#exult@irc.freenode.net logs for 13 Oct 2010 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage

[00:00:39] <Dominus> I traced back the combat issue to 18th of August, my snapshot from the 17th still has the harpies killing me quickly, the 18th has them slowed down
[00:01:48] <Dominus> right around the time when Colourless took on the fullscreen issue
[00:02:52] <Dominus> I'm netting on Revision #6329 breaking it...
[00:02:56] <Dominus> betting
[00:06:31] <Dominus> exult/trunk/schedule.cc has a combat related change
[00:06:44] <Dominus> - seek_combat = (bool)(qual&32);
[00:06:44] <Dominus> + seek_combat = (qual&32)!=0;
[00:18:41] <Dominus> at least this change alone is not the culprit
[00:30:03] <Dominus> the invisibility issue is a no issue, same thing happens in the original
[00:31:22] <Dominus> and Rev#6329 definitely broke combat, Rev#6328 is behaving normally
[00:51:17] <Dominus> funny, when an NPC starves he doesn'T get health penalty but instead gains hit points :)
[00:53:01] <Dominus> might be a really old bug...
[01:21:53] <Dominus> interesting savegame in https://sourceforge.net/tracker/?func=detail&aid=3040803&group_id=2335&atid=102335
[01:22:31] <Dominus> the savegame is almost 3 MB big with the file u7ireg over 1.4 GB big
[01:22:53] <Dominus> the file u7ireg33, I mean
[01:23:40] <Dominus> how can this get so big?
[01:27:08] <Dominus> Again this is on 64bit Windows...
[02:16:10] <Dominus> so, had my way with the bug tracker. The only show stopping bugs seem to me the combat bug and the starving-is.healthy bug
[02:24:30] <-- Dominus has left IRC (Quit: Leaving.)
[03:14:58] --> Zxcvb has joined #exult
[03:15:24] <Zxcvb> how well does the dingoo port work?
[03:33:43] <-- Zxcvb has left IRC (Quit: Leaving)
[03:52:09] <-- Marzo has left IRC (Ping timeout: 265 seconds)
[03:59:56] <-- jvlee has left IRC (Ping timeout: 255 seconds)
[05:46:15] <-- Kirben has left IRC ()
[05:52:37] --> Kirben has joined #exult
[05:52:38] --- ChanServ gives channel operator status to Kirben
[06:26:58] <-- Morde has left IRC (Read error: Connection reset by peer)
[07:32:48] <wjp> fixed combat
[07:36:41] <wjp> Dominus: thanks for tracking it down to a specific commit :-)
[07:41:30] <Colourless> so you undid my fix and like me have no idea why the code is the way it was
[07:42:09] <Colourless> i changed it to what i think the code was trying to do
[07:42:56] <Colourless> but as the person who first implemented that code 'thought' it was working, the fixed code doesn't do what was intended at all
[07:43:38] <Colourless> sue is a no issue, same thing happens in the original
[07:43:39] <Colourless> [11:01] <Dominus> and Rev#6329 definitely broke combat, Rev#63
[07:43:45] <Colourless> uh..
[07:44:02] <Colourless> ignore last two lines
[07:44:49] <Colourless> !rand()%3 and !rand()%4 are logically equivilant to rand()!=0
[08:10:13] <wjp> yes
[08:10:27] <wjp> which technically matches the comment, but, well... :-)
[08:14:53] <wjp> but you're right, I have no idea what it's doing
[08:15:02] <wjp> or trying to do, rather
[08:27:40] <wjp> it was added in r5842 back in 2007
[08:41:03] <-- Rottingbeef has left IRC ()
[08:44:20] <Colourless> re windows x64 issues. i don't believe they exist. The 32bit environment in x64 is virtually identical to the 32 bit environment in 32bit windows.
