ScummVM IRC Log - 10 Jul 2017

[Back to Index]

[00:03] <m_kiewitz> they should have called it SVGA
[00:03] <m_kiewitz> or VESA or whatever
[00:03] <m_kiewitz> or well (hires)
[00:13] <-- Stormkeeper left irc: Remote host closed the connection
[00:15] --> Stormkeeper joined #scummvm.
[00:26] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[00:27] <tsoliman> media was indeed corrupt ... I didn't think this would happen ... makes me lose faith in other dumps
[00:27] <tsoliman> this could very well be HDD corruption since some of these datafiles were rescued from a failing HDD .. I think I'll just redump everything to be sure
[00:28] <snover> next time, write some par2s :)
[00:30] <tsoliman> some of these detection tables are very funny
[00:30] <snover> what do you mean?
[00:30] <tsoliman> resource.000 - 006 are all the same first-5000 md5
[00:32] <tsoliman> https://github.com/scummvm/scummvm/blob/master/engines/sci/detection_tables.h#L2192-L2195
[00:32] <tsoliman> https://github.com/scummvm/scummvm/blob/master/engines/sci/detection_tables.h#L2222-L2228
[00:32] <tsoliman> and so on .. you get the idea :)
[00:34] <snover> yes, the resources were often duplicated across disks
[00:37] <snover> i am not sure that such early games required hard drives
[00:38] <tsoliman> they didn't - often there would savegames in the floppies too
[00:38] <tsoliman> or edited settings
[00:39] <tsoliman> but these dumps aren't from floppies
[00:39] <tsoliman> they're from sierra collection CDs
[00:39] <tsoliman> and GOG
[00:39] <tsoliman> I was dumping GOG to see if it is any different than some CDs
[00:39] <tsoliman> and to my surprise, GOG datafiles had savegames in them .. often
[00:39] <tsoliman> and sometimes there were older versions of the game .. which is odd
[00:40] <tsoliman> they also seem to have edited the datafiles to remove phone numbers
[00:44] <snover> i would not wish to inherit an old sierra bbs phone number
[00:44] <snover> or tech support number
[00:45] <tsoliman> https://www.youtube.com/watch?v=6ktqiwck1tc
[00:48] <tsoliman> yeah .. GOG edited the VERSION file which used to contain 800 numbers and put "XXX XXX-XXXX" instead
[00:51] <snover> hey, thats the password to my luggage
[01:00] <tsoliman> hmm .. adding lsl6 and lsl6-cd gives the same description
[01:02] <tsoliman> (different targets but same "name" field)
[01:03] <tsoliman> ah there's a missing field
[01:15] --> GitHub14 joined #scummvm.
[01:15] <GitHub14> [scummvm] tsoliman pushed 1 new commit to master: https://git.io/vQi5m
[01:15] <GitHub14> scummvm/master b0ecbc9 Tarek Soliman: SCI: Fix detected name for lsl6-cd
[01:15] GitHub14 (GitHub14@192.30.252.40) left #scummvm.
[01:16] <tsoliman> I am assuming the "extra" field is populated with "CD" only when non-CD versions exist .. which is why GK2 doesn't have it ... in this case lsl6 has both normal (floppy?) and "CD" and "Hi-res"
[01:17] <tsoliman> but ignore the "Hi-res" variant since it has a different base target name
[01:32] <snover> the extra field is a string, the second field in the ADGameDescription
[01:33] <snover> then there is a separate ADGF_CD that puts -cd on the game id
[01:36] <snover> oh, sorry, i need to read logs from the top down :)
[01:37] <snover> assuming the comments are accurate, that seems like an appropriate change to make.
[01:37] Nick change: Stormkeeper -> Storm-AFK
[01:37] <snover> tsoliman: ^
[01:45] <tsoliman> yeah I was gonna just fix the one "CD" entry I owned
[01:46] <tsoliman> and then I decided to apply it to all the ones that have ADGF_CD
[01:46] <tsoliman> so the assumption isn't just the comments but the ADGF flags
[01:59] <-- Dominus left irc: Ping timeout: 260 seconds
[02:00] --> Dominus joined #scummvm.
[02:21] <-- dreammaster left irc:
[03:05] --> GitHub29 joined #scummvm.
[03:05] <GitHub29> [scummvm] criezy pushed 1 new commit to master: https://git.io/vQiNi
[03:05] <GitHub29> scummvm/master afc0ac6 Thierry Crozat: I18N: Regenerate translations data file
[03:05] GitHub29 (GitHub29@192.30.252.42) left #scummvm.
