ScummVM IRC Log - 11 Apr 2019

[Back to Index]

[00:10] <-- SylvainTV left irc: Read error: Connection reset by peer
[00:31] <-- criezy left irc: Quit: criezy
[00:32] <-- thesquib left irc: Ping timeout: 245 seconds
[00:34] <-- timofonic left irc: Ping timeout: 255 seconds
[00:34] --> thesquib joined #scummvm.
[00:34] --> timofonic joined #scummvm.
[01:38] --> DominusExult joined #scummvm.
[01:38] <-- DominusExult left irc: Changing host
[01:38] --> DominusExult joined #scummvm.
[01:42] <-- Dominus left irc: Ping timeout: 245 seconds
[01:42] Nick change: DominusExult -> Dominus
[02:15] --> dreammaster joined #scummvm.
[02:15] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[03:27] <-- dreammaster left irc:
[04:59] <-- timofonic left irc: Ping timeout: 245 seconds
[05:00] --> timofonic joined #scummvm.
[05:10] <-- BeefEats left irc: Ping timeout: 255 seconds
[05:11] --> BeefEats joined #scummvm.
[05:18] <-- timofonic left irc: Ping timeout: 255 seconds
[05:18] --> timofonic joined #scummvm.
[05:20] <-- Lightkey left irc: Ping timeout: 240 seconds
[05:22] <-- eriktorbjorn left irc: Quit: Leaving.
[05:33] --> Lightkey joined #scummvm.
[05:49] --> stroggoff joined #scummvm.
[05:54] <-- timofonic left irc: Ping timeout: 252 seconds
[07:24] <-- SupSuper left irc: Read error: Connection reset by peer
[07:38] --> whiterandrek joined #scummvm.
[07:38] #scummvm: mode change '+o whiterandrek' by ChanServ!ChanServ@services.
[08:15] <-- whiterandrek left irc: Quit: Leaving
[08:23] <-- JohnnyonFlame left irc: Ping timeout: 246 seconds
[08:30] --> jamm joined #scummvm.
[08:41] <stroggoff> Hi guys!
[08:42] <stroggoff> The Debian-x86 build has been failing some tests for a few days now
[08:42] <stroggoff> something about MathTestSuite::test_rad2deg and common/math.h (?)
[08:43] <stroggoff> * test/common/math.h
[08:48] --> JohnnyonFlame joined #scummvm.
[08:50] <stroggoff> It looks like it could be related to the M_PI define for that toolchain?
[09:03] <-- JohnnyonFlame left irc: Read error: Connection reset by peer
[09:07] --> mwillcock joined #scummvm.
[09:12] --> ldevulder_ joined #scummvm.
[09:15] <-- ldevulder left irc: Ping timeout: 268 seconds
[09:20] <-- jamm left irc: Ping timeout: 245 seconds
[09:31] --> TAS-2012v joined #scummvm.
[09:31] <-- TAS_2012v left irc: Ping timeout: 255 seconds
[09:31] Nick change: TAS-2012v -> TAS_2012v
[09:35] --> waltervn joined #scummvm.
[09:35] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[09:49] --> jamm joined #scummvm.
[10:05] --> rootfather joined #scummvm.
[10:05] <-- rootfather left irc: Changing host
[10:05] --> rootfather joined #scummvm.
[10:05] #scummvm: mode change '+o rootfather' by ChanServ!ChanServ@services.
[10:11] <rootfather> hi peeps
[10:33] --> TAS-2012v joined #scummvm.
[10:34] <-- TAS_2012v left irc: Ping timeout: 255 seconds
[10:34] Nick change: TAS-2012v -> TAS_2012v
[11:27] --> Littleboy joined #scummvm.
[11:27] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[11:32] --> Stormkeeper joined #scummvm.
[11:39] <Scummette> [scummvm] dafioram pushed 1 new commits to master: https://git.io/fjqV9
[11:39] <Scummette> scummvm/master edb93c4 dafioram: TESTS: Use Less than or equal to for math tests
[11:54] --> yuv422 joined #scummvm.
