[Back to Index]
[00:05] <Scummette> scummvm/master 621ce16 IlDucci: I18N: Update translation (Spanish)
[00:05] <Scummette> scummvm/master 76bbba2 antoniou79: I18N: Update translation (Greek)
[00:05] <Scummette> [scummvm] criezy pushed 1 new commits to master: https://git.io/JtXQk
[00:05] <Scummette> scummvm/master 075cd31 criezy: AGS: Do case insensitive search for game files and sub-directories
[00:32] --> Mia joined #scummvm.
[00:32] <-- Mia left irc: Changing host
[00:32] --> Mia joined #scummvm.
[00:34] <-- Axy left irc: Ping timeout: 240 seconds
[00:40] <-- criezy left irc: Remote host closed the connection
[00:41] --> criezy joined #scummvm.
[00:41] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[00:44] <grogbot> <timofonic> igdb.com is used by Playnite, but there's a zillion of similar services. Maybe mobygames.com could be used for it too.
[00:46] <-- criezy left irc: Ping timeout: 260 seconds
[00:47] <grogbot> <timofonic> @BeleG I was half sleep while writing. You were faster than me! ð
[00:58] --> criezy joined #scummvm.
[00:58] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[00:59] <-- criezy left irc: Remote host closed the connection
[00:59] --> criezy joined #scummvm.
[00:59] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[01:09] <-- |Cable| left irc: Ping timeout: 264 seconds
[01:15] <Tea> That last AGS commit prevented The Shivah from loading with scummvm: common/fs.cpp:112: Common::String Common::FSNode::getPath() const: Assertion `_realNode' failed.
[01:18] --> |Cable| joined #scummvm.
[01:20] <-- criezy left irc: Remote host closed the connection
[01:21] --> criezy joined #scummvm.
[01:21] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[01:26] <-- criezy left irc: Ping timeout: 272 seconds
[01:38] --> criezy joined #scummvm.
[01:38] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[01:48] <-- criezy left irc: Ping timeout: 260 seconds
[02:18] <-- TMM left irc: Remote host closed the connection
[02:19] --> TMM joined #scummvm.
[02:19] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[02:25] <-- |Cable| left irc: Ping timeout: 240 seconds
[02:52] --> |Cable| joined #scummvm.
[02:54] --> DominusExult joined #scummvm.
[02:54] <-- DominusExult left irc: Changing host
[02:54] --> DominusExult joined #scummvm.
[02:57] <-- Dominus left irc: Ping timeout: 272 seconds
[02:57] Nick change: DominusExult -> Dominus
[02:59] --> criezy joined #scummvm.
[02:59] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[03:12] <Scummette> [scummvm] criezy pushed 1 new commits to master: https://git.io/JtXNA
[03:12] <Scummette> scummvm/master c86355d criezy: AGS: Fix assert when trying to open file that does not exist
[03:14] <-- criezy left irc: Remote host closed the connection
[03:14] --> criezy joined #scummvm.
[03:14] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[03:18] <-- criezy left irc: Remote host closed the connection
[03:19] --> criezy joined #scummvm.
[03:19] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[03:27] <-- criezy left irc: Remote host closed the connection
[03:29] --> criezy joined #scummvm.
[03:29] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[03:34] <-- criezy left irc: Ping timeout: 272 seconds
[03:47] --> criezy joined #scummvm.
[03:47] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[03:52] <-- criezy left irc: Ping timeout: 264 seconds
[04:17] --> criezy joined #scummvm.
[04:17] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[04:29] <-- criezy left irc: Ping timeout: 272 seconds
[05:29] <ScummBot> Port build status changed with c86355d5: Failure: master-amigaos4
[05:57] <-- cd left irc: Quit: cd
[06:16] <-- SylvainTV left irc: Read error: Connection reset by peer
[06:17] --> Axy joined #scummvm.
[06:17] --> SylvainTV joined #scummvm.
[06:17] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[06:20] <-- Mia left irc: Ping timeout: 272 seconds
[06:21] --> criezy joined #scummvm.
[06:21] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[06:25] <-- criezy left irc: Ping timeout: 272 seconds
[06:52] --> criezy joined #scummvm.