[03:08] <Lightkey> https://twitter.com/christinelove/status/884078805610106880 :-)
[04:00] <-- Lightkey left irc: Ping timeout: 258 seconds
[04:13] --> Lightkey joined #scummvm.
[04:20] --> omer_mor_ joined #scummvm.
[04:21] <-- omer_mor left irc: Ping timeout: 268 seconds
[04:38] <-- ced117 left irc: Ping timeout: 276 seconds
[04:39] --> ced117 joined #scummvm.
[04:41] --> Mia joined #scummvm.
[04:41] <-- Mia left irc: Changing host
[04:41] --> Mia joined #scummvm.
[05:21] --> waltervn joined #scummvm.
[05:21] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[05:23] <waltervn> morning
[05:43] <-- tsoliman left irc: Quit: I've been banished!
[05:44] --> tsoliman joined #scummvm.
[05:44] #scummvm: mode change '+o tsoliman' by ChanServ!ChanServ@services.
[05:46] <-- tsoliman left irc: Client Quit
[05:47] --> tsoliman joined #scummvm.
[05:47] #scummvm: mode change '+o tsoliman' by ChanServ!ChanServ@services.
[05:48] <-- tsoliman left irc: Client Quit
[05:49] --> tsoliman joined #scummvm.
[05:49] <-- tsoliman left irc: Client Quit
[05:50] --> tsoliman joined #scummvm.
[05:50] #scummvm: mode change '+o tsoliman' by ChanServ!ChanServ@services.
[06:12] --> GitHub4 joined #scummvm.
[06:12] <GitHub4> [scummvm] bgK pushed 1 new commit to master: https://git.io/vQPfe
[06:12] <GitHub4> scummvm/master 1141bfc Bastien Bouclet: MOHAWK: Riven: Clear the effects screen at the beginning of credits...
[06:12] GitHub4 (GitHub4@192.30.252.41) left #scummvm.
[06:15] <bgK> snover: I only noticed palette issues in Phantasmagoria in the basement and chapel during chapter 7
[06:47] <-- tsoliman left irc: Quit: I've been banished!
[06:48] --> tsoliman joined #scummvm.
[06:48] #scummvm: mode change '+o tsoliman' by ChanServ!ChanServ@services.
[07:01] <-- LittleToonCat left irc: Remote host closed the connection
[07:19] --> Strangerke|work joined #scummvm.
[07:20] <Strangerke|work> hi guys
[07:26] <-- klusark left irc: Ping timeout: 268 seconds
[07:32] --> klusark joined #scummvm.
[07:50] <-- TMM left irc: Quit: Ex-Chat
[08:27] <-- klusark left irc: Ping timeout: 255 seconds
[08:27] <-- Cheeseness left irc: Remote host closed the connection
[08:34] <wjp> morning
[08:36] --> klusark joined #scummvm.
[08:38] --> TMM joined #scummvm.
[08:38] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[08:40] --> t0by joined #scummvm.
[08:40] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[08:41] --> _sev|work joined #scummvm.
[08:41] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[08:59] <-- _sev|work left irc: Quit: This computer has gone to sleep
[09:00] --> _sev|work joined #scummvm.
[09:00] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[09:06] <-- _sev|work left irc: Quit: This computer has gone to sleep
[09:41] --> ajax16384 joined #scummvm.
[09:41] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[09:50] --> _sev|work joined #scummvm.
[09:50] <-- _sev|work left irc: Changing host
[09:50] --> _sev|work joined #scummvm.
[09:50] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[09:54] --> criezy|Work joined #scummvm.
[09:54] #scummvm: mode change '+o criezy|Work' by ChanServ!ChanServ@services.
[10:00] <-- |Cable| left irc: Ping timeout: 260 seconds
[10:12] --> |Cable| joined #scummvm.
[10:43] Nick change: Storm-AFK -> Stormkeeper
[11:34] <-- Mia left irc: Ping timeout: 240 seconds
[11:40] <-- LePhilousophe left irc: Ping timeout: 240 seconds
[11:41] --> LePhilousophe joined #scummvm.
[11:55] <-- _sev|work left irc: Quit: This computer has gone to sleep
[12:02] --> _sev|work joined #scummvm.
[12:02] <-- _sev|work left irc: Changing host
[12:02] --> _sev|work joined #scummvm.
