ScummVM IRC Log - 27 Feb 2017

[Back to Index]

[00:01] <-- 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.
[00:02] --> GitHub72 joined #scummvm.
[00:02] <GitHub72> [scummvm] Strangerke pushed 1 new commit to master: https://git.io/vykpo
[00:02] <GitHub72> scummvm/master e4b7f69 Strangerke: CRYO: Fix crash when restarting the game
[00:02] GitHub72 (GitHub72@192.30.252.42) left #scummvm.
[00:08] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[00:14] <-- nutron left irc: Remote host closed the connection
[00:15] --> nutron joined #scummvm.
[00:21] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[00:41] --> RLP__ joined #scummvm.
[00:44] <-- D0SFreak left irc: Ping timeout: 240 seconds
[01:30] <-- RLP__ left irc: Quit: Leaving
[02:31] <-- exmensa_ left irc: Ping timeout: 240 seconds
[02:36] <-- TMM left irc: Ping timeout: 240 seconds
[02:37] <-- Dominus left irc: Ping timeout: 240 seconds
[02:39] --> Dominus joined #scummvm.
[02:39] <-- Dominus left irc: Changing host
[02:39] --> Dominus joined #scummvm.
[02:42] --> TMM joined #scummvm.
[02:42] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[02:42] --> Vampire0 joined #scummvm.
[02:45] <-- Vampire0_ left irc: Ping timeout: 260 seconds
[03:07] --> exmensa joined #scummvm.
[03:50] --> vinayakvivek joined #scummvm.
[03:58] <-- SylvainTV left irc: Read error: Connection reset by peer

[04:01] <-- heroux left irc: Read error: Connection reset by peer
[04:15] --> heroux joined #scummvm.
[04:19] <-- heroux left irc: K-Lined
[04:21] --> Axy joined #scummvm.
[04:21] <-- Axy left irc: Changing host
[04:21] --> Axy joined #scummvm.
[04:24] <-- Mia left irc: Ping timeout: 268 seconds
[04:47] <-- dreammaster left irc:
[05:32] --> te_lanus joined #scummvm.
[05:41] --> diehumblex joined #scummvm.
[06:18] --> waltervn joined #scummvm.
[06:18] #scummvm: mode change '+o waltervn' by ChanServ!ChanServ@services.
[06:18] <waltervn> morning
[06:52] <wjp> morning
[07:49] --> ajax16384 joined #scummvm.
[07:49] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[08:01] <-- Lightkey left irc: Ping timeout: 240 seconds
[08:14] --> Lightkey joined #scummvm.
[08:31] <-- te_lanus left irc: Read error: Connection reset by peer
[08:54] <-- TMM left irc: Quit: Ex-Chat
[08:55] <-- LittleToonCat left irc: Remote host closed the connection
[08:57] --> _sev|work joined #scummvm.
[08:57] <-- _sev|work left irc: Changing host
[08:57] --> _sev|work joined #scummvm.
[08:57] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[09:44] <-- vinayakvivek left irc: Quit: Connection closed for inactivity
[09:49] <-- wanwan left irc: Remote host closed the connection
[09:52] --> wanwan joined #scummvm.
[09:56] --> Henke37 joined #scummvm.
[10:09] --> TMM joined #scummvm.
[10:09] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[10:11] <-- TMM left irc: Client Quit
[10:11] --> Strangerke|work joined #scummvm.
[10:12] <Strangerke|work> hi guys
[10:13] --> TMM joined #scummvm.
[10:13] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[10:15] <Strangerke|work> Today Google will announce which projects are selected for the GSoC
[10:15] <Strangerke|work> Fingers crossed
[10:38] --> h00ligan joined #scummvm.
[10:38] #scummvm: mode change '+o h00ligan' by ChanServ!ChanServ@services.
[10:42] <-- ajax16384 left irc: Ping timeout: 260 seconds
[11:22] --> pegartillo joined #scummvm.
[11:49] <-- pegartillo left irc: Quit: http://www.kiwiirc.com/ - A hand crafted IRC client
[12:16] <-- Henke37 left irc: Quit: ERR_SHUTDOWN
[12:19] --> jamm joined #scummvm.
[12:48] --> Vampire0_ joined #scummvm.
[12:49] <-- Vampire0 left irc: Ping timeout: 252 seconds
[12:59] --> m_kiewitz joined #scummvm.
[12:59] <-- m_kiewitz left irc: Changing host
[12:59] --> m_kiewitz joined #scummvm.