[06:52] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[06:57] <-- criezy left irc: Ping timeout: 256 seconds
[07:45] <grogbot> <timofonic> Tea There's a ScummVM Discord server with an AGS channel
[07:48] --> palasso joined #scummvm.
[08:04] --> borosky joined #scummvm.
[08:19] --> Shine_ joined #scummvm.
[09:16] --> ldevulder_ joined #scummvm.
[09:20] <-- ldevulder left irc: Ping timeout: 246 seconds
[10:37] --> ldevulder joined #scummvm.
[10:40] <-- ldevulder_ left irc: Ping timeout: 272 seconds
[11:04] <grogbot> <MarioKartFan> Got excited (probably foolishly) that recent submission from @ccawley2011 might have fixed 3DS nightlies. Unfortunately they still crash on launch. Still, good to see support for the platform!
[11:20] --> JohnnyonFlame joined #scummvm.
[11:24] <Scummette> [scummvm] ScummVM-Translations pushed 1 new commits to master: https://git.io/Jt1GM
[11:24] <Scummette> scummvm/master 2d782d0 losernator: I18N: Update translation (Korean)
[11:25] <grogbot> <antoniou79> @MarioKartFan did you build ScummVM on your own or tried the daily build? Maybe the development build from the buildbot does not yet have the fix?
[11:27] <-- Lightkey left irc: Ping timeout: 240 seconds
[11:35] --> criezy joined #scummvm.
[11:35] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[11:40] <grogbot> <lephilousophe> no, it's not related
[11:40] <grogbot> <lephilousophe> it's just a change of namespace name
[11:40] <-- criezy left irc: Ping timeout: 272 seconds
[11:40] --> Lightkey joined #scummvm.
[11:41] <grogbot> <lephilousophe> @MarioKartFan I looked at the crash dump yesterday (ripped off some code from LUMA to parse it...) but it's not obvious what fails
[11:41] <grogbot> <lephilousophe> can you tell us which version you used? that's present in the folder name of the package
[11:43] <ScummBot> Port build status changed with 2d782d0d: Success: master-amigaos4
[12:03] <-- JohnnyonFlame left irc: Read error: Connection reset by peer
[12:04] --> JohnnyonFlame joined #scummvm.
[12:04] <-- JohnnyonFlame left irc: Read error: Connection reset by peer
[12:09] --> criezy joined #scummvm.
[12:09] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[12:11] --> vliaskov joined #scummvm.
[12:17] <-- criezy left irc: Ping timeout: 260 seconds
[12:24] --> criezy joined #scummvm.
[12:24] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[12:31] <-- criezy left irc: Ping timeout: 265 seconds
[12:39] --> cd joined #scummvm.
[12:43] --> criezy joined #scummvm.
[12:43] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[12:47] <-- criezy left irc: Ping timeout: 246 seconds
[13:17] --> criezy joined #scummvm.
[13:17] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[13:22] <-- criezy left irc: Ping timeout: 264 seconds
[13:46] --> criezy joined #scummvm.
[13:46] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[13:50] <-- criezy left irc: Ping timeout: 272 seconds
[13:51] --> criezy joined #scummvm.
[13:51] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[14:13] <-- criezy left irc: Remote host closed the connection
[14:13] --> criezy joined #scummvm.
[14:13] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[14:18] <-- criezy left irc: Ping timeout: 246 seconds
[14:26] --> criezy joined #scummvm.
[14:26] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[14:28] <-- criezy left irc: Remote host closed the connection
[14:29] --> criezy joined #scummvm.
[14:29] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[14:33] <-- criezy left irc: Ping timeout: 240 seconds
[14:46] --> criezy joined #scummvm.
[14:46] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[14:51] <-- criezy left irc: Ping timeout: 272 seconds
[15:20] <-- _sev left irc: Quit: This computer has gone to sleep
[15:23] --> _sev joined #scummvm.
[15:23] #scummvm: mode change '+o _sev' by ChanServ!ChanServ@services.
[15:31] <-- palasso left irc: Remote host closed the connection
[15:35] <-- _sev left irc: Quit: This computer has gone to sleep
[15:35] --> criezy joined #scummvm.