[12:02] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[12:04] <-- _sev|work left irc: Client Quit
[12:46] <-- Stormkeeper left irc: Quit: KVIrc 4.9.2 Aria http://www.kvirc.net/
[12:57] --> Littleboy joined #scummvm.
[12:57] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[13:42] --> kgbme joined #scummvm.
[13:43] <kgbme> o/ hi, guys, need a hand: registered on the forums and received the pwd email, however no (other) email to confirm my account; can't log in, either (inactive message)
[13:44] <kgbme> wanted to comment about using Ghost Commander to add games on Android 6.0 with SD card formatted as internal storage, a pain. :)
[13:45] --> _sev|work joined #scummvm.
[13:45] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[13:50] <kgbme> _sev|work 3:43 PM <kgbme> o/ hi, guys, need a hand: registered on the forums and received the pwd email, however no (other) email to confirm my account; can't log in, either (inactive message)
[13:51] <kgbme> .. if you don't mind, pls. :)
[13:53] <criezy|Work> Hi kgbme
[13:54] <criezy|Work> I would advise just to stay in here and be patient. Hopefully somebody who knows how the forum registration works will notice and solve your issue eventually.
[13:54] <kgbme> hi!
[13:54] <criezy|Work> Unfortunately I have no idea how it works myself.
[13:54] <kgbme> np, ty (if irccloud doesn't throw me. :))
[13:55] <kgbme> *free acct. meh
[14:09] <-- kgbme left irc: Quit: Updating details, brb
[14:09] --> kgbme joined #scummvm.
[14:12] <-- kgbme left irc: Client Quit
[14:12] --> kgbme joined #scummvm.
[14:16] <Strangerke|work> kgbme: done
[14:17] <kgbme> ty, sec
[14:19] <kgbme> Strangerke|work: kgbme3@gmail.com ?
[14:19] <Strangerke|work> yes, that's the mail mentioned in the user profile
[14:19] <kgbme> yessir, won't let me in. hm
[14:21] <Strangerke|work> maybe there's a delay somewhre, it's weird. Didn't you receive a confirmation mail?
[14:22] <kgbme> https://usercontent.irccloud-cdn.com/file/qZtwAnDy/Screenshot_20170710-162157.png
[14:22] <kgbme> no email. :( only one came, on register, containing my acct. pwd
[14:23] <kgbme> said that confirmation is on its way
[14:24] <kgbme> it's been a while, but I can wait, all good - if that's the case; server cron having a coffee break :)
[14:30] <kgbme> ah, Jesus Christ
[14:31] <kgbme> Strangerke|work: sorry, guys.. it *did* end up in spam omg
[14:32] <kgbme> oh by no link, only text: "Your account on "ScummVM :: Forums" has now been activated, you may login using the username and password you received in a previous email."
[14:32] <kgbme> *bt
[14:33] <Strangerke|work> hum. And is it working at the end or not?
[14:33] <kgbme> inactive, so strange
[14:34] <kgbme> :(
[14:34] <Strangerke|work> damn it :(
[14:34] <kgbme> :f gah!
[14:36] --> Begasus joined #scummvm.
[14:38] <Strangerke|work> I was able to send you a pm, and everything is set to active... I don't get it
[14:38] <Strangerke|work> And it tells me you're already active when I try to force it
[14:39] <kgbme> test pm came, but can't log in :))))
[14:39] <kgbme> got the email
[14:39] --> Begas_VBox joined #scummvm.
[14:39] <kgbme> *all I did was reg through the browser, nothing else
[14:40] <Strangerke|work> yes, that's what you're supposed to do
[14:40] <kgbme> yeah....
[14:40] <Strangerke|work> I may try to disable it and activate it again... just in case
[14:40] <kgbme> *regular phpbb thing came up, no errors or anything
[14:41] <Strangerke|work> I just did that, could you try again?
[14:42] <kgbme> sec
[14:43] <kgbme> ok, PLS don't kill me.. dunno WTH is wrong with me today: was using email instead of username
[14:44] <kgbme> everything is cool, tyvm!
[14:44] <Strangerke|work> Arf :) Cool
[14:45] <Strangerke|work> Good thing you didn't hit an obscure bug in phpbb
[14:45] <Strangerke|work> (and your pm worked, too) ;)
[14:47] <kgbme> I'm just going to hide for a bit now and hope it blows over. xD
[14:49] <kgbme> gotta figure out why irccloud is disconnecting me, when free accts are supposed to have a 2hr window, bah
[14:50] <Strangerke|work> no need to hide. what's important is that it works at the end
[14:51] <kgbme> :))
[15:10] --> LittleToonCat joined #scummvm.
