IRC logs for #opendune on OFTC at 2009-11-19
⏴ go to previous day
00:06:25 <Xaroth> anyways, i hacked the authentication api a bit
00:06:39 <Xaroth> file download urls should work directly
00:07:05 <Xaroth> not sure on other browsing
00:07:28 <Xaroth> but it'll try to log in as anonymous if not logged in, unless at a login page
07:16:09 *** TrueBrain has joined #openDune
10:19:47 <TB2> I have to attend a few presentations .. you got to love laptops and wireless
10:20:06 <TB2> Xaroth: tnx for fixing up Mantis :) Didn't try it yet, but I am sure you did well ;) :p
10:23:14 *** mib_1cf4x5 has joined #openDune
10:23:25 *** mib_1cf4x5 is now known as TB3
10:23:33 <TB3> oops .. wireless switch ;)
10:28:47 <TB3> Pff, you guys are boring :p
10:38:46 <TB3> clearly 2 people got addicted to the game :p
15:20:10 <Xaroth> fed2k is down again? :o
15:38:28 *** Xaroth sets mode: +o TrueBrain
15:49:06 <DorpsGek> SVN: truebrain (r524) [JIT] -Add: mapped another 12 functions (crash-bins provided by glx)
15:50:09 <TrueBrain> planetmaker: I dunno what you did with #15, but something went terrible wrong
15:51:03 <planetmaker> The website complained about not having chosen a report category.
15:51:11 <planetmaker> I didn't notice. crash.bin is lost
15:51:17 <planetmaker> as overwritten by #16
15:51:28 <TrueBrain> euh, no, I meant, in your game something went wrong
15:51:39 <planetmaker> dunno. What's wrong there?
15:51:55 <TrueBrain> Program Termination: jumped to 8F00:8600 <- VERY invalid memory address
15:52:00 <TrueBrain> so .. I am guessing some corruption of some kind
15:52:16 <planetmaker> so you don't fix corruptions? ;-)
15:52:39 * planetmaker tries to think of many way on how to corrupt you :-P
15:53:12 <DorpsGek> SVN: truebrain (r525) [JIT] -Fix (#16): one more problem found by planetmaker
15:54:04 <DorpsGek> SVN: truebrain (r526) -Update (r524, r525, #16): update decompiled code
15:54:25 <planetmaker> ok, so issue 15.5 got lost
15:54:43 <TrueBrain> not lost; it is impossible to fix :p
15:54:52 <TrueBrain> or what you mean with .5? :P
15:55:35 <planetmaker> I had a crash more. Probably it didn't get submitted then at all... due to what I just said before, the missing first selection button
15:55:50 <planetmaker> which I missed to choose from
15:55:51 <TrueBrain> ah; never seen it, so I guess so :p
15:56:09 <planetmaker> yeah... kinda useless IMO to make it necessary (good to have it, though)
15:56:40 <TrueBrain> If it is not required, people never fill it in
15:57:08 <planetmaker> that's not true. And also: it's not needed, if an error description follows. It's duplicate information IMO
15:57:35 <glx> arg new code still contains unresolved stuff
15:57:45 <TrueBrain> sadly, it is very true. Just pick bugs.openttd.org as case-study
15:59:34 <planetmaker> you're erecting a barrier to bug reports there with how mantis works now, with the required entries scattered over several monitor pages
15:59:47 <planetmaker> that's IMO bad interface design and not helping bug reports
16:00:07 <TrueBrain> That you should report to Mantis, but either way: I rather have completely good reports, than half bad reports
16:00:46 <TrueBrain> (and the report issue page is just one single page, so I have no clue what you mean with 'several monitor pages'
16:01:00 <planetmaker> yes, a single page with much scrolling...
16:01:07 <TrueBrain> besides that, only 3 (!) elements are required
16:01:10 <TrueBrain> out of the 10 or so
16:01:16 <planetmaker> and the necessary things scattered, not sequentially
16:01:41 <TrueBrain> I mean, how hard can it be .... worse: the category is the FIRST entry :p
16:02:05 <TrueBrain> either way, settings category to a default value has always given toms of errors and problems on openttd ... how many bugs/patches/features are mislabeled .. pff :p
16:02:08 <planetmaker> All I say is: you miss a perfectly good crash report due to making that a necessary one.
16:02:23 <TrueBrain> and all I say is: too bad; next time, read more carefully :)
16:02:39 <planetmaker> your loss. Not mine
16:02:45 <TrueBrain> it is required because it is a very required field
16:02:49 <TrueBrain> and there is little I can do to fix that ...