[15:35] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[15:38] --> palasso joined #scummvm.
[15:39] --> _sev joined #scummvm.
[15:39] <-- _sev left irc: Changing host
[15:39] --> _sev joined #scummvm.
[15:39] #scummvm: mode change '+o _sev' by ChanServ!ChanServ@services.
[15:43] <-- criezy left irc: Remote host closed the connection
[15:44] --> criezy joined #scummvm.
[15:44] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[15:45] <-- _sev left irc: Quit: This computer has gone to sleep
[15:46] --> _sev joined #scummvm.
[15:46] #scummvm: mode change '+o _sev' by ChanServ!ChanServ@services.
[15:58] <Scummette> [scummvm] ZvikaZ opened pull request #2779: COMMON: Add String::forEachLine and convertBiDiStringByLines (master...z_bidi_multi_line) https://git.io/Jt1gI
[16:13] <grogbot> <chrisis123> Hi, wondering if there are any Nintendo Switch port testers here? The last few daily builds from the buildbot crash silently after the UI shows up for a few seconds (no idea if there is a crash report anywhere I could submit). I use Atmosphere and start it via the Homebrew Launcher. The last version I have tried that still works fine is switch-master-f548bdae (maybe some slightly later build works too, I haven't tried), the error
[16:13] <grogbot> happens at least a few builds ago (I think I've tried builds from Feb 12 or 13 that failed).
[16:15] <grogbot> <lowtraxx> I will try after work.
[16:25] <grogbot> <MarioKartFan> Hi @lephilousophe I ran the .3dsx of the latest buildbot build. I have not tried building it on my own.
[16:25] <-- _sev left irc: Quit: This computer has gone to sleep
[16:29] --> _sev joined #scummvm.
[16:29] <-- _sev left irc: Changing host
[16:29] --> _sev joined #scummvm.
[16:29] #scummvm: mode change '+o _sev' by ChanServ!ChanServ@services.
[17:17] <-- criezy left irc: Remote host closed the connection
[17:18] --> criezy joined #scummvm.
[17:18] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[17:22] <-- criezy left irc: Ping timeout: 240 seconds
[17:25] --> criezy joined #scummvm.
[17:25] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[17:27] --> ny00123 joined #scummvm.
[17:32] <grogbot> <lephilousophe> latest of which date?
[17:34] --> JohnnyonFlame joined #scummvm.
[17:51] <-- criezy left irc: Remote host closed the connection
[17:51] --> criezy joined #scummvm.
[17:51] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[17:56] <-- criezy left irc: Ping timeout: 265 seconds
[18:14] --> Begasus joined #scummvm.
[18:18] --> criezy joined #scummvm.
[18:18] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[18:19] <-- criezy left irc: Remote host closed the connection
[18:20] --> criezy joined #scummvm.
[18:20] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[18:23] <-- TAS_2012v left irc: Remote host closed the connection
[18:24] --> TAS_2012v joined #scummvm.
[18:33] --> ajax16384 joined #scummvm.
[18:33] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[18:33] <-- criezy left irc: Read error: Connection reset by peer
[18:40] --> criezy joined #scummvm.
[18:40] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[18:45] <-- criezy left irc: Ping timeout: 272 seconds
[18:57] --> criezy joined #scummvm.
[18:57] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[18:58] --> Deledrius_ joined #scummvm.
[19:02] <-- Deledrius left irc: Ping timeout: 240 seconds
[19:03] <-- Deledrius_ left irc: Ping timeout: 240 seconds
[19:08] <-- Shine_ left irc: Read error: Connection reset by peer
[19:10] <-- criezy left irc: Ping timeout: 265 seconds
[19:18] <grogbot> <MarioKartFan> @lephilousophe Today.
[19:24] <LePhilousophe> ok!
[19:47] --> criezy joined #scummvm.
[19:47] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[19:51] <-- criezy left irc: Ping timeout: 272 seconds
[19:54] <grogbot> <SupSuper> https://www.vogons.org/viewtopic.php?f=7&p=937740 this looks interesting
[19:55] <grogbot> <SupSuper> pseudo-decompiler for dos exes
[20:04] --> criezy joined #scummvm.
