[Back to Index]

[00:41] <-- Farmboy0 left irc: Remote host closed the connection
[00:49] <rsn8887> snover: the Vita Dockerfile will have to be adjusted because the package sdl_net on dL.vitasdk.org was recently renamed to sdl2_net as part of my commits to update sdl2 and add sdl 1.2 packages. So now sdl_* is all sdl 1.2 and sdl2_* is sdl 2
[01:04] <-- waltervn left irc: Quit: Leaving
[01:41] --> kurtwr joined #scummvm.
[01:42] <-- kurtwr2 left irc: Ping timeout: 260 seconds
[01:43] --> Strangerke_ joined #scummvm.
[01:45] <-- Strangerke left irc: Ping timeout: 260 seconds
[01:45] Nick change: Strangerke_ -> Strangerke
[02:17] --> DominusExult joined #scummvm.
[02:17] <-- Dominus left irc: Ping timeout: 260 seconds
[02:18] Nick change: DominusExult -> Dominus
[02:27] <-- Joefish left irc: Ping timeout: 255 seconds
[02:27] --> Joefish joined #scummvm.
[02:27] #scummvm: mode change '+v Joefish' by ChanServ!ChanServ@services.
[02:30] <-- Dark-Star left irc: Ping timeout: 255 seconds
[02:37] --> Dark-Star joined #scummvm.
[02:54] <-- TMM left irc: Remote host closed the connection
[03:23] <-- tsoliman left irc: Quit: I've been banished!
[03:24] --> tsoliman joined #scummvm.
[03:24] #scummvm: mode change '+o tsoliman' by ChanServ!ChanServ@services.
[03:38] <tsoliman> snover: great job on that maemo worker transplant :)
[03:46] <-- dreammaster left irc:
[04:43] <-- snover left irc: Ping timeout: 240 seconds
[04:49] --> snover joined #scummvm.
[04:49] #scummvm: mode change '+o snover' by ChanServ!ChanServ@services.
[05:07] <rsn8887> snover: I made a PR to new buildbot https://github.com/csnover/scummvm-docker/pull/24 to fix Vita build in the following way: a) dl.vitasdk.org sdl_net was renamed to sdl2_net recently b) added the flac --without-pic to configure command for libflac to fix relocation error
[05:14] <rsn8887> hmm why is the source code on the webpage offered in four different archived versions? .tar.xz .tar.bz2 .tar.gz .zip I would think one is enough
[05:19] --> Begasus joined #scummvm.
[05:27] --> Begas_VBox joined #scummvm.
[06:10] <-- eriktorbjorn left irc: Quit: Leaving.
[06:39] <logix> rsn8887: different platforms - windows users (at least some) will have problems with the .tar versions, so you need a .zip, conversely not all unix systems come with unzip tools per default, but every one can unpack .tar.gz (and then .tar.bz2. and .tar.xz for newer systems because those compression schemes are more efficient)
[06:48] --> eriktorbjorn joined #scummvm.
[06:48] #scummvm: mode change '+o eriktorbjorn' by ChanServ!ChanServ@services.
[07:05] <-- eriktorbjorn left irc: Quit: Leaving.
[07:05] --> eriktorbjorn joined #scummvm.
[07:05] #scummvm: mode change '+o eriktorbjorn' by ChanServ!ChanServ@services.
[07:47] --> Henke37 joined #scummvm.
[07:58] <-- Mia left irc: Read error: Connection reset by peer
[08:00] --> Mia joined #scummvm.
[08:47] <-- LittleToonCat left irc: Remote host closed the connection
[08:48] <-- Begas_VBox left irc: Quit: Vision[0.10.2]: i've been blurred!
[08:49] --> ajax16384 joined #scummvm.
[08:49] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[09:03] <-- Begasus left irc: Ping timeout: 255 seconds
[09:12] --> Strangerke_ joined #scummvm.
[09:14] <-- Strangerke left irc: Ping timeout: 248 seconds
[09:16] --> Begasus joined #scummvm.