16:02:59 <TrueBrain> yes, Mantis has a crappy interface .. but it is the best out there
16:03:16 <planetmaker> honestly... flyspray looks better to me.
16:03:23 <TrueBrain> Flyspray is not under development
16:03:31 <planetmaker> I know. That's the bad news ;-)
16:03:40 <TrueBrain> no, that makes the software unusable (given the bugs)
16:04:12 <TrueBrain> and the amount of misfiling under flyspray is insane
16:04:19 <TrueBrain> people even manage to mistake 'Website' with 'Game'
16:04:30 <TrueBrain> people file things under 0.6.3, "because it is the default value"
16:04:37 <TrueBrain> people don't read .... people never read ...
16:04:59 <planetmaker> and no matter what you do: either you loose bug reports or you deal with that
16:05:08 <planetmaker> you obviously choose the first option.
16:05:19 <planetmaker> I would prefer the latter
16:05:35 <planetmaker> Even though I hat incomplete, badly categorized etc bug reports, too
16:05:45 <TrueBrain> well, really, if you can't fill in 3 required fields (which a clear error tells you when you do not), I can't help it :)
16:06:26 <glx> hmm I need to remember where I added breakpoints last night
16:06:34 <planetmaker> Given the upload time for a crash.bin I cannot be bothered to look at it uploading.
16:06:58 <planetmaker> And then the page is gone in the tabs. And it's too late and never gets submitted. Simple as that
16:07:34 <TrueBrain> don't get me wrong, I really appreciate a good bug report :) But if we make such fields optional (if it is possible at all), it would increase our workload immensely
16:07:40 <planetmaker> And I say: that's sad. As it's IMO caused by bad web interface.
16:07:52 <TrueBrain> So design a better bug-report system :) I think the world would welcome it ;)
16:08:34 <TrueBrain> Tools_AdjustToGameSpeed <- good name?
16:12:09 <glx> hoping it won't add new unresolved stuff :)
16:13:16 <glx> or if it does, not in emu_GameLoop_House ;)
16:18:30 <TrueBrain> will do in a sec .. first I need to make my work commitable I did during my boring lecture today :p
16:18:42 <TrueBrain> it is hard to type and make no noise :p
16:21:34 <TrueBrain> sorry, lazy today :p
16:21:44 <TrueBrain> yippie, my function works ;)
16:23:42 <DorpsGek> SVN: truebrain (r527) -Add: Tools_AdjustToGameSpeed() to C
16:32:04 <Xaroth> had 10 days vacation still left for this year
16:32:09 <Xaroth> so next month i'm taking up 8 of em
16:32:37 <Xaroth> 2 long weekends (fri off and monday after it), and no work from 24 dec till 4 jan :)
16:40:28 <DorpsGek> SVN: truebrain (r528) [JIT] -Add: mapped another 3 functions (crash-bins provided by glx)
16:40:44 <DorpsGek> SVN: truebrain (r529) -Update (r528): update decompiled code
16:41:22 <TrueBrain> more unresolved! :)
16:41:26 <TrueBrain> what cases are they anyway?
16:41:58 <TrueBrain> or you are just adding them without looking what they do? :)
16:42:20 <glx> .,.../* Unresolved jump */ emu_ip = 0x06CE; emu_last_cs = 0x1391; emu_last_ip = 0x06CE; emu_last_length = 0x0019; emu_last_crc = 0x6FCA; emu_call();
16:42:31 <glx> and looks like a never reached one
16:42:31 <TrueBrain> yes .. you gave me 1 unresolved which resolved in another unresolved :)
16:42:34 <TrueBrain> that made me laugh ;)
16:43:02 <TrueBrain> yes, it just never happened that 0x3 was 0xFF
16:43:14 <TrueBrain> (which means there was always a unit building)
16:43:48 <TrueBrain> I guess AI algorithms :p
16:44:18 <glx> luckily I know how to make it "resolve" :)
16:44:23 <TrueBrain> pretty big chunk btw, from 6CE to 720
16:44:31 <TrueBrain> sadly it means I need to do work :)
16:44:39 <TrueBrain> if you hurry, I can do that now, else it will be (much) later today
16:44:48 <TrueBrain> "Additionally, a 404 Not Found error was encountered while trying to use an ErrorDocument to handle the request." (at fed2k)
16:45:13 <TrueBrain> it seems to be code which issues another build (by the fast look at it)
16:45:19 <TrueBrain> so clearly some other code already took care of it :p
16:45:58 <TrueBrain> ha, I was right about it being an AI routine :)
16:47:16 <TrueBrain> k, I have to go; glx, I suggest you collect a lot of your lovely crash logs, and I will handle them all tonight :)
16:47:50 <planetmaker> (from an e-mail which modifies an issue) It seems to work correctly for closed issues
16:57:07 * boekabart really should write that auto-crashlog mailer
16:57:33 <glx> my crash are hand made ;)
16:57:56 <glx> no, breakpoints and memory changes
16:58:39 <glx> resolve unresolved jumps/calls
16:58:58 <glx> it's impossible to convert to C un incomplete function
16:59:44 <boekabart> yes but if you change the 'context' so that they are called, how can you be sure that the context is coherent?