[08:45:42] <Colourless> if an issue is showing up in 64bit windows, there is an issue in exult that needs fixing
[08:46:44] <wjp> sure, but being able to reproduce it would help :-)
[08:55:33] <Colourless> such is the life of a coder, trying to figure how to reproduce bugs... that only a single user has reported
[08:58:18] --> Dominus has joined #exult
[08:58:18] --- ChanServ gives channel operator status to Dominus
[08:58:28] <wjp> hi
[09:09:08] <Dominus> hi
[09:09:32] <Dominus> thanks for finding the culprit in the long commit
[09:10:06] <wjp> it helped it was near the top of the diff :-)
[09:10:28] <Dominus> since I'm keeping the OS X snapshots it's easy to pinpoint the problem
[09:10:34] <wjp> you already did the hard work fortunately
[09:10:39] <Dominus> if it happened between april and now :)
[09:12:49] <Dominus> any idea on the starving gives hit points bug? if that one gets resolved, I *think* we are ready (and on the other hand if it doesn't get resolved we might get lucky again and no one notices it) :)
[09:13:10] <wjp> haven't looked at that yet
[09:13:15] <wjp> can't be too hard I guess
[09:13:44] <Dominus> at least in the changelog I haven't found anything that stands out on first glance
[09:17:48] <wjp> hm, code looks ok at first glance
[09:19:00] <wjp> I can take a closer look tonight
[09:19:19] <Dominus> thanks
[09:24:43] <Dominus> hmm, does anyone have a collection of SI savegames? I'd like to verify how the original handled list field training when a party memeber is invisible but that means I need to play quite a bit :(
[09:40:47] <wjp> I should, but can only check tonight
[09:42:22] <Dominus> thanks again, give me a notice if you have one
[09:44:20] <wjp> that hunger thing isn't something like regen rings regenerating faster than the hunger?
[09:44:56] <Dominus> no, no rings on the hands
[09:46:30] <wjp> if you want to debug it a bit, an interesting thing to do would be adding a printf("reduce_health: delta %d, oldhp %d, val %d\n", delta, oldhp, val); on actors.cc line 2769 (below 'int val = oldhp - delta;')
[09:47:03] <Dominus> ok, will do later today but before tonight :)
[09:47:11] <wjp> and a printf("reduce_health: new value %d\n", val); below 'properties[static_cast<int>(health)] = val;' on line 2845
[09:54:47] --> Rottingbeef has joined #exult
[10:35:05] --> Morde has joined #exult
[11:03:52] <wjp> Dominus: how often did Iolo gain hitpoints? I don't see anything other than the regular very slow gradual healing
[11:04:33] <wjp> and I do lose hitpoints when starving, but only very occasionally
[11:04:42] <Dominus> hmm, strange
[11:05:18] <Dominus> for me it was very obvious that right on the point when he complained about starving (which took some time) he gained about 3 hit points
[11:11:29] <wjp> I sped up the effect a bit to get a complaint every few seconds, but it seems to be working as intended here
[11:11:44] <wjp> (i.e, reduction in hp of either 0, 1 or 2 points)
[11:12:37] <Dominus> I'll add the logging now and will try to see if I can reproduce it somehow...
[11:22:33] <wjp> the only thing I can think of at the moment would be your rand() returning negative values
[11:25:46] <wjp> if you want to speed things up, in npctime.cc, around line 350, change 'if (minutes >= last_time + 60)' to 'if (minutes >= last_time + 1)' and 'add(curtime + 30000, ...)' to 'add(curtime + 1000, ...)'
[11:25:46] <Dominus> hmm, there it did it again
[11:25:52] <wjp> ah :-)
[11:26:16] <Dominus> and it didn't print me log, maybe I need to compile with debug enabled?
[11:26:42] <Dominus> iolo complained and right then hit points jumped up three points
[11:26:45] <wjp> huh
[11:28:15] <Dominus> maybe because it didn't reduce health? :)
[11:31:21] <wjp> can you try with the speed-up?
[11:31:32] <wjp> does it ever print the log?
[11:34:26] <wjp> it doesn't make sense
[11:34:34] <Dominus> i thought I'd see it n the console along with the other output
[11:34:40] <wjp> yes, should be there
[11:34:51] <Dominus> speed up doesn't seem to work...something is wrong, let's see what I'm doing here...