[09:17] <-- Strangerke_ left irc: Ping timeout: 276 seconds
[10:17] <-- ajax16384 left irc: Read error: Connection reset by peer
[10:17] --> Farmboy0 joined #scummvm.
[10:17] <-- Farmboy0 left irc: Changing host
[10:17] --> Farmboy0 joined #scummvm.
[10:49] --> Strangerke joined #scummvm.
[11:13] --> girafe2 joined #scummvm.
[11:16] <-- girafe left irc: Ping timeout: 248 seconds
[12:16] --> ajax16384 joined #scummvm.
[12:16] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[12:49] --> mortalius joined #scummvm.
[12:49] <-- mortalius__ left irc: Ping timeout: 260 seconds
[12:59] <-- Begasus left irc: Ping timeout: 250 seconds
[13:13] --> Begasus joined #scummvm.
[13:20] --> Strangerke_ joined #scummvm.
[13:22] <-- Strangerke left irc: Ping timeout: 240 seconds
[13:22] Nick change: Strangerke_ -> Strangerke
[14:23] --> ccawley2011 joined #scummvm.
[14:38] --> Mataniko joined #scummvm.
[14:39] <Mataniko> good morning
[14:39] <Mataniko> how is the windows release binary built?
[14:39] <Mataniko> the actual numbered release
[14:52] <wjp> snover is using the mingw-w64 toolchain in the docker container he built for the new buildbot to create binaries and an installer
[14:53] <wjp> why do you ask?
[14:53] <Mataniko> i'm looking into https://bugs.scummvm.org/ticket/9804
[14:54] <Mataniko> seems like something happened between 1.8.0 and 1.9.0
[14:54] <Mataniko> but it doesn't happen in the daily or my own builds
[14:55] <Mataniko> even if i build 1.9.0
[15:01] <criezy> This is probably related to http://forums.scummvm.org/viewtopic.php?p=84434 as well
[15:06] <Mataniko> yeah nothing comes out
[15:06] <Mataniko> on any command
[15:06] <Mataniko> the weird part is, redirecting to a file outputs all the text
[15:08] <-- Asterisk left irc: Ping timeout: 255 seconds
[15:09] <Mataniko> i'm using VS, i'll try to repro with a mingw env
[15:13] --> Asterisk joined #scummvm.
[15:13] <criezy> This seems relevant as well: https://github.com/docker/for-win/issues/262
[15:14] --> waltervn joined #scummvm.
[15:14] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[15:14] --> dreammaster joined #scummvm.
[15:14] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[15:19] <dreammaster> Anyone.. is it okay to cherry pick new bugfixes now to the branch-2-0-0 (in expectation of a possible future 2.0.1 release), or should I wait until all the 2.0 releases are built?
[15:21] <Mataniko> criezy: that looks exactly like the same problem
[15:21] <Mataniko> i don't see anything in the scummvm source to cause this, since we're just calling printf
[15:22] <Mataniko> i'm guessing the daily builds are done differently
[15:26] <criezy> dreammaster: you can cherry-pick to branch-2-0 (that will be used for a hypothetic 2.0.1 release), not branch-2-0-0.
[15:30] <Mataniko> snover: I think the solution might be to turn off tty (-t) in docker
[15:31] <dreammaster> Okay dokie. Thanks
[15:37] --> GitHub46 joined #scummvm.
[15:37] <GitHub46> [scummvm] dreammaster pushed 1 new commit to branch-2-0: https://git.io/vbRmx
[15:37] <GitHub46> scummvm/branch-2-0 c9049f6 Paul Gilbert: TITANIC: DE: Fix crash after tapping on bomb window 6 times...
[15:37] GitHub46 (GitHub46@192.30.252.38) left #scummvm.
[16:12] <wjp> for context: docker is used as a container for cross-compilation. Docker and any tty's at build-time should have zero impact on the run-time behaviour
[16:16] <criezy> Also 1.9.0 was not build with docker and exhibits the same issue.