[15:14] --> GitHub12 joined #scummvm.
[15:14] <GitHub12> [scummvm] tobiatesan opened pull request #970: CMD: Fix batch adding of games (master...fix_massadd) https://git.io/vQPdj
[15:14] GitHub12 (GitHub12@192.30.252.45) left #scummvm.
[15:15] <kgbme> speaking of adding games, that's the (crazy) Android 6.0 issue that I'm trying to deal with, with SD card formatted as Internal Storage
[15:17] --> Henke37 joined #scummvm.
[15:19] <t0by> criezy|Work, this reminds me that I've never been able to get the confirmation email either :)
[15:19] <t0by> kgbme, is there a ticket for that already?
[15:20] <t0by> Nevermind, ended up in spam as well
[15:21] <criezy|Work> I created my account more than 10 years ago. So honestly I don't remember if I received a confirmation email or not :P
[15:21] <kgbme> t0by: so far I had just found a forum topic about using something like the Ghost Commander to manually insert the games to dir.
[15:21] <t0by> > Why is this message in Spam? It is in violation of Google's recommended email sender guidelines.
[15:21] <t0by> https://support.google.com/mail/answer/81126?hl=en#authentication
[15:21] <t0by> I'll file a bug for this
[15:21] <kgbme> oh, the forum.. I have no idea, first email came fine; second went to spam
[15:22] <t0by> kgbme, please feel free to open a bug on bugs.scummvm.org
[15:22] <kgbme> bt, it wasn't needed - no confirmation link to click?? I was just trying to log in with email, instead of username. xD
[15:23] <kgbme> t0by: ty, still looking to make sure not to make more mistakes. :)
[15:28] <-- _sev|work left irc: Quit: This computer has gone to sleep
[15:41] <-- demonimin left irc: Ping timeout: 260 seconds
[15:44] --> demonimin joined #scummvm.
[15:44] <-- demonimin left irc: Changing host
[15:44] --> demonimin joined #scummvm.
[15:53] <-- TMM left irc: Quit: Ex-Chat
[15:59] <-- ajax16384 left irc: Quit: Leaving
[16:20] --> ajax16384 joined #scummvm.
[16:20] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[16:25] --> ny00123 joined #scummvm.
[16:42] <-- criezy|Work left irc: Quit: Page closed
[17:06] --> WooShell joined #scummvm.
[17:06] <WooShell> meow =^.^=
[17:35] <-- ny00123 left irc: Quit: Leaving
[17:36] <t0by> woof
[17:43] --> Stormkeeper joined #scummvm.
[17:50] --> Strangerke joined #scummvm.
[17:50] #scummvm: mode change '+o Strangerke' by ChanServ!ChanServ@services.
[18:00] --> m_kiewitz joined #scummvm.
[18:00] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[18:11] <-- t0by left irc: Ping timeout: 240 seconds
[18:24] --> rain1 joined #scummvm.
[18:24] <rain1> hello!
[18:24] <rain1> how is the myst and riven playability?
[18:24] --> t0by joined #scummvm.
[18:24] #scummvm: mode change '+o t0by' by ChanServ!ChanServ@services.
[18:27] --> ny00123 joined #scummvm.
[18:27] <bgK> rain1: they're very playable in daily builds, but have a few minor bugs
[18:27] <rain1> exciting :D
[18:27] <rain1> i might give it a spin
[18:28] <bgK> have fun! if you find bugs that have not been reported yet, please post them on our bug tracker
[18:40] <-- kgbme left irc: Quit: Connection closed for inactivity
[18:49] --> SylvainTV joined #scummvm.
[18:49] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[19:09] <_sev_> criezy: here?
[19:11] <_sev_> let' have buildbot blown up
[19:11] --> GitHub80 joined #scummvm.
[19:11] <GitHub80> [scummvm] sev- closed pull request #967: COMMON: Change way the Singleton instances are instantiated (master...singleton-instantiation) https://git.io/vQuHm
[19:11] GitHub80 (GitHub80@192.30.252.34) left #scummvm.
[19:11] --> GitHub137 joined #scummvm.
[19:11] <GitHub137> [scummvm] sev- pushed 1 new commit to master: https://git.io/vQX0l
[19:11] <GitHub137> scummvm/master eefa72a Thierry Crozat: COMMON: Change way the Singleton instances are instantiated...
