#nuvie@irc.freenode.net logs for 20 Sep 2013 (GMT)

Archive Today Yesterday Tomorrow
Nuvie homepage

[00:17:46] <Malignant_Manor> No, esc never did close gumps. Anyway, I wanted to have cancelAction() close gumps a long time ago but never implemented it.
[00:18:21] <Malignant_Manor> Next revision will have it.
[01:24:45] <-- Malignant_Manor has left IRC (Quit: ChatZilla [Firefox 23.0.1/20130814063812])
[03:47:26] <-- SugarCube has left IRC (Ping timeout: 246 seconds)
[03:59:16] --> SugarCube has joined #nuvie
[08:29:11] --> Yuv422 has joined #nuvie
[09:25:41] <-- Yuv422 has left IRC (Quit: Yuv422)
[09:47:05] <-- SHODAN has left IRC (Read error: Connection reset by peer)
[09:48:57] --> SHODAN has joined #nuvie
[10:50:44] --> Yuv422 has joined #nuvie
[13:01:25] --> Malignant_Manor has joined #nuvie
[13:01:39] <Yuv422> hey Malignant_Manor
[13:01:45] <Malignant_Manor> hi
[13:01:51] <Malignant_Manor> I see you fixed the bug.
[13:02:05] <Malignant_Manor> Do you have any examples of { being used?
[13:06:03] <Yuv422> lots in the fm-towns scripts
[13:06:55] <Yuv422> I need to revisit the towns speech play back
[13:07:10] <Yuv422> as the original doesn't pause the dialog
[13:07:37] <Yuv422> it prints each paragraph as they audio plays
[13:08:01] <Malignant_Manor> ah so that could be the
[13:08:07] <Malignant_Manor> { uses
[13:08:12] <Yuv422> yeah
[13:09:23] <Malignant_Manor> I know the FM-Towns had a bigger screen for talking so had no idea about the line by line.
[13:16:37] <Yuv422> I'm off to bed now
[13:16:39] <Yuv422> cya
[13:16:42] <Malignant_Manor> bye
[13:16:57] <-- Yuv422 has left IRC (Quit: Yuv422)
[13:37:12] <-- Malignant_Manor has left IRC (Quit: ChatZilla [Firefox 23.0.1/20130814063812])
[16:32:58] --> Malignant_Manor has joined #nuvie
[16:34:42] <Malignant_Manor> wjp: The game hangs when looking at generic SE npcs. I fixed a crash when looking at non-existent portrait files. http://pastebin.com/m6euS14a
[16:35:07] <Malignant_Manor> The line shp->load(&shp_lib, 0); is where the hang occurs.
[17:10:46] <wjp> hm, hang as in an infinite-loop-like thing?
[17:11:01] <Malignant_Manor> I'm not sure.
[17:11:15] <wjp> but it doesn't crash and doesn't get past that line?
[17:11:47] <Malignant_Manor> return NULL after it doesn't stop the hang.
[17:12:01] <Malignant_Manor> shp_buf prints out as nothing
[17:12:16] <wjp> let's see if I have SE installed
[17:12:43] <wjp> annoyingly no
[17:12:53] <wjp> let's fix that first ;-)
[17:15:19] <wjp> lazy question: how do I start SE in nuvie?
[17:15:32] <wjp> and how do I trigger this bug afterwards?
[17:15:47] <Malignant_Manor> loadgame setting put se in the field
[17:16:58] <wjp> right
[17:17:50] <Malignant_Manor> To hang, there should be an npc on the far left after exiting the hut. Look at that npc.
[17:18:17] <wjp> how far is far?
[17:18:56] <wjp> oh look, a segfault
[17:19:09] * wjp applies patch
[17:19:36] <Malignant_Manor> alt 404 to npc # 107 will cause the hang too.
[17:20:11] <wjp> alt 404?
[17:20:47] <Malignant_Manor> Yeah, hold alt and press 4 0 4
[17:20:53] <wjp> hm, top-left of the fire is someone with a white portrait which triggers a pause of a second
[17:21:11] <wjp> nothing happens on alt-404
[17:21:56] <Malignant_Manor> oops 414
[17:22:01] <Malignant_Manor> My patch fixes directly opening the portrait with the alt-300 cheat
[17:22:17] <wjp> ah, there we go
[17:22:33] <Malignant_Manor> One of many generic npcs.