[16:17] <criezy> The following suggests this might be related to the use of -mwindows when compiling: https://stackoverflow.com/questions/36171174/c-printf-on-win32-with-mingw
[16:21] <wjp> yeah, it's all a bit of a mess
[16:21] <wjp> also related to automatically opening a console window on launch or not
[16:41] <Mataniko> so what happened between 1.8.0 and 1.9.0?
[16:45] <-- ced117 left irc: Quit: leaving
[16:46] <criezy> a different person built the release: Kirben built 1.8.0 I think and was building previous releases, TMM built 1.9.0 and snover 2.0.0
[16:46] <criezy> and this could lead to a lot of other things being different (maybe different compiler version and different compilation flags)
[16:51] <-- dreammaster left irc:
[16:55] <wjp> also related is our "console" configuration option
[17:01] <snover> i wish i could say that console issue were magically fixed in 2.0 and it is not.
[17:02] --> ny00123 joined #scummvm.
[17:03] <wjp> it seems quite tricky to get sane behaviour from all of cmd (console output and piped output separately), msys, explorer
[17:04] <wjp> the internet seems to be full of tricks of solving some of them, but it doesn't appear like things have converged to a comprehensive solution
[17:04] --> ced117 joined #scummvm.
[17:05] --> girafe3 joined #scummvm.
[17:05] <snover> what are the limitations of AttachConsole(ATTACH_PARENT_PROCESS)?
[17:08] <wjp> according to random web pages, output redirection from cmd.exe doesn't work with that
[17:08] <wjp> haven't tested it myself
[17:08] <-- girafe2 left irc: Ping timeout: 240 seconds
[17:13] <Mataniko> it's really wonky, works in cygwin but not cmd, outputs to file, etc
[17:13] <Mataniko> First I need to know exactly how it's built to repro
[17:17] <snover> its exactly built using the scummvm/buildbot-windows image on docker hub, using configure --host=i686-w64-mingw32
[17:18] <wjp> and --enable-release presumably
[17:19] <Mataniko> i'll give that a go
[17:19] <snover> yes, though it doesnt work on the current buildbot either and that doesnt build in release mode
[17:19] <snover> Mataniko: https://github.com/csnover/scummvm-docker/#using-workers-as-stand-alone-compilers
[17:30] <wjp> snover: as a quick aside, things indeed don't seem to explode when I add some extra space after a SCI script
[17:31] <snover> 👍 !
[17:34] <wjp> I'm working on hooking it up to the script patcher, but this'll probably end up being next weekend rather than this one
[18:02] <rsn8887> The Docker build cache is just too aggressive for my taste: https://stackoverflow.com/questions/38655630/how-does-docker-know-when-to-use-the-cache-during-a-build-and-when-not#38656553
[18:02] <snover> rsn8887: i am sure if you can come up with a better general-purpose solution that they would be happy to hear from you.
[18:02] <rsn8887> Luckily --no-cache=true always works, but it is slow.
[18:03] <rsn8887> snover: Yes it seems impossible to change that, since it is just using a text file of commands to make some kind of inference about whether the resulting binary container file will be same or not :)
[18:04] <rsn8887> container -> image
[18:04] <rsn8887> I still keep confusing the vocabulare.
[18:04] <rsn8887> vocabulary.
[18:05] <rsn8887> But nevertheless the new buildbot is awesome. I can built *ANY* binary for *ANY* system using the same commands without any hassle.
[18:08] --> LittleToonCat joined #scummvm.
[18:12] <-- ajax16384 left irc: Read error: Connection reset by peer
[18:15] <rsn8887> although building psp release gives an error.... ah well.
[18:18] <snover> the main repository (and hence the 2.0 tarballs) are not in a state where their build code will work with the new buildbot in most cases because the old build code is crufty and doesnt function against a generic environment which specifies the locations of all used binaries using environment variables.