[11:34:59] <wjp> are you actually rebuilding? :-)
[11:35:21] <wjp> (note that the first instance of the hunger will still take some time; you can use Ctrl-T to advance time a bit)
[11:37:14] <Dominus> he he, did that twice and iolo is up to full helath after being down to 8 hit points
[11:37:30] <Dominus> reduce_health: delta 2, oldhp 14, val 12
[11:37:31] <Dominus> reduce_health: new value 12
[11:37:41] <wjp> that sounds like spark
[11:37:55] <wjp> and is what should be happening
[11:38:07] <Dominus> hmm, let me try again and get rid of spark
[11:38:17] <wjp> any negative deltas? new values higher than oldhp?
[11:39:24] <Dominus> reduce_health: delta 0, oldhp 18, val 18
[11:39:24] <Dominus> reduce_health: new value 18
[11:39:29] <Dominus> and the same again
[11:39:38] <Dominus> as if iolo doesn'T count
[11:39:49] <Dominus> but his health increased
[11:41:12] <wjp> hm, the hourly healing can be rather big, by the way
[11:41:25] <wjp> up to one third of maxhp
[11:41:57] <Dominus> it's really strange, I see the avatars health decreasing, log does fit him, but nothing happening to iolo
[11:42:08] <wjp> so maybe Iolo is just not starving yet?
[11:43:00] <wjp> after two ctrl-T they are healed pretty much up to full health, and after that it starts decreasing
[11:43:10] <Dominus> his food level is 1
[11:43:13] <wjp> for all three
[11:43:36] <wjp> starving is <= 0
[11:45:37] <Dominus> ok, after setting iolo to 0 he did receive penalty
[11:45:54] <Dominus> but before he just didn't fall below 1
[11:46:28] <Dominus> the avatar with a food level of 2 was already starving to death before Iolo starved at all
[11:46:44] <Dominus> maybe that's what the reporter observed
[11:49:03] <wjp> food level drops by a random amount every hour too
[11:49:20] <wjp> so it's not impossible that the avatar hits 0 before Iolo
[11:49:32] <wjp> (randomly 0, 1, 2 or 3)
[11:50:37] <wjp> that coupled with the hourly healing probably explains everything, then
[11:51:19] <Dominus> yup, setting avatar food level very high, ctrl+t three times and iolo finally starved
[11:51:26] <wjp> and the bug is probably that the hourly healing will typically be larger than any hunger damage...
[11:51:43] <wjp> (which is also hourly)
[11:52:14] <Dominus> yes, that explains it very well
[11:53:59] <wjp> pesky randomness :-)
[11:54:05] <Dominus> getting rid of the bug report then
[11:54:16] <wjp> best leave it, I think
[11:54:22] <wjp> but lower priority
[11:54:38] <wjp> we should still re-evaluate the hunger damage and healing rates it seems
[11:54:55] <Dominus> ok, will also change the subject a bit and write our findings down
[11:54:56] <wjp> or maybe not heal when starving or something
[11:55:00] <wjp> thanks
[12:01:34] <Dominus> ok, then I'll have to take a look at FAQ and docs
[12:03:05] <Dominus> Kirben: if you are reading this, can you stop doing a snapshot after the next one (wjp's combat fix)? I'll update all the necessary version files today or tomorrow and don't want to confuse people :)
[12:03:57] <Dominus> and would you be willing to make a release build when all is ready? Without Studio support and without OpenGL?
[12:08:13] <Dominus> and sometimes I'd really like to know more about your snapshot build system :) how much is automated, what you do manually etc... also how much bandwidth you spent in the last decade (and more) on all of the project snapshots you are providing...
[12:28:50] --> jvlee has joined #exult
[13:01:26] <Kirben> ok, you want Windows snapshots on hold to test a certain fix? or rather an RC build for next release?
[13:02:02] <Kirben> I can handle a release build if required, with specifc features enabled/disabled.
[13:02:37] <Kirben> The Windows snapshots are still all handled manually.
[13:02:58] <Dominus> actually I thought you might do a "last" snapshot with wjp's changes and after that hold off doing snpashots until we have a rc1 out
[13:04:03] <Dominus> and kirben, wow... still doing all those snapshots manually... that is quite a sum of stuff you've done over the years
[13:06:08] <Kirben> ok, I will do.