[20:04] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[20:04] <grogbot> <lephilousophe> @MarioKartFan just tried latest build with Citra and it works ðĶ
[20:06] <grogbot> <SupSuper> i think @rsn8887 is the one responsible for switch port
[20:09] <grogbot> <rsn8887> Will have to test later.
[20:11] <grogbot> <chrisis123> Thanks @rsn8887 just wanted to let you know as I don't think it has been reported yet (I use the daily builds instead of the stable build due to playing The Longest Journey, the build mentioned before from February 8th is working fine for me)
[20:23] <grogbot> <lephilousophe> @MarioKartFan as I don't own a 3DS I won't be able to help further
[20:23] <-- Begasus left irc: Quit: Ex-Chat
[20:24] <grogbot> <MarioKartFan> Ok. Well thanks anyway. I think BgK used to maintain the port but not sure if hes still around.
[20:34] <grogbot> <rsn8887> @chrisis123 I can reproduce the error on Switch. Somebody must have touched code recently that causes Switch to crash. No clue what it is.
[20:34] <grogbot> <rsn8887> The GUI pops up and vanishes after 2 seconds or so, program exits.
[20:35] <grogbot> <lephilousophe> plugins loading?
[20:35] <grogbot> <rsn8887> Switch is a static build
[20:36] <grogbot> <lephilousophe> ok so no ð
[20:38] <-- criezy left irc: Remote host closed the connection
[20:38] <grogbot> <lephilousophe> @MarioKartFan just to be sure it's not a problem with buildbot toolchain, could you please try this package? https://al.scummvm.net/snapshots/master/3ds-master-c86355d5.zip
[20:38] <grogbot> <lephilousophe> it's built using a new toolchain with uptodate devkitARM
[20:38] --> criezy joined #scummvm.
[20:38] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[20:41] <-- ajax16384 left irc: Quit: Leaving
[20:41] <grogbot> <rsn8887> Hmm pretty hard crash, maybe involving GPU
[20:41] <grogbot> <rsn8887> on Switch. Log is empty.
[20:43] <-- criezy left irc: Ping timeout: 272 seconds
[20:44] <grogbot> <lephilousophe> @MarioKartFan if it's still failing, I just see you already did some debug session with bgK earlier (https://bugs.scummvm.org/ticket/11292). We could try to do the same thing, results would be more exploitable. For this I would need to provide you a debug build. Don't know when I would have time for this
[20:52] <grogbot> <chrisis123> Yeah I couldn't find any log either, also I noticed that after the ScummVM crash the whole system becomes unstable and I get regular system crashes that force me to reboot (it's fine after the reboot though)
[20:53] --> criezy joined #scummvm.
[20:53] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[20:56] --> Shine_ joined #scummvm.
[20:57] <grogbot> <MarioKartFan> @lephilousophe Thanks for offering, but the package you linked to crashes at the same time.
[21:01] --> Deledrius joined #scummvm.
[21:03] <grogbot> <SupSuper> @rsn8887 a git bisect might help
[21:05] <-- Shine_ left irc: Read error: Connection reset by peer
[21:05] <grogbot> <chrisis123> Hmm, I just tried the current build with a new filename/folder and it didn't crash immediately (without any games in it), I was able to add a game and it ran. When I opened it again (with a game in the list) it crashed again... not sure what the reason might be.
[21:08] --> Deledrius_ joined #scummvm.
[21:08] <-- Deledrius left irc: Ping timeout: 264 seconds
[21:13] --> Deledrius joined #scummvm.
[21:15] <-- Deledrius_ left irc: Ping timeout: 272 seconds
[21:23] <grogbot> <rsn8887> I build a version from before the many commits by @Henke37 and that works.
[21:23] <grogbot> <rsn8887> EG f1d753c1c4f061c43a works
[21:24] <grogbot> <rsn8887> current master crashes
[21:24] <grogbot> <Henke37> you are referring to the PauseToken refactoring i presume?