[12:59] #scummvm: mode change '+o m_kiewitz' by ChanServ!ChanServ@services.
[13:47] --> dreammaster joined #scummvm.
[13:47] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[13:47] <-- marcus_c left irc: Ping timeout: 240 seconds
[13:56] --> GitHub150 joined #scummvm.
[13:56] <GitHub150> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vyIp0
[13:56] <GitHub150> scummvm/master 93cd3ec Paul Gilbert: TITANIC: Implemented CStarControl doAction
[13:56] GitHub150 (GitHub150@192.30.252.41) left #scummvm.
[13:56] --> GitHub177 joined #scummvm.
[13:56] <GitHub177> [scummvm] criezy pushed 1 new commit to master: https://git.io/vyIpg
[13:56] <GitHub177> scummvm/master 2dab510 Ben Castricum: I18N: Update translation (Dutch)...
[13:56] GitHub177 (GitHub177@192.30.252.42) left #scummvm.
[13:56] <-- _sev|work left irc: Quit: This computer has gone to sleep
[14:00] --> _sev|work joined #scummvm.
[14:00] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[14:08] --> GitHub90 joined #scummvm.
[14:08] <GitHub90> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vyIjJ
[14:08] <GitHub90> scummvm/master 1f99c01 Paul Gilbert: TITANIC: Added remainder of CStarControl class
[14:08] GitHub90 (GitHub90@192.30.252.40) left #scummvm.
[14:21] --> GitHub186 joined #scummvm.
[14:21] <GitHub186> [scummvm] dreammaster pushed 1 new commit to master: https://git.io/vyLeg
[14:21] <GitHub186> scummvm/master 0c64023 Paul Gilbert: TITANIC: Fix initialization of FRect
[14:21] GitHub186 (GitHub186@192.30.252.42) left #scummvm.
[14:21] <-- GeekShadow left irc: Ping timeout: 245 seconds
[14:23] --> marcus_c joined #scummvm.
[14:29] --> GeekShadow joined #scummvm.
[14:29] <-- GeekShadow left irc: Changing host
[14:29] --> GeekShadow joined #scummvm.
[14:29] <-- dreammaster left irc:
[15:03] --> GitHub38 joined #scummvm.
[15:03] <GitHub38> [scummvm] wjp pushed 1 new commit to master: https://git.io/vyLmT
[15:03] <GitHub38> scummvm/master 8bac883 Willem Jan Palenstijn: SCI: Fix up readString writing into too small buffer...
[15:03] GitHub38 (GitHub38@192.30.252.34) left #scummvm.
[15:03] --> GitHub22 joined #scummvm.
[15:03] <GitHub22> [scummvm] criezy pushed 1 new commit to master: https://git.io/vyLmI
[15:03] <GitHub22> scummvm/master 7e933e5 Thierry Crozat: I18N: Update translations templates
[15:03] GitHub22 (GitHub22@192.30.252.34) left #scummvm.
[15:08] <m_kiewitz> wjp: do you have a tool to dump resource.aud resources?
[15:08] <wjp> no
[15:08] <m_kiewitz> i noticed that several audio recordings are missing in King's Quest 6, and I would like to make sure that they are really missing.
[15:09] <m_kiewitz> (text/subtitles exist, but audio doesn't)
[15:12] <-- _sev|work left irc: Quit: This computer has gone to sleep
[15:16] --> Vampire0 joined #scummvm.
[15:17] <-- Vampire0_ left irc: Ping timeout: 260 seconds
[15:18] --> _sev|work joined #scummvm.
[15:18] <-- _sev|work left irc: Changing host
[15:18] --> _sev|work joined #scummvm.
[15:18] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[15:24] <wjp> waltervn: do you have time to walk me through SCI2 scis today?
[15:25] <wjp> (not having much luck with your getangle.s in SCI2 so far)
[15:27] --> Vampire0_ joined #scummvm.
[15:28] <-- Vampire0 left irc: Ping timeout: 260 seconds
[15:29] <wjp> (missing c_Str, s_data, s_format or something?)