[13:07:37] <Dominus> I hope to be done with all the versions stuff and last docs polish today/tonight, then we will have to see how to actaully release these days on Sourceforge. I'll also look into updating the whatsnew.txt with all the changes we've done
[13:07:40] <Dominus> Thanks kirben
[13:08:19] <Dominus> so, Windows version will be Kirben, Mac- me, other *nix stuff... maybe wjp and/or marzo, I hope :)
[13:08:41] <Dominus> I'll report on the ML how far I'm done with stuff
[13:11:11] <-- Kirben has left IRC ()
[13:31:53] --> Marzo has joined #exult
[14:36:36] <Dominus> hmm, about changing version number, I'm not sure whether tagging ES the same (/win32/exultstudioico.rc) but since we keep them in sync, I guess I should, even though we will not release ES...
[14:37:00] <wjp> yeah, might as well
[14:37:11] <wjp> some people might build from source and enable ES
[14:37:22] <Dominus> k
[14:43:52] <Dominus> sometimes we have "Copyright 1998-2010" (keyactions.cc) and sometimes "Copyright 2000-2010" (infoplist.in), I guess this should be always 1998-2010, or?
[14:44:54] <wjp> each source file has its own date range
[14:45:09] <wjp> but the ranges for the whole of Exult should probably be 1998-2010
[14:46:09] <wjp> (so that's anything in resource files, about dialogs and the like)
[14:46:14] <Dominus> yeah, I'm talking about the displayed one (right click on exult.app shows the range specified in the infoplist.in)
[14:46:24] <wjp> right
[14:50:28] <Dominus> our readme.txt also mentions msvcstuff/exconfig/exconfig.rc - does this still need changes? I think it's for the old installer (installshield) and not the Inno Setup one. Maybe Colourless can answer this :)
[14:53:56] <Colourless> exconfig stuff is still used and required by the installer. it allows the installer to configure exults paths
[14:54:23] <Colourless> its in the win32 dir now
[14:54:53] <Dominus> that one I'm changing, just wondering about the ones in msvc folders
[14:54:56] <Colourless> it needs updating
[14:55:07] * Colourless shrugs
[14:55:12] <Dominus> :)
[14:55:21] <Colourless> the msvc one probably shouldn't exist
[14:55:49] <Dominus> k
[14:57:17] * Colourless looks at the time
[14:57:28] <Colourless> i really should be in bed. l8tr
[14:57:32] <Dominus> :)
[14:57:35] <Dominus> niht
[14:59:20] <wjp> good night
[15:11:44] <Dominus> hmm, where to put Notebook, Autonotes and Screenshots in the docs?
[15:13:57] <Dominus> in the section "Stuff that doesn't fit anywhere else" :)
[15:30:06] <Dominus> For the autonotes feature of exult I'm "hard" linking the SVN files in the documentation (e.g. http://exult.svn.sourceforge.net/viewvc/exult/exult/trunk/data/bg/autonotes.txt), mentioning that these are mostly giving the barebones of information, especially the SI file
[15:38:52] <Dominus> he he he, pcx is really a very old format... OS X does not come with a viewer for that format :)
[15:39:15] <Dominus> maybe we should switch to PNG at some point :)
[16:22:15] --> SiENcE has joined #exult
[16:51:26] <-- SiENcE has left IRC (Quit: @all: cya)
[18:41:11] --> SiENcE has joined #exult
[19:40:39] <Dominus> I *think* I got the versions done, now looking at the news.txt file :)
[20:07:55] <Dominus> bah... brain farts on writing news.txt...
[21:08:23] <wjp> do let me/us know if we can do anything
[21:10:48] <Dominus> will do, I've written some stuff, will commit now and then post to the ML
[21:15:11] --> Cahaan has joined #exult
[21:40:45] <-- Morde has left IRC (Read error: Connection reset by peer)
[21:44:39] <-- SiENcE has left IRC (Quit: cya @all)
[21:44:53] --> Fingolfin has joined #exult
[21:44:53] --- ChanServ gives channel operator status to Fingolfin
[22:05:04] <-- Rottingbeef has left IRC ()
[22:16:33] --> Kirben has joined #exult
[22:16:33] --- ChanServ gives channel operator status to Kirben
[22:48:57] <-- Fingolfin has left IRC (Quit: Fingolfin)
[22:51:54] --> Rottingbeef has joined #exult
[23:57:58] <-- Cahaan has left IRC ()