[19:11] GitHub137 (GitHub137@192.30.252.45) left #scummvm.
[19:11] --> GitHub147 joined #scummvm.
[19:11] <GitHub147> [scummvm] criezy pushed 1 new commit to master: https://git.io/vQX0B
[19:11] <GitHub147> scummvm/master c61c0cb Thierry Crozat: I18N: Update translations templates
[19:11] GitHub147 (GitHub147@192.30.252.34) left #scummvm.
[19:13] <t0by> :)
[19:15] <-- ny00123 left irc: Quit: Leaving
[19:15] <_sev_> criezy: okay, that broke things badly
[19:16] <_sev_> ConfigManager::getBool(dump_scripts,): '' is not a valid bool!
[19:17] <_sev_> and in general, ConfigManager is broken
[19:18] --> GitHub5 joined #scummvm.
[19:18] <GitHub5> [scummvm] sev- pushed 1 new commit to master: https://git.io/vQXER
[19:18] <GitHub5> scummvm/master 940b2a2 Eugene Sandulenko: Revert "COMMON: Change way the Singleton instances are instantiated"...
[19:18] GitHub5 (GitHub5@192.30.252.34) left #scummvm.
[19:18] --> GitHub53 joined #scummvm.
[19:18] <GitHub53> [scummvm] criezy pushed 1 new commit to master: https://git.io/vQXEE
[19:18] <GitHub53> scummvm/master a32ee25 Thierry Crozat: I18N: Update translations templates
[19:18] GitHub53 (GitHub53@192.30.252.42) left #scummvm.
[19:46] Nick change: _sev_ -> _sev
[19:48] <-- Strangerke|work left irc: Ping timeout: 260 seconds
[19:49] --> sh4rm4^bnc joined #scummvm.
[19:50] <sh4rm4^bnc> i downloaded "Igor Objetivo Uikokahonia (DOS CD, Spanish)" game, but scummvm doesnt recognize it. is it supported only in trunk, or is it mistakenly labeled as a scummvm supported game ?
[19:56] <logix> to quote the topic: "no warez"
[19:58] <m_kiewitz> No warez, no Game Requests, read the FAQ!
[20:22] <sh4rm4^bnc> i'm not talking about warez :) only asking if the game is supposed to be supported...
[20:23] --> GitHub198 joined #scummvm.
[20:23] <GitHub198> [scummvm] bgK opened pull request #971: VIDEO: Allow setting the mixer sound type used to play audio tracks (master...videodecoder-sound-type) https://git.io/vQXKv
[20:23] GitHub198 (GitHub198@192.30.252.34) left #scummvm.
[20:23] #scummvm: mode change '+b sk4!*@*' by _sev!~sev@scummvm/undead/sev
[20:25] #scummvm: mode change '+b *!*sh4rm4@*unaffiliated/sh4rm4' by _sev!~sev@scummvm/undead/sev
[20:25] sh4rm4^bnc kicked from #scummvm by _sev: No warez
[20:25] #scummvm: mode change '-b sk4!*@*' by _sev!~sev@scummvm/undead/sev
[20:31] <-- Strangerke left irc: Ping timeout: 240 seconds
[20:31] <criezy> I though Pendulo had made Igor freeware...
[20:32] <criezy> Ah, this is mentionned here: http://forums.scummvm.org/viewtopic.php?t=4859&postdays=0&postorder=asc&start=0
[20:32] <_sev> criezy: so, the patch did not work :(
[20:33] <criezy> So downloading it might be legal I suppose.
[20:33] <_sev> criezy: it compiles fine, though
[20:33] --> Strangerke joined #scummvm.
[20:34] <criezy> _sev: what was the problem?
[20:34] <criezy> I tested it quickly and it seemed to work for me. There was multiple instances in the binary but they are weakly linked and it uses only one of those I far as I could see, effectively making it a singleton.
[20:34] <_sev> criezy: maybe it is because I have plugins separatedly compiled?
[20:34] <_sev> nothing works for me
[20:35] <criezy> Oh, you are using plugins? I didn't try that...
[20:37] <wjp> hm, I did try to build with plugins, but maybe not run
[20:39] <-- ajax16384 left irc: Read error: Connection reset by peer
[20:42] <_sev> it built nicely
[20:42] <_sev> but not run
[20:42] <_sev> it cannot even read the gam path from the config
[20:44] Action: wjp tries something
[20:48] <Lightkey> https://www.youtube.com/watch?v=W68M26UjgAU yay, let's totally talk about piracy!