[15:29] <waltervn> wjp: right, that's probably because I use symbol files that are written by a hacked scummvm
[15:31] <waltervn> https://github.com/waltervn/scummvm/tree/debugger Unfortunately, that branch hasn't been updated in almost 4 years
[15:32] <-- TMM left irc: Quit: Ex-Chat
[15:34] <waltervn> so c_Str is the Str class id, s_format is the 'format' selector. Those constants are output by a 'dump_symbols' debug command in that branch
[15:35] <wjp> ah, that does ring a bell
[15:35] <waltervn> scis then uses cpp to include that file... it's a bit of a hack I suppose, but it worked at some point in time :P
[15:36] <wjp> ah right, 3 years ago you talked me through it for qg1new
[15:40] <wjp> the good news is that the debugger branch merges with master without too much difficulty
[15:41] <waltervn> the sci2 stuff in scis probably needs some fixed, like that exports thing that was figured out recently
[15:41] <waltervn> *fixes
[15:45] <waltervn> unfortunately, it can be rather tricky to navigate the scummvm heuristics. Often you need to include certain things in the code to get scummvm to detect the right SCI version
[15:50] <wjp> the merge of debugger with master is the branch sci_dump_symbols in my fork, if you're interested
[15:52] <wjp> now trying to figure out how to actually get SSCI to run this
[15:56] <wjp> gk1 doesn't appear to react to the presence of this new 0.scr/0.hep at all
[15:56] <-- jamm left irc: Ping timeout: 240 seconds
[15:56] <waltervn> are there are other patches?
[15:56] <wjp> yes
[15:58] <waltervn> it's quite possible that I only tested it with SQ6
[15:59] <wjp> let me try that
[16:01] <wjp> ok, that works :-)
[16:05] <snover> wjp: what are you testing again? fileio? ive forgotten already.
[16:05] <wjp> (whole bunch of SCUMM games on sale on GOG, by the way)
[16:05] <wjp> snover: was thinking of trying these kRandom corner cases
[16:06] <snover> ah, right.
[16:06] <-- h00ligan left irc: Read error: Connection reset by peer
[16:07] <wjp> snover: I pushed a readstring patch, by the way
[16:09] <snover> i saw that! i need to clean up some stuff i was working on, then ill see what happens with Inside the Chest
[16:09] <snover> i suspect it will run into the (int)bytesRead > dest_r.maxSize path, but well see
[16:19] --> pegartillo joined #scummvm.
[16:20] <wjp> SQ6 kRandom doesn't behave like the comments in kRandom say SCI16 kRandom behaves
[16:20] <wjp> kRandom(5, 10) -> [5, 10]
[16:20] <wjp> kRandom(10, 5) -> (5, 10)
[16:20] <wjp> kRandom(0, -1) -> 0
[16:20] <wjp> ( [] = inclusive on both ends, () = exclusive on both ends)
[16:24] --> ajax16384 joined #scummvm.
[16:24] #scummvm: mode change '+o ajax16384' by ChanServ!ChanServ@services.
[16:25] <snover> hm, no, i have a bad memory. inside the chest uses kFileIOReadRaw, not ReadString.
[16:33] <-- _sev|work left irc: Quit: This computer has gone to sleep
[16:34] --> TMM joined #scummvm.
[16:34] #scummvm: mode change '+o TMM' by ChanServ!ChanServ@services.
[16:42] <wjp> any idea if bytesRead < size there too?
[16:43] <snover> it is not
[16:43] <snover> bytesRead == size
[16:45] <snover> but the whole thing is pretty messed up; it does a raw read into an int16 array, which blows up on master because int16 arrays arent considered raw so it tries to set reg_t values and eventually overflows. fixing that so int16s are treated as raw, nothing happens
[16:45] <snover> it reads in 8 bytes and then no more reads
[16:47] --> hurricos joined #scummvm.
[16:48] <wjp> so it might be ok on the script side but not on ours?
[16:48] <wjp> (related: are int16 arrays LE, BE, SE or native?)
[16:49] --> Vampire0 joined #scummvm.
[16:50] <-- Vampire0_ left irc: Ping timeout: 252 seconds
[16:51] <snover> im not sure, the scripts seemed to be setting up objects with an element size smaller than the amount of data to read (elems: 8, bytes: 20)
[16:51] <snover> (so 16 bytes of storage for 20 bytes of read)
[16:52] --> WooShell joined #scummvm.
[16:54] <snover> oh. actually that change to make int16 arrays raw is wrong
[16:54] <snover> int16 arrays are internally reg_ts
[16:55] <snover> i forgot, that change had to be made because games would request int arrays and then store memids in them
[16:56] --> ny00123 joined #scummvm.
[16:58] <-- hurricos left irc: Quit: ERC (IRC client for Emacs 24.5.1)
[16:58] <WooShell> meow =^.^=
[16:58] --> Vampire0_ joined #scummvm.