[12:11] ist5shreawf[m] <-- (ist5shreaw@gateway/shell/matrix.org/x-ezbcuqatdtdfwhxa) left irc: Ping timeout: 250 seconds
[12:11] ad5twoknebor[m] <-- (ad5twokneb@gateway/shell/matrix.org/x-ctqreimjxcyulnlp) left irc: Ping timeout: 250 seconds
[12:11] olmvnec[m] <-- (olmvnected@gateway/shell/matrix.org/x-xoidyuowxnqkadcp) left irc: Ping timeout: 252 seconds
[12:11] divi[m] <-- (divivipsem@gateway/shell/matrix.org/x-obhzqcwzqgraakmq) left irc: Ping timeout: 252 seconds
[12:11] Xorbb[m] <-- (pamatrixst@gateway/shell/matrix.org/x-bawxawqrzrzltrwx) left irc: Ping timeout: 250 seconds
[12:11] <-- Ironically_Serio left irc: Ping timeout: 250 seconds
[12:11] hylshols7qui[m] <-- (hylshols7q@gateway/shell/matrix.org/x-cwjdbxzsbydonzox) left irc: Ping timeout: 265 seconds
[12:12] alx604[m] <-- (alx604matr@gateway/shell/matrix.org/x-bzdzgevttqnoqumx) left irc: Ping timeout: 264 seconds
[12:18] <-- yuv422 left irc: Remote host closed the connection
[12:34] <-- jamm left irc: Ping timeout: 245 seconds
[13:06] <-- stroggoff left irc: Read error: Connection reset by peer
[13:06] --> stroggoff joined #scummvm.
[13:14] <-- TAS_2012v left irc: Ping timeout: 246 seconds
[13:15] --> TAS_2012v joined #scummvm.
[13:18] --> wazimuth joined #scummvm.
[13:18] --> jamm joined #scummvm.
[14:50] --> NULUSIOS joined #scummvm.
[15:22] <-- wazimuth left irc: Ping timeout: 256 seconds
[15:39] <-- ST left irc: Read error: Connection reset by peer
[15:40] --> ajax16384 joined #scummvm.
[15:40] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[15:53] <-- stroggoff left irc: Quit: Looking for a flock of swans
[15:58] --> ny00123 joined #scummvm.
[16:12] <-- Tkachov left irc: Quit: q
[16:12] --> Tkachov joined #scummvm.
[16:12] #scummvm: mode change '+o Tkachov' by ChanServ!ChanServ@services.
[16:46] --> SupSuper joined #scummvm.
[17:21] <-- cd left irc: Quit: cd
[17:25] --> cd joined #scummvm.
[17:30] --> _athrxx joined #scummvm.
[17:31] <-- mwillcock left irc:
[17:43] --> SylvainTV joined #scummvm.
[17:43] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[17:50] --> eriktorbjorn joined #scummvm.
[17:50] #scummvm: mode change '+o eriktorbjorn' by ChanServ!ChanServ@services.
[18:05] <-- girafe left irc: Read error: Connection reset by peer
[18:10] <-- mlk50 left irc: Quit: The Lounge - https://thelounge.chat
[18:11] --> mlk joined #scummvm.
[18:13] <-- jamm left irc: Ping timeout: 252 seconds
[18:16] --> wazimuth joined #scummvm.
[18:23] <-- madmoose left irc: Ping timeout: 244 seconds
[18:48] --> girafe joined #scummvm.
[18:52] --> lavosspawn joined #scummvm.
[18:55] <-- BeefEats left irc: Ping timeout: 255 seconds
[18:56] --> BeefEats joined #scummvm.
[19:04] --> JohnnyonFlame joined #scummvm.
[19:20] --> DJWillis joined #scummvm.
[19:20] #scummvm: mode change '+o DJWillis' by ChanServ!ChanServ@services.
[19:26] <-- lavosspawn left irc: Ping timeout: 250 seconds
[19:27] <peterkohaut> please, on android, how can i access scummvm menu when i have no menu button?
[19:29] <rootfather> if memory serves me right, a two- or three-finger-tap will launch the menu
[19:30] --> stroggoff joined #scummvm.
[19:31] <peterkohaut> ah, that worked, thanks!
[19:32] <rootfather> which was it? two or three fingers? ;)
[19:33] <peterkohaut> that i'm not sure of, i just smash my phone few time till the menu came up.... :)
[19:35] <peterkohaut> but i only have adb console as my pixel format is not supported and everything is black
[19:36] <peterkohaut> wait, i was lying, no idea what i did but now it doesnt work
[19:38] <Scummette> [scummvm] bgK opened pull request #1583: COMMON: Use a prefix table to speed up the Huffman decoder (master...huffman-prefix-table) https://git.io/fjqDh
[19:39] --> athrxx joined #scummvm.