[21:25] <grogbot> <rsn8887> I am not sure. Something between 2d782d0d5258c1061b6a89731d755772e6adad1d and f1d753c1c4f061c43a broke the switch build
[21:25] <grogbot> <Henke37> can't say i've memorized the commit hashes...
[21:28] <grogbot> <Henke37> no, that's much more recent.
[21:29] <grogbot> <Henke37> this is likely about the hashmap and confman refactoring
[21:30] <grogbot> <Henke37> it's stricter than before, so previous use may now be considered misuse and detected with an error()
[21:36] <grogbot> <Henke37> which is naturally going to be rather difficult to diagnose if it happens before said error() function is fully initialized.
[21:38] <grogbot> <rsn8887> Let me check one more thing
[21:39] <grogbot> <rsn8887> You are saying new confman behavior is not compatible to old?
[21:39] <grogbot> <rsn8887> I am trying the commit after your commits to see if it works
[21:41] <grogbot> <rsn8887> There was a modern GCC fix afterwards, which is needed to make any of those compile on Switch
[21:42] <grogbot> <Henke37> i suppose i am yes, that some previous behavior is no longer supported.
[21:42] <grogbot> <Henke37> in particular, the changes requires that a called be explicit about how to handle a read of a value that is not present
[21:43] <grogbot> <Henke37> does it expect it to always be present or does it expect a default value.
[21:44] <grogbot> <SupSuper> @Henke37 maybe you missed some code outside of the msvc project? i had to fix the tests for example
[21:45] <grogbot> <Henke37> not impossible. the buildbot should check that it at least compiles, but it can't check that the correct behavior was selected
[21:45] <grogbot> <rsn8887> I can compile it just crashes after GUI appears for 2 secs
[21:45] <grogbot> <Henke37> that is unusual. you'd expect it to fail right away, not after a short amount of time
[21:46] <grogbot> <rsn8887> yeah maybe something else
[21:46] <grogbot> <rsn8887> I am still trying that other commit
[21:46] <grogbot> <Henke37> not just for my changes, but for errors in general. they don't tend to happen while the program is "stable"
[21:47] <grogbot> <Henke37> there is nearly always some sort of perturbation
[21:47] <grogbot> <criezy> @rsn8887 If that is related to the changes by @Henke37, that would be the relevant pull request for context: https://github.com/scummvm/scummvm/pull/2548
[21:49] <grogbot> <rsn8887> Ok it is something between f27bf85bf761b895 and f1d753c1c4f0 so Henke stuff
[21:49] <grogbot> <rsn8887> But cannot easily bisect because needs that gcc fix.
[21:50] <grogbot> <Henke37> i presume that the "gcc fix" is a commit related to my pr too
[21:50] <grogbot> <rsn8887> This thing is massive and should have been ensured to stay compatible
[21:50] <grogbot> <rsn8887> I dont get why it was merged honestly
[21:51] <grogbot> <Henke37> being clearer about desired behavior is a desirable thing.
[21:52] <grogbot> <Henke37> it is supposed to catch bugs which were previously hidden by the unclear desired behavior.
[21:52] <grogbot> <Henke37> it is expected that it has teething problems.
[21:54] <grogbot> <rsn8887> What am I looking for?
[21:54] <grogbot> <rsn8887> GetVal? There's no getVal here. https://github.com/scummvm/scummvm/blob/5a313198532b96aacfe5ed9637a9f0fbf76f8a93/backends/platform/sdl/switch/switch.cpp
[21:55] <grogbot> <rsn8887> In the PR it said getVal can crash? I don't see any Switch specific geval
[21:55] <grogbot> <Henke37> you are correct, that is the member function targeted by the pr
[21:56] <grogbot> <Henke37> however, it is also called indirectly by member functions such as getBool
[21:59] <grogbot> <rsn8887> I use a ton of getBool on Switch, what shall I do?
[21:59] <grogbot> <rsn8887> Oh no only setBool
[22:00] <grogbot> <criezy> Doing a quick grep on ConfMan in backends/platform/sdl/switch/ doesn't show any suspicious one. The only one that could crash is he return ConfMan.getBool("touchpad_mouse_mode"); in getFeatureState(), but there is also a ConfMan.registerDefault("touchpad_mouse_mode", false); in initBackend(), so that should be fine.