[20:50] <wjp> hm, our build system doesn't really support building plugin objects differently from main objects I guess
[20:54] <rain1> well what about unbanning that fellow then?
[20:55] <m_kiewitz> the ban was already removed
[20:56] <logix> the nick-ban was, but not the user@host ban
[20:56] <m_kiewitz> ah wait, no it wasnt
[20:56] <logix> I doubt that guy was aware that it was freeware
[20:56] <_sev> the naming suggests it comes as part of a bigger arez collection
[20:59] <wjp> _sev: oh, I had enabled plugins, but forgot to turn the engines into plugins...
[20:59] <wjp> oops
[20:59] <_sev> wjp: hehe
[21:00] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[21:01] <wjp> hm, it does work here it seems
[21:02] <wjp> (64 bit linux with gcc 4.9.3)
[21:03] <wjp> (and also with 6.3.0)
[21:04] <_sev> weird
[21:04] <_sev> Mac here
[21:04] <_sev> criezy: may you try it then?
[21:04] <_sev> clang
[21:04] <wjp> it does use linking magic to mark the symbols as globally unique
[21:05] <criezy> I can try it.
[21:05] <criezy> How do you turn engines into plugins? You need to list them all with --enable-engine-dynamic=...?
[21:05] <wjp> --default-dynamic
[21:06] <wjp> but you may want to just build one engine anyway
[21:06] <-- Begas_VBox left irc: Quit: Vision[0.9.8]: i've been blurred!
[21:07] <-- Begasus left irc: Quit: Ex-Chat
[21:09] <m_kiewitz> wjp: is there support so that all engines are in separate DLLs on Windows?
[21:09] <criezy> Indeed, it doesn't work here
[21:09] <criezy> with Apple LLVM version 6.0 (clang-600.0.57) (based on LLVM 3.5svn)
[21:09] <m_kiewitz> or I mean is there a reason for not doing so?
[21:11] <wjp> criezy: could you try this? template<class T> __attribute__ ((visibility("default"))) T *Singleton<T>::_singleton = 0;
[21:12] <wjp> m_kiewitz: I think the question should be reversed. Why do you want to?
[21:13] <m_kiewitz> just an idea. I mean we do support a plugin mode, so why not?
[21:13] <criezy> wjp: same result. It's still broken
[21:14] <m_kiewitz> I mean the .exe is currently around 43 mb, at some point it may be 100 mb or 200 mb.
[21:15] <m_kiewitz> i have never looked into the plugin system, so i have no idea how much work that would be
[21:20] <-- ST left irc: Disconnected by services
[21:20] --> ST1 joined #scummvm.
[21:28] <wjp> can anybody explain to me how this Titanic game works? I've just successfully reproduced 9880 by randomly clicking and saying things but I can't quite make sense of it (and fonts seem to randomly disappear?)
[21:28] <wjp> (this engine is really horribly slow in valgrind, by the way)
[21:31] <wjp> (_HORRIBLY SLOW_)
[21:31] <-- t0by left irc: Remote host closed the connection
[21:32] <wjp> time between typing a letter and it appearing is about a minute
[21:37] <wjp> hm, memory management in wave streams in titanic is fishy
[21:38] <wjp> and tt_parser accesses data after deleting it
[21:48] <waltervn> this reminds me that clone2727 pointed out an issue with the audio steams at some point, did we ever look into that?
[21:51] <-- waltervn left irc: Quit: Leaving
[22:06] <-- Dark-Star left irc: Ping timeout: 258 seconds
[22:07] <wjp> oh wow
[22:10] --> Dark-Star joined #scummvm.
[22:11] --> Cheeseness joined #scummvm.
[22:14] <wjp> so this is a pretty bad stack smashing bug
[22:23] <-- rain1 left irc: Quit: WeeChat 1.9
[22:26] <-- WooShell left irc: Quit: If you understand or if you don't, if you believe or if you doubt - There's a universal justice, and the eyes of truth are always watching you.
[22:28] --> deep-book-gk_ joined #scummvm.
[22:28] deep-book-gk_ (1wm_su@50.118.215.100) left #scummvm.
