts at 1. globeradius[] = { 76, 116, 180, 240, 270, 330 };
[08:31:07] <wjp> hm? I just mean the header says uint8 *shading_globe[6];
[08:31:54] <Malignant_Manor> static const sint32 in cpp file
[08:33:06] <Malignant_Manor> for the value it takes.
[08:33:28] <Malignant_Manor> sqr(globeradius[i]
[08:33:30] <-- SugarCube has left IRC (Ping timeout: 252 seconds)
[08:36:38] <CIA-75> nuvie: 03malignantmanor * r1523 10/nuvie/trunk/ (ChangeLog screen/Screen.cpp): *Fix segmentation fault from partially initialized value (thanks WJP)
[08:37:58] <-- Malignant_Manor has left IRC (Quit: ChatZilla [Firefox 14.0.1/20120713134347])
[08:47:11] --> SugarCube has joined #nuvie
[08:49:31] <-- Yuv422 has left IRC (Quit: Yuv422)
[13:24:08] <-- Kirben has left IRC ()
[21:22:31] --> Yuv422 has joined #nuvie
[21:33:04] --> Malignant_Manor has joined #nuvie
[21:36:32] <Yuv422> I'm going to add the amulet of submission logic
[21:37:18] <Malignant_Manor> I didn't even know it wasn't working.
[21:37:30] <Yuv422> gargs should be spawned with the worktype 0xc when the amulet is on the avatar
[21:38:04] <Yuv422> I also need the shamino says... logic
[21:38:36] <Yuv422> I might also tighten up the egg spawn / clean range
[21:38:46] <Yuv422> so it is more inline with the original
[21:39:01] <Yuv422> as we only have 31 temp actors slots
[21:39:28] <Malignant_Manor> Yeah, too bad the engine wasn't created to handle more npcs at the beginning.
[21:39:38] <Yuv422> yeah
[21:40:05] <Yuv422> the strange part is there is a gap between the end of the npcs and the temp actors
[21:44:53] <Malignant_Manor> Can Nuvie use those?
[21:49:03] --> Malignant_Manor_ has joined #nuvie
[21:49:42] <Yuv422> I was wondering about that
[21:50:32] <Yuv422> maybe we can investigate that after we get the original game logic finished off
[21:50:36] <-- Malignant_Manor has left IRC (Ping timeout: 256 seconds)
[21:50:42] --- Malignant_Manor_ is now known as Malignant_Manor
[22:15:56] <Dominus> Yuv422: I'm having this patch for the autotools setup for having codesigning a configure option on OS X
[22:16:26] <Dominus> I'm unsure whether it is too hackish and hoping wjp will take a look as I want to do this for the other projects I'm on as well
[22:16:50] <Dominus> what do you think if you have an opinion? :)
[22:17:11] <Yuv422> is the patch on sf.net?
[22:17:30] <Dominus> sorry I meant to paste it http://pastebin.com/vwen0iCU
[22:18:38] <wjp> is the codesign binary always present on OS X?
[22:19:08] <Dominus> yes, at least if you have developer tools installed which you need to do any compiling :)
[22:19:22] <Yuv422> I'd imagine it is only there on os x 10.8+
[22:19:37] <Dominus> nope it's there since forever (I think :)
[22:20:09] <wjp> what does it have as input/output?
[22:20:14] <wjp> does it need setting up? (key-wise)
[22:21:26] <Dominus> https://developer.apple.com/library/mac/#documentation/Darwin/Reference/ManPages/man1/codesign.1.html
[22:21:38] <Dominus> seems it is available since 10.5
[22:21:55] <Dominus> and needs no setting up
[22:22:31] <Dominus> but will fail if the "Developer ID Application" certificate is not there
[22:23:21] <Dominus> the problem with my way is that it will use any certificate that begins with "Developer ID Application"
[22:23:34] <Dominus> in my case it is "Developer ID Application: Dominik Reichardt"
[22:25:21] <Dominus> but if one really wants to codesign, one should have the certificate
[22:26:24] <wjp> hm
[22:26:29] <Dominus> so input is the certificate and whatever you want to sign (could be a several things, binaries, app bundles, disk images)
[22:27:57] <wjp> would it be nice to have configure take the identity as an argument?
[22:29:03] <Dominus> hmm, maybe yes, since in theory one could also use this to build an iOS app and there it is "Developer ID iOS" or so
[22:29:46] <wjp> and I think it should maybe also be the last step in "bundle" rather than the first step in "osxdmg"
[22:31:33] <Dominus> problem is, as we discussed this in #pentagram, that then when osxdmg is being run the bundle gets made again
[22:31:44] <Dominus> oh, I see
[22:32:01] <Dominus> doh
[22:32:08] <Dominus> you are right
[22:32:33] <Dominus> too many trees that blocked my view of the forest
[22:38:51] * Dominus is looking into how to verify a signature via codesign or any other way to maybe test in configure
[22:42:28] <Yuv422> could you call the switch --enable-osx-codesign
[22:43:09] <Dominus> yes, but it is only showing if the host is actually OS X
[22:44:50] <wjp> no, it'll always show up
[22:45:17] <Dominus> oh, I though if test x$HOST_SYSTEM == xMACOSX; then would take care of this
[22:45:34] <wjp> nah... one of the more annoying pitfalls of autoconf
[22:45:43] <Dominus> :(
[22:45:51] <Dominus> thanks for pointing that out
[22:45:53] <wjp> shell ifs and macro expansion often interact counter-intuitively :-(
[22:47:43] <Dominus> ok, so it is going to be --enable-osx-codesign and --with-osx-signature="name of signature"
[22:48:00] <Dominus> (with default being "Developer ID Application" when --enable-osx-codesign) <- does that sound acceptable?
[22:49:29] <wjp> you could also just have the --with-osx-signature option (and have that automatically enable signing if it is present)
[22:49:58] <Dominus> true, makes more sense
[22:50:39] <wjp> (and it can still have a default if you just pass --with-osx-signature without an argument)
[22:51:20] <Dominus> yes, I think I can manage that
[22:52:03] <Dominus> I'll see if I can cook it up and make a new patch tomorrow evening
[22:52:45] <Dominus> if I start this now I'm going to regret it tomorrow morning when I have the first shift, getting up with the baby... :)
[22:53:24] * Dominus is still annoyed that this will show for every system
[22:55:24] <Malignant_Manor> I hate Nuvie's obj ownership system.
[22:55:50] <Yuv422> Malignant_Manor: it's better than it was
[22:56:13] <Malignant_Manor> I don't think objects moved to containers are assigned to the parent actor.
[22:56:45] <Yuv422> they are assigned to their container
[22:56:59] <Yuv422> there is a call that can get you the actor holding the container
[22:57:46] <wjp> Dominus: yeah... but I don't think there's a way to make the help string conditional (since it can't run any tests before it is generated, as far as I can tell)
[22:57:47] <Dominus> ok, going to bed, thanks Yuv422 and wjp for getting me on the right track
[22:57:50] <Yuv422> Obj::get_actor_holding_obj()
[22:57:55] <wjp> Dominus: good night
[22:58:06] <Yuv422> cya
[22:58:16] <Yuv422> I'm off now too
[22:58:28] <Yuv422> cya
[22:58:30] <-- Yuv422 has left IRC (Quit: Yuv422)
[22:58:45] <Dominus> good night