[22:01] <grogbot> <criezy> The issue might be outside of backends/platform/sdl/switch/.
[22:01] --> Shine_ joined #scummvm.
[22:01] <grogbot> <Henke37> i checked getBool in particular, it chains to getValOrDefault via ConfigManager::get
[22:04] <grogbot> <Henke37> of course, there is the issue of this member function: cpp const String &ConfigManager::get(const String &key) const { if (_transientDomain.contains(key)) return _transientDomain[key]; else if (_activeDomain && _activeDomain->contains(key)) return (*_activeDomain)[key]; else if (_appDomain.contains(key)) return _appDomain[key]; return _defaultsDomain.getValOrDefault(key); }
[22:05] <grogbot> <Henke37> operator [] here assumes that the key does exist.
[22:05] <grogbot> <Henke37> which is rather reasonable, since it was just verified that it does in fact exist.
[22:08] <grogbot> <rsn8887> It is not the touchpad mouse mode getbool just tested that
[22:10] <grogbot> <Henke37> it could also be a generic hashmap, those were affected too
[22:12] <grogbot> <rsn8887> are there backend specific ones?
[22:13] <grogbot> <criezy> Do you have the ability to see the standard output somewhere when running on the switch?
[22:15] <grogbot> <criezy> Trying to access a key that does not exist should call unknownKeyError(), which calls error(). And since you see the GUI for a few seconds I assume the OSystem instance is initialized, so this should call OSystem::logMessage.
[22:15] <grogbot> <rsn8887> it can be done with nxlink, But I have to set it up in the code.
[22:15] <grogbot> <criezy> And on the Switch this seems to do a printf.
[22:15] <grogbot> <rsn8887> But the log is basically empty
[22:16] <grogbot> <rsn8887> The .log file I mean
[22:16] <grogbot> <criezy> Rather than look blindly this might be a better approach because the error should tell you which key is the culprit.
[22:16] <grogbot> <rsn8887> Ok
[22:16] <grogbot> <rsn8887> I will try to setup nxlink then
[22:18] <grogbot> <criezy> Do you usually have things in the log file on Switch?
[22:18] <grogbot> <criezy> OSystem_Switch reimplements logMessage and doesn't print in the log file.
[22:21] <grogbot> <criezy> Isn't the implementation from OSystem_SDL working on the switch? I don't really understand why it is reimplemented for the switch.
[22:24] <grogbot> <criezy> It seems to me if you have a log file on the switch, you should let logMessage() print the message to it. That could help to debug these kind of issues...
[22:24] <grogbot> <rsn8887> You can see it is just overwriting a few functions?!?
[22:24] <grogbot> <rsn8887> It is inheriting I think, I forgot.
[22:24] <Scummette> [scummvm] somaen pushed 7 new commits to master: https://git.io/Jt15P
[22:24] <Scummette> scummvm/master f9c1f62 somaen: TINSEL: Partially implement scene loading for Discworld Noir
[22:24] <Scummette> scummvm/master 0831a53 somaen: TINSEL: Map EVENT
[22:24] <Scummette> scummvm/master e583392 somaen: TINSEL: Map and stub SET3DTEXTURE
[22:24] <grogbot> <rsn8887> Also it is surface SDL afaik
[22:24] <Scummette> [scummvm] ScummVM-Translations pushed 1 new commits to master: https://git.io/Jt15X
[22:24] <Scummette> scummvm/master e817a2f : I18N: Update translation (Japanese)
[22:24] <grogbot> <rsn8887> Including OpenGL etc
[22:25] <grogbot> <criezy> I can see that logMessage(LogMessageType::Type type, const char *message) is reimplemented in OSystem_Switch, which means that the implementation in OSystem_SDL is not called.
[22:26] <grogbot> <criezy> And the implementation in OSystem_Switch only print to standard output (while the one in OSystem_SDL also print to the log file.
[22:26] <grogbot> <rsn8887> Probably to speed things up and allow nxlink if needed for debugging. Not sure.
