#pentagram@irc.freenode.net logs for 26 Feb 2009 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage


[00:36:35] --> Fingolfin has joined #pentagram
[00:36:35] --- ChanServ gives channel operator status to Fingolfin
[00:56:19] <-- watt has left IRC ()
[01:26:47] --> watt has joined #pentagram
[01:26:47] --- ChanServ gives channel operator status to watt
[01:29:13] <-- watt has left IRC (Client Quit)
[01:46:07] --> watt has joined #pentagram
[01:46:07] --- ChanServ gives channel operator status to watt
[02:11:00] --> ShadwChsr has joined #pentagram
[02:11:08] <ShadwChsr> 'Evenin
[02:11:10] <ShadwChsr> :)
[02:11:22] <watt> Howdy.
[02:22:05] <-- Fingolfin has left IRC ()
[02:38:33] <watt> ?logs
[02:38:33] <exultbot> Logs are available at http://log.usecode.org/pentagramlog.php
[03:33:40] <ShadwChsr> I'm seeing some rendering flaws at the shrine of the ancients
[03:34:03] <ShadwChsr> not just the missing shape - but floor tiles are positioned wrong and the entire zelan sequence shows only the roof
[04:43:52] <watt> Neat... just ran into a case of version differences with ShadwChsr. I appear to have that 2.13 version mentioned way back when - http://log.usecode.org/pentagramlog.php?log=30Jul2005
[04:56:35] <watt> Good think I didn't disable debug in my builds :-)
[04:56:36] <watt> ignore_savegame_mismatch=yes
[05:23:13] <ShadwChsr> Anyone alive here?
[05:23:18] <ShadwChsr> I need a save for mythran's book :(
[06:26:41] <-- ShadwChsr has left IRC ()
[07:52:41] --> pupnik has joined #pentagram
[08:10:12] <-- pupnik_ has left IRC (Read error: 101 (Network is unreachable))
[13:09:21] <-- Kirben has left IRC (Read error: 60 (Operation timed out))
[14:47:41] <-- Mrrrrz0r has left IRC (Read error: 110 (Connection timed out))
[14:49:45] <-- watt has left IRC ()
[15:13:57] --> watt has joined #pentagram
[15:13:57] --- ChanServ gives channel operator status to watt
[15:18:11] <watt> BTW - EA owned u8 download site for the patches to the game. http://www.uo.com/archive/ultima8/download.html
[16:07:25] --> Mrrrrz0r has joined #pentagram
[17:03:19] --> peewee_RotA has joined #pentagram
[17:03:25] <peewee_RotA> hello
[17:04:00] <peewee_RotA> Just out of curiousity, does anyone know if the concept of GUMPs is copyrighted?
[17:07:00] <wjp> a gump is just a window really
[17:15:36] <peewee_RotA> Yeah, that does make sense. But the major difference in the technology is that in normal inventory, you have an item and a quantity.
[17:15:48] <peewee_RotA> A gump is an inventory that also assigns an XY
[17:16:13] <peewee_RotA> position
[17:17:08] <peewee_RotA> It just seems like the kind of thing that may be patented. Seeing as they invented it and named it, it's not too much of a stretch that they may have patented it.
[17:17:57] <peewee_RotA> I can't think of any non-Ultima titles that takes advantage of the idea. The closest is the slots in diablo.
[17:28:42] <wjp> hm, file management windows use the same concept I think
[17:28:53] <wjp> I wouldn't worry about it
[17:31:17] <peewee_RotA> I do doubt that EA would activly persue anyone if this was a problem.
[17:31:19] --> ShadwChsr has joined #pentagram
[17:31:41] <wjp> hi ShadwChsr
[17:31:59] <ShadwChsr> Hey :)
[17:32:14] <ShadwChsr> Watt and I were investigating the sorter a bit last night
[17:32:27] <wjp> yes, very good to hear :-)
[17:32:32] <wjp> did you see my mail?