[17:00] <-- Vampire0 left irc: Ping timeout: 260 seconds
[17:01] <Strangerke|work> Guys, we have been accepted for GSoC 2017
[17:02] <Strangerke|work> I'll write an announcement in the evening to advertise about it
[17:02] <wjp> Great!
[17:02] <fydo> Awesome!
[17:03] <Strangerke|work> :)
[17:03] <m_kiewitz> wjp: interesting, so they must have changed some things (and/or the RNG used in sci32 behaves differently)
[17:04] <m_kiewitz> but wait, kRandom(0, -1) -> 0
[17:04] <m_kiewitz> while at the same time kRandom(10,5) -> (5,10)
[17:04] <m_kiewitz> ???
[17:04] <m_kiewitz> ah, so maybe they use a signed int32
[17:07] <Strangerke|work> 201 orgs this year, which means GSoC 2017 is the largest GSoC ever
[17:10] <m_kiewitz> wjp: can you please try what it does on for example 60.000? i guess it should read the 16-bit value as unsigned
[17:11] <Strangerke|work> So, the student period start now and ends on 20/3 with the student application forms deadline. We should all be ready to answer student's questions
[17:12] <wjp> m_kiewitz: that doesn't sound consistent
[17:12] <-- pegartillo left irc: Quit: http://www.kiwiirc.com/ - A hand crafted IRC client
[17:12] <m_kiewitz> huh?
[17:12] <wjp> because then it would treat kRandom(0,-1) as kRandom(0,65535)
[17:13] <m_kiewitz> that would be a major change then
[17:13] <m_kiewitz> larry 5 requires it to be unsigned
[17:13] <m_kiewitz> i don't know how you tested it out
[17:13] <m_kiewitz> i thought maybe you used a -1 sint32)
[17:14] <wjp> I'm not sure how I would pass a sint32 to kRandom in SSCI
[17:14] <m_kiewitz> i don't know that you passed it that way
[17:15] <wjp> in any case, kRandom(0, 60000) returns values between 60000 and 65536 when treating the output as uint16
[17:15] <m_kiewitz> anyway, so kRandom(0, 32767) -> (0...32767)
[17:15] <wjp> (don't have enough samples to see if it's inclusive or exclusive)
[17:16] <m_kiewitz> that's definitely completely different to kRandom in sci16
[17:16] <m_kiewitz> kRandom(0, 32768) would then return (32768 ... 65535)?
[17:17] <m_kiewitz> which one did you test?
[17:17] <m_kiewitz> SQ6?
[17:17] <wjp> SQ6
[17:18] <m_kiewitz> <wjp> kRandom(10, 5) -> (5, 10)
[17:19] <m_kiewitz> () exclusive on both ends
[17:19] <m_kiewitz> does this mean effectively 6 up to 9? so it won't return 5 nor 10?
[17:19] <wjp> indeed
[17:19] <m_kiewitz> really weird behavior
[17:19] <wjp> kRandom(0, 32768) indeed returns values between 32768 and 65536 (i.e., -32768 up to 0)
[17:20] --> FleetAdmiral joined #scummvm.
[17:21] <FleetAdmiral> Hey guys! I'm interested in participating through GSoC this year, i have a few questions regarding the project
[17:22] <wjp> m_kiewitz: yeah, that (10, 5) doesn't seem to make much sense
[17:22] <FleetAdmiral> I'd like to work on the GUI part, could i know what technology is used for the same?
[17:22] <m_kiewitz> i guess they used some RNG which behaved so weird and also sees the parameters as signed
[17:22] <m_kiewitz> this definitely was not the case in sci16
[17:23] <m_kiewitz> but good that we found out. this would be one of the things that can create really weird issues and would take ages to figure out
[17:24] <Strangerke|work> FleetAdmiral: hi :) And welcome!
[17:25] <Strangerke|work> FleetAdmiral: We have two GUI tasks, both of them would be mentored by _sev
[17:27] <Strangerke|work> FleetAdmiral: He may not be available right now, because at this time of the day he may be on his way back home.
[17:27] <FleetAdmiral> Oh alright, any pointers on where i could start working for the same?
[17:28] <Strangerke|work> sorry, I don't understand "for the same?"
[17:29] --> _sev|work joined #scummvm.
[17:29] #scummvm: mode change '+o _sev|work' by ChanServ!ChanServ@services.