[22:26] <grogbot> <rsn8887> file IO is tremendously slow on Switch
[22:26] <somaen> There, now the pre-menu credits show (well the upper half) for Discworld Noir.
[22:29] <grogbot> <criezy> In the iOS backend I am using a system notification to display the critical errors. That way I don't have to fish for the log file or to connect to a computer to attach a debugger. I found this quite useful and maybe this is an approach that could be used in other backends as well.
[22:29] <grogbot> <Henke37> i'd go as far as to say it's a defect to not write to the log file
[22:32] <grogbot> <criezy> If writing to a file is really slow, maybe something that you could do for the switch is only write the error messages to the log.
[22:36] <-- Shine_ left irc: Read error: Connection reset by peer
[22:44] <-- palasso left irc: Quit: I am not a quitter!
[22:50] <grogbot> <rsn8887> Hu it crashes when I touch the joystick
[22:50] <grogbot> <rsn8887> So something to do with keymapper I gues
[22:50] <grogbot> <rsn8887> Since joymouse uses it
[22:54] <grogbot> <Henke37> there's the pertubation.
[22:55] <grogbot> <rsn8887> Gotta check the pre-Henke commit again ð
[22:55] <grogbot> <rsn8887> To make sure it doesn't crash on joymouse
[23:02] <-- ny00123 left irc: Quit: Leaving
[23:05] <grogbot> <rsn8887> Works.
[23:05] <grogbot> <rsn8887> Last working commit on Switch: Last working: commit f1d753c1c4f061c43ab97b8591d7e292f5943a94 (HEAD -> testsw) Author: Thierry Crozat <criezy@scummvm.org> Date: Thu Feb 11 23:38:26 2021 +0000 AGS: Fix out of bound memory access with getpixel The allegro implementation returns -1 when called with a pixel that lies outside the bitmap. Doing the same in the implementation in ScummVM fixes a crash in some games.
[23:06] <grogbot> <rsn8887> Non-working commit (crashes on joystick motion): Non-working: commit f27bf85bf761b895723a12e35d98db9a63b6594d Author: Eugene Sandulenko <sev@scummvm.org> Date: Fri Feb 12 00:55:16 2021 +0100 COMMON: Fix for modern GCC
[23:06] <grogbot> <rsn8887> Inbetween are all the Henke commits.
[23:07] <grogbot> <rsn8887> and none of them inbetween commits compiles
[23:09] <grogbot> <criezy> Maybe this is related to https://bugs.scummvm.org/ticket/12161
[23:10] <grogbot> <criezy> This one also seems to be caused by Henke's commits (the Unknown key error is a good indication of that).
[23:10] <grogbot> <rsn8887> I suspect Henke PR successfully broke joymouse keymapper or all of keymapper
[23:10] <grogbot> <rsn8887> Probably was tested only with mouse keyboard? Not Gamepad?
[23:11] <grogbot> <criezy> It's a shame that the error in this case does not seem to indicate which 'key' is unknown.
[23:13] <grogbot> <rsn8887> Yup crashes on Windows 10 with Dualshock 4 too, as soon as I move left stick
[23:14] <grogbot> <rsn8887> So this has nothing to do with Switch. Just happens to be less common to use controlelr on Windows.
[23:14] <grogbot> <rsn8887> So I pass the ball back to @Henke37
[23:15] <grogbot> <criezy> At least this should be easier to debug if that also happens on desktop.
[23:16] <grogbot> <criezy> I don't have any controler though, so I will let others take a look.
[23:17] <grogbot> <Henke37> i don't have one either. at least none i use often. i'm more worried that i'm not even building with keymapper support
[23:20] <grogbot> <rsn8887> Could be this commit no clue: https://github.com/scummvm/scummvm/pull/2548/commits/130c3ac6b93bbc7cd081347cfd1f00e26686ba26
[23:25] <grogbot> <rsn8887> On Vita, Switch etc joymouse is basically the only thing used.
[23:25] <grogbot> <Henke37> it's a clue for sure. but the thing is, all the replacements are ones that do not call error
[23:28] <grogbot> <rsn8887> Here is my output on Win10: Using game controller: PS4 Controller Using game controller: PS4 Controller Unknown key!