[17:32:38] <ShadwChsr> not yet, let me check
[17:32:54] <wjp> I sent it to your SF address
[17:33:34] <ShadwChsr> Ahhh, great
[17:33:38] <ShadwChsr> I'll investigate tonight
[17:33:52] <ShadwChsr> It looks the fix for Mythrans book caused a few of the other render issues you're tracking
[17:34:36] <wjp> interesting
[17:34:37] <ShadwChsr> I commented out the fix (ztops check) and also changed the issue I found - it seemed to solve the roof issue at the Zealan Shrine, The Zealan Shrine "glowing" sprite bug, as well as the Ogre/Avatar sort issue
[17:34:55] <ShadwChsr> Of course, that probably leaves Mythrans book broken again, but there might be a better fix.
[17:35:37] <ShadwChsr> The problem is it's checking ztops before the more obvious overlapping cases - any objects of varying height will override cases where objects are 'barely' overlapping on x or y
[17:35:51] <ShadwChsr> At very least it probably should be changed to zbottoms, but I'm skeptical if it's needed at all.
[17:39:53] <peewee_RotA> I've noticed a couple bugs in the past couple weeks. Should I post them in SF or would you rather hear them on here?
[17:40:32] <wjp> peewee_RotA: please post them to our tracker
[17:41:07] <ShadwChsr> I looked through the #Pentagram logs about the sorting bug I posted - I think there was some discussion about whether it was a real bug.
[17:43:09] <ShadwChsr> ItemSorter.cpp(387)
[17:43:10] <ShadwChsr> if (si1.x + si1.y != si1.x + si2.y) return (si1.x + si1.y < si2.x + si2.y);
[17:43:45] <ShadwChsr> In a way it's a trinary state - equivalent, less than, greater than.
[17:44:05] <ShadwChsr> The "if" statement handles the equivalence, while the "return' statement handles the < or >
[17:44:38] <ShadwChsr> it's kind of like:
[17:45:00] <ShadwChsr> if (si1.x + si1.y == si2.x + si2.y) return 0;
[17:45:16] <ShadwChsr> else if (si1.x + si1.y < si2.x + si2.y) return -1;
[17:45:18] <ShadwChsr> else return 1;
[17:46:51] <wjp> no, it doesn't return if ==
[17:47:39] <wjp> it's just like with string compares: if the objects differ in this property, we can say which is in front, otherwise we have to look further
[17:49:16] * ShadwChsr nods
[17:49:37] <ShadwChsr> My if example wasn't great, was meant show the trinary state a bit better.
[17:50:13] <ShadwChsr> Either way, shouldn't the statements match other than the operator?
[17:50:13] <wjp> I'm sure we all understand the concept, but why did you bring it up?
[17:50:55] <wjp> because you completely lost me...
[17:51:02] <ShadwChsr> In the logs there was some discussion about whether it was a real bug or not
[17:51:09] <wjp> what is "it" ?
[17:51:35] <ShadwChsr> (si1.x + si1.y != ---> si1.x <--- + si2.y)
[17:51:47] <wjp> that's clearly a bug
[17:51:47] <ShadwChsr> Should be si2.x
[17:52:17] <ShadwChsr> Sorry, I should have linked you to the tracker first. :)
[17:53:41] <wjp> referring to "the sorting bug I posted" after mentioning 4 of them earlier was a bit confusing :-)
[17:53:51] <ShadwChsr> Sorry :)
[17:54:17] <ShadwChsr> There's another strange issue at the Zealan shrine - the carpets are misplaced
[17:54:28] <wjp> np :-)
[17:54:47] <ShadwChsr> It's not a sorter issue - the actual coordinates are wrong. I'm starting to suspect they're wrong in the actual map itself, but there's a difference in rounding or position calculation between Pentagram and U8
[17:55:09] <ShadwChsr> I think Pentagram is technically correct but U8's behavior hid the bug.
[17:56:11] <wjp> I think I reached a similar conclusion when I looked at it, but can't remember the specifics
[17:57:02] <ShadwChsr> Is it worth fixing at all? Is there a 'shim' system in place to cleanly 'hack' maps and usecode as they're loaded?