[18:25] <-- girafe3 left irc: Read error: Connection reset by peer
[18:28] --> dreammaster joined #scummvm.
[18:28] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[18:31] --> Strangerke_ joined #scummvm.
[18:31] <-- Strangerke left irc: Ping timeout: 240 seconds
[18:31] Nick change: Strangerke_ -> Strangerke
[18:39] #scummvm: mode change '+o Strangerke' by ChanServ!ChanServ@services.
[18:49] <-- Strangerke left irc: Ping timeout: 240 seconds
[19:01] --> Strangerke joined #scummvm.
[19:19] --> girafe joined #scummvm.
[19:29] <-- Strangerke left irc: Ping timeout: 258 seconds
[19:30] --> Strangerke joined #scummvm.
[19:38] <-- borosky left irc: Ping timeout: 248 seconds
[19:39] --> borosky joined #scummvm.
[19:43] <-- ny00123 left irc: Remote host closed the connection
[19:44] --> ny00123 joined #scummvm.
[19:50] <-- ccawley2011 left irc: Ping timeout: 260 seconds
[19:52] --> ccawley2011 joined #scummvm.
[19:55] <-- Strangerke left irc: Ping timeout: 248 seconds
[20:12] --> Strangerke joined #scummvm.
[20:20] --> Strangerke_ joined #scummvm.
[20:22] <-- Strangerke left irc: Ping timeout: 240 seconds
[20:22] Nick change: Strangerke_ -> Strangerke
[20:23] <snover> dockers product ecosystem is confusing. docker hub is a thing, and it has optional automatic builds, and then theres docker cloud which is the exact same thing except has the word cloud in it, which runs atop hub. cloud has some option marked beta for using swarm mode which hub doesnt, plus one extra feature for specifying the working directory for automatic builds which hub does not.
[20:23] <snover> docker swarm is a thing, intended for deploying clusters of containers across multiple hosts in a coordinated manner, and it used to be standalone and then it was integrated, and there are still some instructions for legacy standalone mode, and now in some places the *whole* docker swarm is legacy and docker stacks replaces it, except docker stacks is marked as experimental.
[20:25] --> Strangerke_ joined #scummvm.
[20:27] <-- Strangerke left irc: Ping timeout: 268 seconds
[20:27] Nick change: Strangerke_ -> Strangerke
[20:27] <wjp> I will pretend to have read and understand all that and just nod, if that's ok? :-)
[20:28] <snover> im not much better in understanding what is going on with all of this. it is pretty baffling.
[20:28] <snover> its like they dont actually have anyone in charge of product who is making sure that they are not duplicating the same things over and over again.
[20:29] <snover> im just trying to figure out if there is a way i can join vm and vm2 together in a single network so i can avoid having to do extra work to set up a separate network overlay or tunneling for the buildbot command channel.
[20:31] <wjp> we can do an own private vpn between the two if we need to
[20:35] <snover> indeed. its just that it seems that there is supposed to be a way to do this with docker itself, which allows the isolated networks of the containers to remain isolated, instead of needing to expose them to the host.
[20:35] <snover> so i would like to do that if i can
[20:36] <snover> generally speaking once i figure out how to do something its easy to do, its just the process of wading through their documentation that can be problematic.
[20:39] <rsn8887> snover: I like that docker allows to do it. However, however security using Docker seems somewhat limited since, on each connected computer running docker container, the docker host application has to be running with full root access (wtf?)
[20:40] <snover> docker host application? you mean the daemon?
[20:40] <rsn8887> yes
[20:40] <-- Strangerke left irc: Ping timeout: 260 seconds
[20:40] <rsn8887> I was baffled that it needed my root password, maybe that's ok I am a bit concerned about it
[20:41] <rsn8887> There's probably a way of avoiding that.
[20:41] <snover> anything that can perform privileged kernel operations needs superuser access.
[20:42] <rsn8887> Hmm you are right I just checked and Virtualbox also needs such access. I guess it is fine then.