[22:42] <snover> wjp: at some point in the near future i will hopefully be running through titanic with ASan, which at least finds some memory bugs without being unusably slow
[22:44] <wjp> the main issues are a dreadful lack of inlining in titanic's surface functions and Common::Bitstream stuff
[22:44] <wjp> or in particular virtual function calls in our Common::Bitstream even when used on a MemoryReadStream
[22:45] <wjp> it manages to rack up several billion calls to things like Titanic::CTransparencySurface::getPixel, isPixelOpaque, ... and hundreds of millions to MemoryReadStream::eos, pos, read
[22:45] <wjp> in a minute or so
[22:45] <wjp> which is really just crazy
[22:47] <wjp> Common::Bitstream is pretty bad
[22:47] <wjp> that should be redesigned with some template magic to allow inlining MemoryReadStream functions
[22:48] <wjp> and Titanic::CTransparencySurface could use basic some inlining
[22:49] <wjp> a billion calls to a non-inlined avClipUint8...
[22:49] <wjp> I'm guessing nobody ran this through a profiler yet :-)
[22:50] <snover> i would imagine not. i noticed performance problems when i tested it during development, but figured those would get cleared up at some later date, since things were pretty incomplete at that point (videos still werent rendering transparency correctly, etc.)
[22:52] <snover> its too bad to hear about bitstream. i would be interested to learn more of the template magic you are thinking about. a specialisation for MemoryReadStream, or&?
[22:52] <wjp> yeah
[22:52] <wjp> haven't looked at it for more than a minute though
[22:52] <wjp> basically it now just has a SeekableReadStream*, so everything goes via virtual function calls
[22:53] <snover> i thought i was crazy for using CRTP on span. i mean, i might still be have been crazy for doing that, but i thought that CPUs would handle virtual calls fairly reasonably these days.
[22:54] <wjp> it's not inlining them that really hurts here I think
[22:54] <wjp> which would lead to re-use of intermediates
[22:55] --> dreammaster joined #scummvm.
[22:55] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[22:55] <wjp> hi
[22:55] <wjp> this basic inlining I'm doing now makes it at least semi-playable in valgrind
[22:56] <wjp> have either of you seen disappearing characters when typing or deleting (backspace) text rapidly?
[22:57] <wjp> (characters as in letters, not actors)
[22:58] <wjp> ok, I take back semi-playable. Video sequences are still horrible due to the indeo codec + bitstream
[22:58] <snover> i never got past the embarkation lobby so i cannot confirm that at the moment, though i am building now so i may be able to verify in a minute.
[22:59] <wjp> I'm playing from this savegame at 9880
[22:59] <wjp> which I think is still in the embarkation lobby?
[22:59] <snover> whoops. somehow im down to 3.2GiB of free disk space, so make that a few minutes.
[22:59] <dreammaster> Oh.. you're talking about Titanic. There may well be minor issues with rendering the characters due to recent changes I did to match the original to only redraw "dirty" parts of the scene
[22:59] <wjp> dreammaster: I left a fairly extensive comment on 9880
[22:59] <wjp> bad stack smashing happening there
[23:01] <dreammaster> Ah. Thanks. That certainy may help in tracking down the problem. Particularly given the trouble I had replicating it under Visual Studio
[23:01] <wjp> so the most called function is now actually Titanic::CDropTarget::DropObjectMsg
[23:02] <wjp> also at hundreds of millions of times a minute
[23:03] <dreammaster> That's ironic, considering I've only just been working on it. The nose holder in the parrot room uses a CDropTarget, and there are various graphical issues I've been trying to resolve
[23:03] <dreammaster> That's definitely excessive
[23:05] <dreammaster> Is i doing something specifically?
[23:05] <wjp> I'm a bit suspicious of this profiling run
[23:06] <wjp> doing a clean build now
[23:07] <dreammaster> It may be worthwhile. Unless I'm missing something, a CDropObjectMsg instance is only created when dropping an object on another one, and it only calls a single designated target
[23:07] <wjp> I guess something got out of sync. It's MemoryReadStream::read now
[23:08] <wjp> in any case, you may want to move all functions from transparency_surface.cpp except the constructor to the header to have them inlined
[23:08] <wjp> those cost significant amounts of time for no reason now
[23:09] <dreammaster> Ok, if you think that will help
[23:09] <wjp> s/think/know/ :-)
[23:09] <wjp> I've been profiling it
[23:10] <dreammaster> Then consider it done. I'll stash my current work and do it now
[23:10] <wjp> no hurry
[23:10] <snover> ugh, spotify is such a f'ing dumpster fire of software. found 12GiB of space! building now.