[19:40] <rootfather> hm
[19:41] <rootfather> maybe it was a swipe gesture?
[19:41] <-- _athrxx left irc: Ping timeout: 264 seconds
[19:44] <Lightkey> let the cat handle it
[19:47] <-- BeefEats left irc: Ping timeout: 252 seconds
[19:48] <peterkohaut> first i need to get some image output from the game :)
[19:49] --> BeefEats joined #scummvm.
[19:49] <peterkohaut> otherwise cat will get bored before touching it
[19:54] <-- Mia left irc: Read error: Connection reset by peer
[19:59] --> Mia joined #scummvm.
[19:59] <-- Mia left irc: Changing host
[19:59] --> Mia joined #scummvm.
[20:06] <stroggoff> For Android there's an option in launch menu to swap back button with menu button. Then you just press the back button from Android menu and the in-game ScummVM menu will appear
[20:08] <stroggoff> I think that option is enabled by default (it was for me, but my android scummvm build is from last August
[20:14] --> ccawley2011 joined #scummvm.
[20:14] #scummvm: mode change '+o ccawley2011' by ChanServ!ChanServ@services.
[20:17] --> criezy joined #scummvm.
[20:17] #scummvm: mode change '+o criezy' by ChanServ!ChanServ@services.
[20:20] <peterkohaut> i didnt fint that option, i tried accessiblity menu but the "menu" button is not there either
[20:21] <Scummette> [scummvm] ccawley2011 pushed 1 new commits to master: https://git.io/fjqy9
[20:21] <Scummette> scummvm/master 88bca6a ccawley2011: MUTATIONOFJB: Fix warning
[20:24] <peterkohaut> differrent question, any opinion on calling surface::convertTo for every frame? from 555 to 5551. or convert the engine to do it in background? or why is android pixel format support so limited?
[20:28] <peterkohaut> *to do it in background for every resource used
[21:04] <-- ajax16384 left irc: Quit: Leaving
[21:10] --> madmoose joined #scummvm.
[21:10] #scummvm: mode change '+o madmoose' by ChanServ!ChanServ@services.
[21:17] <madmoose> peterkohaut: When would you do the conversion if not on every frame?
[21:53] --> kensan joined #scummvm.
[21:55] <-- madmoose left irc: Ping timeout: 250 seconds
[22:20] --> madmoose joined #scummvm.
[22:22] <-- waltervn left irc: Quit: Leaving
[22:23] --> dafioram joined #scummvm.
[22:25] <-- madmoose left irc: Ping timeout: 264 seconds
[22:25] <dafioram> is it obvious to any1 why the debian is failing? Here is the log: https://buildbot.scummvm.org/builders/master-debian-x86/builds/1021/steps/test/logs/stdio
[22:26] <dafioram> the first build that it failed for was: https://buildbot.scummvm.org/builders/master-debian-x86/builds/997/steps/test/logs/stdio
[22:27] <dafioram> so it shouldn't be an M_PI issue
[22:30] <SupSuper> precision issues?
[22:32] --> ST joined #scummvm.
[22:32] #scummvm: mode change '+o ST' by ChanServ!ChanServ@services.
[22:32] <dafioram> the comparison precision is set to be a floating value of 0.000001
[22:35] --> timofonic joined #scummvm.
[22:35] <timofonic> Hello
[22:36] <timofonic> Escoria = Scumm (from Spanish to English). Is it a coincidence? https://github.com/godotengine/escoria
[22:42] <criezy> That would be scum and not scumm.
[22:42] <criezy> But yeah, I doubt it is a coincidence.
[22:42] <criezy> I actually saw a presentation of Escoria a few years back at FOSDEM, but I don't remember if that was mentionned.
[22:42] <SupSuper> ""Escoria" is a Spanish word that means "scum" (the creators where not very original when naming the framework :-)"
[22:42] <SupSuper> seems intentional
[22:43] <criezy> I will try to remember to ask the next time we meat the Godot developers :-P
[22:43] <Deledrius> I'm pretty sure I saw some mention on one of their issues that it's intentional

[22:54] --> madmoose joined #scummvm.