16:59:59 <boekabart> so that strange stuff won't happen
17:00:20 <boekabart> anyway still - even those crashes could be auto-sent-in, I suppose?
17:03:18 <glx> and decompiler converts them to emu_ C
17:03:41 <glx> changing context just triggers some ifs
17:47:01 <Xaroth> i'll have a look at it :/
17:51:04 <planetmaker> Was my assumption correct that it's your task?
17:51:43 <Xaroth> yes, I'm more useful with php than I am with C :P
20:17:27 <Xaroth> planetmaker: think i fixed it
20:17:29 <Xaroth> or something like that
20:17:43 <Xaroth> else i'll really have to go try to find it
20:18:01 <planetmaker> well... hard to verify for me now ;-)
23:04:18 <TrueBrain> boekabart: auto-sent would be useful ;)
23:04:28 <TrueBrain> boekabart: stuff might not be coherent for that run, but for sure it decompiles the code ;)
23:11:30 <TrueBrain> glx: ghehe .. guess what .. it introduces another unresolved :p
23:11:34 <TrueBrain> let me see if I can fix that ;)
23:12:03 <TrueBrain> no .. seen the big new blob it introduces, it is very logic ;)
23:12:54 <glx> @base 10 16 (calc 0x720 - 0x6ce)
23:12:54 <DorpsGek> glx: base <fromBase> [<toBase>] <number>
23:13:07 <glx> @base 10 16 [calc 0x720 - 0x6ce]
23:13:22 <TrueBrain> he, that works?! Nice :)
23:13:57 <TrueBrain> why do you know more about my bot then I do .....
23:15:45 <TrueBrain> haha, fair point :)
23:15:58 <TrueBrain> but that was much more then I ever did ... I never realised it WOULD be possible :p
23:16:20 <DorpsGek> SVN: truebrain (r530) [JIT] -Add: mapped another 4 functions (crash-bins provided by glx)
23:16:31 <DorpsGek> SVN: truebrain (r531) -Update (r530): update decompiled code
23:16:34 <glx> anyway all unresolved stuff is in structure loop :)
23:16:37 <TrueBrain> there, I solved all the unresolved ;)
23:16:58 <TrueBrain> glx: and the fact it is unresolved, means it was never really called ;)
23:17:28 <glx> we will know why when I'll rewrite it :)
23:18:20 <TrueBrain> well .. in the structure loop I had the same, some unused blob, which seems to have reasonable code .. but they just implemented it somewhere else too, and that was called sooner :p
23:18:30 <TrueBrain> either way, good luck rewriting ;)
23:19:09 <glx> btw I can already say local variable 0A is write only :)
23:19:59 <TrueBrain> all those 7 locals are timers, like in Structure GameLoop
23:20:10 <TrueBrain> it is a pretty clever way :)
23:20:17 <TrueBrain> well .. 'timers' .. if they get 1, the timer is triggered
23:21:02 <glx> there's a part analysing reinforcement
23:21:40 <TrueBrain> the 0A timer goes every 5 ticks .. but indeed never used
23:22:19 <TrueBrain> btw, keep in mind that emu_dx is in fact the lower part of a 32bit integer
23:23:12 <TrueBrain> a bit .. annoying, in relevance to other ax/dx stuff
23:23:14 <TrueBrain> hmm .. my connection is dropping :(
23:23:20 <TrueBrain> oeh, after 49second no reply, it is back
23:23:23 <glx> some csip32 are "easy" to spot
23:23:28 <TrueBrain> IRC can recover big blackouts :p
23:24:03 <TrueBrain> lol, here the debug mode makes it drop out at a logic moment
23:24:10 <glx> ,...emu_ax = emu_get_memory16(emu_ss, emu_bp, -0x18);
23:24:10 <glx> ,...emu_orw(&emu_ax, emu_get_memory16(emu_ss, emu_bp, -0x16));
23:24:28 <TrueBrain> if (csip32 == 0) ;)
23:25:03 <TrueBrain> also the same for 32bit integers btw
23:25:22 <TrueBrain> which is confusing at first, to seperate csip32 from uint32 ;)
23:25:54 <glx> when you have a FindNext just before it's easier
23:26:07 <TrueBrain> I love FindFirst / FindNext
23:26:12 <TrueBrain> even more what I did with it to make it Find ;)
23:29:49 <TrueBrain> I do have to rewrite Free(), as it now takes on a csip32, which is stupid :p
23:29:53 <glx> hmm if I remove emu_GameLoop_House from 1391.c there's no function left
23:30:13 <TrueBrain> then remove the file ;)
23:30:32 <TrueBrain> I do that _all_ the time :p
23:33:14 <planetmaker> eh...? A fix for a "won't fix" issue, TrueBrain ? :-D
23:33:30 <planetmaker> or what is svn r530?