[17:23:40] <wjp> it'll snap out of it after a minute or so
[17:23:55] <Malignant_Manor> U6Lib_n::get_item probably returns a blank buff from malloc initializations.
[17:23:59] <wjp> it thinks there are rather a lot of entries in that data file somehow
[17:28:46] <wjp> (or maybe a little more than a minute)
[17:29:39] <wjp> only to end up with garbage
[17:30:01] <wjp> so we mis-parse this data file. Bah
[17:36:42] <wjp> one possibility is that the shp_lib.open below your patch needs a 2 instead of a 4 for SE
[17:37:22] <wjp> or maybe not
[17:37:54] <Malignant_Manor> The shp_buf.open call is where the issue happens.
[17:40:58] <Malignant_Manor> The faces.lzc index shouldn't exist for the portrait.
[17:43:13] <Malignant_Manor> I can exclude indexes in Portrait::get_portrait_num but that is a hard coded solution and I would have likely have to do the same to MD.
[17:44:00] <wjp> oh, I think I see the problem
[17:44:20] <wjp> this is a bit annoying
[17:45:35] <wjp> yes...
[17:45:46] <Malignant_Manor> The bad workaround id here. (not sure if all are excluded offhand) http://pastebin.com/4RL9T7ET
[17:46:04] <wjp> nah, this is a bug in the file loader
[17:46:23] <Malignant_Manor> Yeah
[17:46:24] <wjp> it thinks the first item in the file is the file with index 0
[17:46:35] <wjp> and that isn't true for this one
[17:46:52] <wjp> and some offset computations get horribly confused by this
[18:00:41] <Malignant_Manor> Dominus: Is the commit that closes gumps on space and esc working how you think it should?
[18:00:43] <wjp> ok, fixed
[18:02:03] <wjp> will commit in a minute
[18:02:47] <Malignant_Manor> Do the portraits come out correctly?
[18:03:05] <Malignant_Manor> Like blond haired Avatar
[18:03:08] <wjp> if I press ctrl-1 I get a broken portrait
[18:03:16] <wjp> but the first couple of NPCs I run into seem ok
[18:03:53] <Malignant_Manor> Those portraits never were working.
[18:05:13] <Malignant_Manor> SE and MD need a lot of work.
[18:05:40] <wjp> this is faces.lzc you said? (For the ChangeLog entry)
[18:06:26] <Malignant_Manor> Yeah
[18:08:06] <Malignant_Manor> I screwed up an earlier changelog/svn log.
[18:08:31] <Malignant_Manor> http://pastebin.com/fv4XptJb
[18:08:48] <wjp> oh, I just committed
[18:08:54] <Malignant_Manor> oh well
[18:12:22] <Malignant_Manor> I think it still needs my Portraits.ccp fix for the crash.
[18:13:18] <wjp> yes, should be unrelated
[18:14:00] <Malignant_Manor> Did it think there were no skipped indexes?
[18:14:27] <wjp> with the hang, you mean?
[18:14:46] <Malignant_Manor> Yeah
[18:15:13] <wjp> it was assuming that the data of the index 0 entry was right after the offset table
[18:15:25] <wjp> and so that you could compute the length of the offset table from the position of the data of entry 0
[18:15:51] <wjp> but in this file, entry 0 was at the very end, so it though there were a _lot_ of entries in that file
[18:16:06] <wjp> half a million or so, and trying to find all of those took some time
[18:16:45] <wjp> (it uses a rather inefficient algorithm, but that isn't an issue for 100 entries)
[18:17:18] <wjp> it's now a bit smarter :-)
[18:19:33] <Malignant_Manor> MD and SE also have hacks to deal with.
[18:20:59] <Malignant_Manor> I know I have a saved file with look description indexes screwed up.
[18:36:54] <-- Malignant_Manor has left IRC (Quit: ChatZilla [Firefox 23.0.1/20130814063812])
[19:38:48] <Dominus> maligant_manor, yes, it now behaves as I expected. Thanks for implementing
[20:22:12] --> Yuv422 has joined #nuvie
[20:55:01] --> Malignant_Manor has joined #nuvie
[20:58:28] <Malignant_Manor> Yuv422: I finally added the multiuse (doubleclick catch all) function to a key binding setting.
[21:18:06] <-- Malignant_Manor has left IRC (Quit: ChatZilla [Firefox 23.0.1/20130814063812])
[21:48:30] <Yuv422> cool