[22:59] <-- madmoose left irc: Ping timeout: 268 seconds
[23:01] <-- dafioram left irc: Quit: Leaving
[23:01] <Scummette> [scummvm] dafioram pushed 1 new commits to master: https://git.io/fjqHr
[23:01] <Scummette> scummvm/master 9b993ae dafioram: TESTS: reduce precision for math tests
[23:08] <-- wazimuth left irc: Ping timeout: 256 seconds
[23:15] <-- athrxx left irc: Quit: Leaving
[23:21] ced117 (~ced117@opensuse/member/ced117) got netsplit.
[23:24] <criezy> That error for the rad2deg test is strange.
[23:25] <criezy> I am wondering if it might be related to the template type it ends up using...
[23:26] --> dafioram joined #scummvm.
[23:27] <dafioram> Even though it still failed the number of tests that failed went from 3 to 1 so somehow the precision is the answer?
[23:27] ced117 (~ced117@opensuse/member/ced117) returned to #scummvm.
[23:28] <criezy> Yes, it seems so.
[23:29] <dafioram> but what is debian-86 doing differently?
[23:29] <criezy> Hmmm, if I run it on my side it uses the deg2rad with float and not double...
[23:30] <criezy> And since the error is absolute, ona value of 180 or 360 we are beyond the precision of float...
[23:30] <criezy> Which would explain why the test with 0, and the test on deg2rad work, as they are on smaller absolute value.
[23:31] <-- ny00123 left irc: Quit: Leaving
[23:32] <-- Mia left irc: Read error: Connection reset by peer
[23:34] <criezy> I don't know why, but I need to use Common::rad2deg<double>(M_PI) for it to use the double precision version of deg2rad.
[23:34] <criezy> If I don't specify the template type it uses the single precision version (but still doesn't fail for me).
[23:37] <dafioram> i think its good that we know the limitations of the conversion as float
[23:38] <dafioram> i still don't understand why the float is beyond precision. it does a conversion from 2pi to ~360 then subtracts 360 and compares that to 1e-5
[23:40] <-- ccawley2011 left irc: Quit: Page closed
[23:40] <criezy> You have about 6 or 7 digit precision, so on 360 your precision is at about 3 or 4 digit after the decimal point.
[23:41] <criezy> If you substract for example 360.0001 and 360.0004, your difference if 0.0003.
[23:41] <criezy> But it's late so maybe my logic is completely wrong.
[23:41] <criezy> My brain no longer works properly after midnight usually ;-)
[23:44] <dafioram> that makes sense. let me think how to revise test
[23:45] <criezy> You can maybe make sure to use input and output type of double with something like this:
[23:45] <criezy> TS_ASSERT_ALMOST_EQUALS((Common::rad2deg<double, double>)(M_PI), 180.0, MAX_ERROR);
[23:45] <SupSuper> shouldn't MAX_ERROR also be a double?
[23:46] <dafioram> it will just upconvert to double
[23:49] --> Littleboy joined #scummvm.
[23:49] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[23:51] <criezy> Hmmm, interestingly using deg2rad<double, double>() actually uses the float version...
[23:52] <criezy> So the only one that does use the double version is deg2rad<double>(value) when value is a double.
[23:53] <dafioram> hmm
[23:53] <Scummette> [scummvm] dafioram pushed 1 new commits to master: https://git.io/fjqQ3
[23:53] <Scummette> scummvm/master cf0e9bf dafioram: TESTS: further reduce precision for math tests
[23:54] <dafioram> the float versions probably sufficient
[23:55] --> madmoose joined #scummvm.
[23:55] <criezy> I think I am starting to understand a bit how it chooses. Anything rad2deg<type>(value) with value a double with use the double version and then cast the result to type.
[23:55] <criezy> Anything else uses the float version.
[23:56] <criezy> The issue with reducing the error like that, is that now on the deg2rad test we have a huge relative error since the values we test are much smaller.
[23:57] <criezy> Maybe we should use different errors depending on the absolute value of the value tested.
[23:57] <dafioram> alright
[23:57] <dafioram> shall i test the different templates of the function?
[23:58] <criezy> Or treat is as a relative error (like passing 180 * MAX_ERROR as the error for example)?
[23:58] <criezy> when testing rad2deg(PI)
[23:58] <criezy> and *360 when testing rad2deg(2 * PI)
[23:59] <criezy> You mean addingmore tests with different template types?
[23:59] <dafioram> yes
[23:59] <dafioram> so your saying MAX_ERROR = 1 + 1e-5 or something?
[23:59] <-- madmoose left irc: Ping timeout: 252 seconds
[00:00] --- Fri Apr 12 2019