[23:11] <dreammaster> Meh. It'll be a welcome procrastination. I'm currently in a tug of war with how the movie handles transparencies.
[23:11] <wjp> after transparency_surface, the next major bottleneck is the indeo/Bitstream stuff
[23:12] <dreammaster> Either I break how the Doorbot in the original looks (because it has partial transparencies over several frames), or I end up with the Nose item in the parrot room staying there even after I pick it up..
[23:12] <dreammaster> ..because it sets a frame that's simply completely transparent, and with my current code, that means the previous display of the item remains in place :P
[23:15] <wjp> next one, writeChar() is called an average of 2000 times per writeString() call? (If I'm reading this right? Can that make sense?)
[23:15] <dreammaster> Hmmm.. depends on the size of the text. On average, writeString handles full text responses, which can be multiple sentences at a time.
[23:17] <dreammaster> In fact, for the conversation log, it would be rendering up to a full screen's worth of text from the history of conversations, so 2000 is probably reasonable
[23:19] <wjp> writeString/writeChar is called with tons of text that I don't see ending up on screen
[23:22] <dreammaster> Hmmm.. been a while since I worked on the conversation log. It may be the way it figures out what part of the conversation log is being displayed, Since writeChar returns the status of whether writing the char would be in or out of bounds
[23:22] <wjp> not really a problem since it's quite far below the graphics stuff in run time, but it sticks out a bit in number of calls
[23:23] <dreammaster> At the very least, it doesn't call the copyRect function in it to actually blit anything if the character would be outside the text rendering area
[23:24] --> GitHub162 joined #scummvm.
[23:24] <GitHub162> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vQXh7
[23:24] <GitHub162> scummvm/master 7fc69eb Paul Gilbert: TITANIC: Inline CTransparencySurface methods at wjp's suggestion
[23:24] GitHub162 (GitHub162@192.30.252.40) left #scummvm.
[23:24] <wjp> great, thanks :-)
[23:28] <dreammaster> And thanks likewise for the suggestion
[23:30] <criezy> _sev, wjp: the plugin issue on macOS with my singleton change is due to the use of an exported symbol list
[23:30] <criezy> https://github.com/scummvm/scummvm/blob/master/configure#L3650
[23:30] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[23:30] <criezy> As a result it exports only 4 symbols: _PLUGIN_getVersion, _PLUGIN_getType, _PLUGIN_getTypeVersion and _PLUGIN_getObject
[23:31] <criezy> Everything works well if I remove that -exported_symbol_list "$(srcdir)/plugin.exp"
[23:33] <wjp> interesting. That's what I was trying (slightly more specifically :-) ) with that visibility attribute, but I guess this symbols list could override that
[23:38] <snover> wjp: yes, i see characters being half-erased (or half-drawn), though i am having a hard time taking a screenshot since it corrects itself a few frames after i stop holding down a key
[23:39] <snover> ok, here we are. https://zetafleet.com/i/5964103d287f5.jpg
[23:39] <wjp> criezy: I'm starting to think this singleton clean up is not worth the linker headaches
[23:40] <wjp> criezy: especially on platforms less easily accessible to us
[23:40] <dreammaster> Rightio.. I'll look into it
[23:40] <wjp> snover: right, that looks like what I'm seeing too
[23:40] <criezy> wjp: I will comment on the PR with my findings, but I am starting to think the same thing.
[23:40] <criezy> It looks like we do the same for elf with --retain-symbols-file,$(srcdir)/backends/plugins/elf/plugin.syms but I don't know if this has the same consequences.
[23:41] <dreammaster> Hah. dafioram created a bug for it an hour ago. :)
[23:43] <wjp> oh indeed
[23:43] --> GitHub150 joined #scummvm.
[23:43] <GitHub150> [scummvm] wjp pushed 1 new commit to master: https://git.io/vQ1eu
[23:43] <GitHub150> scummvm/master 4ec3c28 Willem Jan Palenstijn: IMAGE: Inline two simple indeo functions
[23:43] GitHub150 (GitHub150@192.30.252.45) left #scummvm.
[23:43] <snover> i guess dafioram needs to get two cakes now
[23:44] <wjp> now we just need a volunteer for BitStream :-)
[23:44] <wjp> ...but look at the time
[23:44] <wjp> night!
[23:46] <dreammaster> Cya
[00:00] --- Tue Jul 11 2017