[20:47] <snover> running an application in a container is only marginally less secure than running an application in a full virtual machine, both of which are way safer than running an application on the host itself where any privilege escalation means the entire host is owned.
[20:47] <rsn8887> I see.
[20:49] <wjp> but you do have to trust the container daemon and/or hypervisor
[20:49] --> Strangerke joined #scummvm.
[20:49] <rsn8887> that's why I have only installed the stable branch daemon, although it has some hilarious bug on osx that causes it to eat all diskspace until hd is full.
[20:50] <snover> no, it just collects slack space on a timer instead of immediately when the data is released.
[20:51] <snover> so clear away some images/containers and wait several minutes and the disk space will be returned
[20:51] <snover> it should probably be running the collection immediately after purge, and it is not so bad as needing to constantly wipe everything as it was in the past.
[20:53] <rsn8887> I think what I am struggling with is this bug ".qcow2 file grows forever", not fixed, which I think is different from what you mean: https://github.com/docker/for-mac/issues/371#issuecomment-350555006
[20:53] <snover> it doesnt grow forever, once the cleanup task triggers it shrinks.
[20:53] <rsn8887> I waited a whole day, all images cleared, system purged, and still diskspace was gone.
[20:53] <rsn8887> Hmm let me try again.
[20:55] <snover> system purge wont get rid of any volumes since they are intended to be persistent, so if you have been putting a lot of data into volumes then you will still have disk space used for those volumes.
[21:00] <rsn8887> Docker system df shows only 996 MB used by Image (Cache?) on 0B in volumes, but ls -al ~/Library/Containers/com.docker.docker/Data/com.docker.driver.amd64-linux/Docker.qcow2 shows the qcow2 file is 5.9 GB (!)
[21:00] <-- Strangerke left irc: Ping timeout: 248 seconds
[21:01] <rsn8887> hmmm, "docker system prune" reduced it to 3.4 GB so it does do something like you said, maybe if I just run docker system prune every once in a while and clear images, it will be fine.
[21:05] <rsn8887> Looks like the .qcow2 is some kind of vm disk needed to hack docker to "appear" to run on Darwin when it is in reality running in a vm. The official recommendation from Docker, is to remove all images if possible, and just delete the .qcow2 file, until this problem is fixed.
[21:06] <rsn8887> Apparently after deleting .qcow2, Docker generates a new, small one.
[21:18] <rsn8887> Hmm I deleted it and everything works fine. I guess it only gets created when needed, maybe some kind of overflow drive for data that doesn't fit into ram.
[21:19] <rsn8887> At least I got my diskspace back, some via docker system prune, some by just deleting the .qcow2 file...
[21:19] <snover> the OS wont release a file until there are no more open file handles, even if you delete it
[21:21] <snover> if you feel really compelled to wipe the data for some reason there is a button to do that in the docker preferences.
[21:22] <rsn8887> Ah yes, that does try to delete the .qcow2 file. It just complained that my .qcow2 file was not found and that I should reset to factory defaults. fine.
[21:22] --> nh_cham joined #scummvm.
[21:22] <-- nh_cham left irc: Quit: AtomicIRC: The nuclear option.
[21:23] <rsn8887> After restarting it created a new one with size 1.2 GB so quite a bit smaller.
[21:23] --> nh_cham joined #scummvm.
[21:24] <rsn8887> snover: Is there any point in me working on buildbot-psp? Currently it aborts compilation with some errors at png_loader.c
[21:25] <rsn8887> The old buildbot built an image just fine at the end of november
[21:29] --> Strangerke joined #scummvm.
[21:41] <Mataniko> So i'm in builtbot-windows and compiled using make, but it's not a windows binary
[21:41] <Mataniko> what am i missing
[21:51] <wjp> how did you run configure?