[23:29] <grogbot> <rsn8887> as soon as I touch left stick, and back to commandline
[23:33] <grogbot> <rsn8887> Buttons seem to work, just joy axes crash
[23:35] <grogbot> <rsn8887> Any axes, when trying to map, including analog shoulders crash with unknown key. Digital buttons work
[23:35] <Scummette> [scummvm] somaen pushed 8 new commits to master: https://git.io/Jt1bA
[23:36] <grogbot> <criezy> Can you set a breakpoint in unknownKeyError() on line 310 of common/hashmap.h and get a backtrace when it gets there?
[23:36] <grogbot> <SupSuper> should an unknown key really be a fatal error?
[23:37] <grogbot> <Henke37> there is a variant for when you want a default value instead
[23:38] <grogbot> <rsn8887> Me? No. I cant compile Windows and Switch doesnt have debugging like that. If you are a ScummVM dev maybe get a $10 usb controller so you can test your mammoth refactoring Pr ðĪŠ
[23:38] <grogbot> <rsn8887> I think one of us probably has Windows debugging ANd controller capability ð
[23:39] <grogbot> <criezy> That was not my PR ð
[23:39] <grogbot> <criezy> And I am indeed hoping that some devs already have a controller.
[23:39] <grogbot> <rsn8887> Joy axes is not an unknown key on Windows. Maybe it was even before, but error was ignored?
[23:40] <grogbot> <rsn8887> I remember analog axes were handled kind of differently than other keys I thinkZ
[23:40] <grogbot> <criezy> 'key' here is not related to a keyboard key (or other hardware). This is a key in a hashmap, which could be anything.
[23:41] <grogbot> <Henke37> could it be keymap.cpp:220?
[23:41] <grogbot> <SupSuper> i can dig out my controller later tonight
[23:42] <grogbot> <SupSuper> i don't use it for anything but debugging ð
[23:42] <grogbot> <Henke37> looks rather suspicious to have different behavior in the if body vs the else body
[23:44] <grogbot> <criezy> That might be a good guess.
[23:45] <grogbot> <criezy> Everywhere else it is using getValOrDefault() to get a value from the _hwActionMap, so it should probably do that here as well and not use the operator[]()
[23:47] <grogbot> <criezy> @rsn8887 If you want to test on the switch you could change line 220 in backends/keymapper/keymap.cpp to use _hwActionMap.getValOrDefault(hardwareInput) instead of _hwActionMap[hardwareInput]. That might fix your crash.
[23:48] <grogbot> <Henke37> it is also consistent with the fact that the error fails to output the offending key.
[23:50] <grogbot> <criezy> Indeed. Since a HardwareInput is not something it knows how to print.
[23:50] <grogbot> <rsn8887> I can test immediately one sec
[23:51] <grogbot> <SupSuper> wait how do you actually use a controller on scummvm
[23:51] <grogbot> <SupSuper> do i just plug it in and watch the magic happen?
[23:52] <grogbot> <criezy> I think you used to have to set something in the config file, but that might no longer be the case.
[23:53] <grogbot> <Henke37> hmm, there's no nice way to get a string for the Key template argument, is there? the strigification preprocessor command would run too early.
[23:54] <grogbot> <criezy> I think it was joystick_num. But I see there is a default registered for it now.
[23:54] <grogbot> <SupSuper> yup got it to crash
[23:54] <grogbot> <SupSuper> any unbound control crashes it
[23:55] <grogbot> <SupSuper> an assert would be nicer, then the debugger would stop there
[23:56] <grogbot> <SupSuper> anyways your guess is correct, it's this line: actions.push_back(_hwActionMap[hardwareInput]);
[23:57] <grogbot> <Henke37> perhaps the error function should also do the magic incantation to make the debugger trip?
[23:58] <grogbot> <SupSuper> we'd need a portable version
[23:58] <grogbot> <Henke37> or let each backend provide the magic
[23:59] <grogbot> <SupSuper> it's compiler-dependant, not backend-dependant
[00:00] --- Wed Feb 17 2021