[17:58:44] <wjp> not really... I'm hacking some maps to fix holes in the ground already, though
[17:59:00] <wjp> (see Map::loadFixed)
[18:00:10] <wjp> one thing that is possibly related: when spawning globs, the low bits of the item coordinates are ignored. Maybe a similar effect affects things here?
[18:00:31] <wjp> (see GlobEgg::enterFastArea, coordmask)
[18:05:09] <ShadwChsr> Hrmmmmmmm, could be.
[18:05:31] <ShadwChsr> might also be something like (going by memory)
[18:05:37] <ShadwChsr> Pentagram does:
[18:05:44] <ShadwChsr> (x + y) / 4
[18:06:30] <ShadwChsr> I could see the U8 guys originally designing for a massive map, and in that case maybe they wanted to reduce the potential of an overflow, or maybe they were using 16 bit values in places
[18:06:33] <ShadwChsr> they could have done:
[18:06:37] <ShadwChsr> x/4 + y/4
[18:06:49] <wjp> could be
[18:06:54] <ShadwChsr> The result would be different between the two due to integer division
[18:07:05] <ShadwChsr> maybe just enough to throw some tiles off *shrug*
[18:08:29] <wjp> the effect would be quite subtle since items are typically aligned to a larger than 4x4 grid
[18:08:50] <wjp> so it does sound like a potential explanation
[18:30:31] <ShadwChsr> I'll play with it a bit tonight :)
[18:50:45] <ShadwChsr> I think I know why the Orge sort bug was difficult to find
[18:51:05] <ShadwChsr> There's another bug in the ">>" operator overload (strangely) used to show a string output of the comparison process in the console
[18:51:09] <ShadwChsr> COMPARISON_RETURN_TF(si1.ztop,<,si2.z,true,"");
[18:51:23] <ShadwChsr> That's the line that sorted the ogre wrong - should be si2.ztop
[18:51:27] <ShadwChsr> on the right-hand side
[20:01:18] <ShadwChsr> Would you be open to giving me SVN access to fix some of these things?
[20:01:35] <ShadwChsr> I could always work in a branch for review - probably easier than a bunch of out of date patches
[20:07:56] <ShadwChsr> I don't mind being on a short leash ;)
[20:11:41] <wjp> in time, yes, but let's start with a number of patches first
[20:16:54] <peewee_RotA> Aparently you must learn confusion blast before you can learn branch revision.
[20:18:23] <peewee_RotA> (the first reagent for the "revision branch" spell is a tortoise LOL)
[20:20:11] <watt> Kal Vas Flam Corp Xen
[20:26:11] <ShadwChsr> https://sourceforge.net/tracker2/?func=detail&aid=2532651&group_id=53819&atid=471706
[20:26:49] <ShadwChsr> Posted a patch for the obvious ItemSorter issues - left the ZTops for now until I can debug the book too
[20:27:51] <wjp> thanks!
[20:28:00] <ShadwChsr> n/p :)
[20:56:59] <-- ShadwChsr has left IRC ()
[21:41:32] <peewee_RotA> I have to note that the "Chests hanging against wall" but sounds much more graphic than what the bug really is. Can we just add bloody torso's to the wall anyway?
[21:41:37] <peewee_RotA> *bug
[21:44:20] <wjp> heh
[22:03:39] <peewee_RotA> well nice chatting. Have a good evening
[22:03:43] <-- peewee_RotA has left IRC ("Leaving")
[22:25:50] --> Kirben has joined #pentagram
[22:25:50] --- ChanServ gives channel operator status to Kirben
[22:30:35] <-- Kirben has left IRC (Client Quit)
[22:37:21] --> Kirben has joined #pentagram
[22:37:21] --- ChanServ gives channel operator status to Kirben
[22:49:33] --> ShadwChsr has joined #pentagram
[23:51:59] <-- watt has left IRC ()