[21:57] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[22:03] <Mataniko> just ./configure
[22:04] <-- girafe left irc: Read error: Connection reset by peer
[22:04] <-- waltervn left irc: Quit: Leaving
[22:05] <Mataniko> ah, --host
[22:06] <Mataniko> er, that's only for other platforms, not win32
[22:08] <wjp> --host=i686-w64-mingw32
[22:12] <-- ny00123 left irc: Quit: Leaving
[22:15] #scummvm: mode change '+o Strangerke' by ChanServ!ChanServ@services.
[22:18] <Mataniko> it's trying to link x11 now to build
[22:18] <Mataniko> libx11-dev is instlaled
[22:19] <snover> Mataniko: oops, i always forget this about that builder. you must PATH=/usr/i686-w64-mingw32/bin:$PATH configure --host=i686-w64-mingw32
[22:20] <snover> or else explicitly give paths to sdl2-config, freetype-config, and &whatever the third one is
[22:20] <snover> using --with flags
[22:20] <snover> (the easy way is to just PATH when configuring)
[22:22] <snover> rsn8887: as i said already before, its not possible to build all the ports from the release tarball/master code because it is outdated and broken
[22:23] <snover> psp builds fine, from my buildbot branch, which is not ready
[22:25] <-- Farmboy0 left irc: Remote host closed the connection
[22:26] <Mataniko> same error snover
[22:28] <snover> what does configure say about where it is finding sdl2-config?
[22:31] <Mataniko> /usr/bin/sdl2-config
[22:32] <snover> you are running exactly `PATH=/usr/i686-w64-mingw32/bin:$PATH configure --host=i686-w64-mingw32`?
[22:32] <snover> and not just `configure --host=i686-w64-mingw32`?
[22:32] <Mataniko> i got the PATH order mixed, so it loaded /usr/bin first
[22:34] <Mataniko> got a different error now, backends/platform/sdl/win32/win32.cpp: In member function 'virtual bool OSystem_Win32::openUrl(const Common::String&)':
[22:34] <Mataniko> backends/platform/sdl/win32/win32.cpp:142:52: warning: format '%u' expects argument of type 'unsigned int', but argument 2 has type 'uint64 {aka long long unsigned int}' [-Wformat=]
[22:34] <snover> thats not an error.
[22:34] <snover> if it was an error it would have the word error in it. :)
[22:34] <Mataniko> ./common/forbidden.h:57:89: error: expected initializer before 'SYMBOL'
[22:34] --> GitHub45 joined #scummvm.
[22:34] <GitHub45> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vbRw2
[22:34] <GitHub45> scummvm/master bd904b0 Paul Gilbert: XEEN: Updating map dialog outdoors rendering to use constants
[22:34] GitHub45 (GitHub45@192.30.252.39) left #scummvm.
[22:37] <wjp> you may want to put a full build log from configure all the way up to the error on pastebin or such
[22:46] <Mataniko> https://gist.github.com/anonymous/0bc83f1e4c89ff92d80052bf494a0713
[22:49] <wjp> and which code exactly are you building?
[22:49] --> DJWillis joined #scummvm.
[22:49] #scummvm: mode change '+o DJWillis' by ChanServ!ChanServ@services.
[22:51] <-- Lightkey left irc: Ping timeout: 246 seconds
[22:59] <Mataniko> v1.9.0 tag
[22:59] <wjp> scummvm.exe builds fine for me here using that docker image
[22:59] <wjp> oh, you may have to fiddle with things for that
[22:59] <wjp> v2.0.0 works here
[23:03] --> Lightkey joined #scummvm.
[23:05] <Mataniko> success
[23:05] <Mataniko> and repro of the output issue
[23:15] --> Strangerke_ joined #scummvm.
[23:17] <-- Strangerke left irc: Ping timeout: 240 seconds
[23:17] Nick change: Strangerke_ -> Strangerke
[23:19] <-- Begasus left irc: Quit: Ex-Chat
[23:40] <-- criezy left irc: Quit: criezy
[00:00] --- Mon Dec 11 2017