23:33:30 <TrueBrain> where/what/when/why?
23:33:41 <TrueBrain> r530 is a fix for problems glx gave me
23:33:51 <TrueBrain> (as the commit says)
23:34:01 <glx> and not a "real" problem ;)
23:34:24 <planetmaker> I got an e-mail...
23:37:57 <TrueBrain> that I use the number '4'
23:38:00 <TrueBrain> doesn't mean I refer to bug 4 :p
23:38:14 <TrueBrain> lets make it a rule the # is mandatory :)
23:38:24 <planetmaker> ah, that's it :-)
23:38:34 <TrueBrain> sorry about that planetmaker :)
23:38:40 <glx> ok first step done ("split" in emu_house.c and house.c) and it still compiles
23:38:51 <Xaroth> er, TrueBrain, it parses EVERY #<number> :/
23:38:55 <Xaroth> that's how mantis works
23:38:56 <planetmaker> no worries. I just wondered... and thought that something weired like that might have happened :-)
23:39:06 <TrueBrain> Xaroth: in the svn commit log :p
23:39:42 <TrueBrain> Xaroth: lets put it an other way .. php /var/www/bugs.opendune.org/htdocs/scripts/svn_opendune_checkin.php --msg="[JIT] -Add: mapped another 4 functions (crash-bins provided by glx)" --debug --author=truebrain --revision=530
23:40:55 <Xaroth> might need to fix the reges
23:41:21 <TrueBrain> I think we are now on the same page ;)
23:41:21 <planetmaker> with redmine we only link, if the number is preceeded by #
23:41:32 <TrueBrain> that is what I just suggested too, yes ;)
23:41:48 <planetmaker> and to revisions, if preceeded by r
23:43:18 <TrueBrain> k .. what would I do next ... 'script' or 'strings' (GetString and stuff)
23:46:42 <Xaroth> made it search up front!
23:47:03 <glx> ,...if (g_global->debugScenario) return; <-- I guess it makes sense ;)
23:47:10 <TrueBrain> sorry, it doesn't look fixed ;)
23:47:24 <glx> begining of GameLoop_House
23:47:26 <TrueBrain> glx: yup :) In debugScenario, the tickers do not run :)
23:47:40 <TrueBrain> but weird is that in GameLoop_Structure it was inside the loop ... I moved it before it, but it is after timer-handlers ..
23:47:51 <TrueBrain> in GameLoop_House I believe it is before the tickers?
23:47:56 <Xaroth> that's what i'm trying to fix now
23:48:51 <TrueBrain> lovely inconsistant ..... you got to love :p
23:49:27 <TrueBrain> sleep well planetmaker
23:56:50 <Xaroth> TrueBrain: how about now?
23:57:01 <TrueBrain> looks much better ;)
23:57:02 <Xaroth> i made it -require- ( ) around the fix list
23:57:13 <Xaroth> and they can only be RIGHT after the fix/add
23:58:07 <TrueBrain> I truly hate regexp
23:58:27 <TrueBrain> but .. scripts of strings? Do tell?
23:58:49 <TrueBrain> [00:43] <TrueBrain> k .. what would I do next ... 'script' or 'strings' (GetString and stuff)
23:59:13 <Xaroth> script will be more challenging, so i'd say that
23:59:30 <TrueBrain> but strings can be useful to understand a few key elements .. now they are just numbers :)
23:59:47 <TrueBrain> but scripts are more challenging :p
continue to next day ⏵