[17:29] <FleetAdmiral> for GUI related tasks
[17:29] <Strangerke|work> There are some details about the GUI tasks on the wiki
[17:29] Action: Strangerke|work pokes _sev|work
[17:30] <_sev|work> I'll be able to talk only in 2-3 hours or so
[17:30] --> GitHub12 joined #scummvm.
[17:30] <GitHub12> [scummvm] sev- pushed 2 new commits to master: https://git.io/vyLoz
[17:30] <GitHub12> scummvm/master 9278380 Eugene Sandulenko: DIRECTOR: Improved debug output
[17:30] <GitHub12> scummvm/master 63f168c Eugene Sandulenko: DIRECTOR: Fix frame script execution...
[17:30] GitHub12 (GitHub12@192.30.252.45) left #scummvm.
[17:30] <Strangerke|work> FleetAdmiral: Could you wait 2 or 3 hours and ask more detailed questions to _sev?
[17:31] <FleetAdmiral> Yea sure i'll do that
[17:31] <FleetAdmiral> i'll look about the wiki in the meantime
[17:31] <Strangerke|work> that sounds perfect :)
[17:31] <FleetAdmiral> Awesome, thanks for the help guys :D
[17:31] <Strangerke|work> yw :)
[17:33] <-- _sev|work left irc: Client Quit
[17:39] <-- diehumblex left irc: Ping timeout: 240 seconds
[17:41] --> diehumblex joined #scummvm.
[17:43] <-- Strangerke|work left irc: Quit: Bbl
[17:45] --> LittleToonCat joined #scummvm.
[17:45] --> Littleboy joined #scummvm.
[17:45] #scummvm: mode change '+o Littleboy' by ChanServ!ChanServ@services.
[17:46] <_sev> FleetAdmiral: all the GUI code is custom. You may find it in gui/ directory in our repository. Also, there is a custom XML description file, you may find all info on our Wiki in http://wiki.scummvm.org/index.php/GUI_Themes/Specs
[17:46] <_sev> FleetAdmiral: perhaps you have more specific question?
[17:46] <FleetAdmiral> No, i just wanted leads to the GUI content, i'll let you know if i want more, thanks :D
[17:49] --> Henke37 joined #scummvm.
[17:59] <rsn8887> sev: I am thinking of doing the rebase myself and then simply do a pull request from my rebased branch to cpasjuste's branch? If he pulls it, the changes should show up on the current Vita PR, and everything should be ok.
[17:59] <rsn8887> But he will have to do a "forced" merge I guess. I am not familiar how that works
[17:59] <rsn8887> sev: I am not the owner of cpasjuste's branch
[18:05] <_sev> rsn8887: yes, but he might give you access to that, so then you can do it
[18:05] <_sev> rsn8887: but yes, do it in your branch and then he could fetch it from you
[18:22] <wjp> m_kiewitz: thinking about it some more, this kRandom behaviour does make some sense if you assume (positiveNumber % negativeNumber) returns a value in (negativeNumber, 0]
[18:23] <wjp> then it's just "if (to - from + 1 == 0) return from; else return from + (random() % (to - from + 1))"
[18:23] <wjp> depending on the implementation-defined behaviour of %
[18:24] <wjp> or maybe in [negativeNumber, 0)
[18:24] <wjp> yeah, that last one
[18:48] --> vinayakvivek joined #scummvm.
[18:59] <-- FleetAdmiral left irc: Ping timeout: 260 seconds
[19:03] --> SylvainTV joined #scummvm.
[19:03] #scummvm: mode change '+o SylvainTV' by ChanServ!ChanServ@services.
[19:27] --> Farmboy0 joined #scummvm.
[19:27] <-- Farmboy0 left irc: Changing host
[19:27] --> Farmboy0 joined #scummvm.
[19:34] <rsn8887> sev: Ok I will do it in my branch.
[19:44] <rsn8887> sev: so quick question, I rebase everything, then merge it into a local copy of master, do the manual merges, push it to my repo, make a PR for Cpasjuste so he can fetch it?
[20:07] <wjp> that sounds like the gist of it
[20:07] <rsn8887> Hmm I am getting merge conflicts on my rebase on old commits I have nothing to do with
[20:07] <rsn8887> How is this possible?
[20:07] <rsn8887> Those have all been merged into master already.
[20:07] <wjp> then you may be doing the wrong thing
[20:08] <wjp> it's relative easy to accidentally rebase a bunch of old commits onto the wrong branch
[20:09] <wjp> if you don't already do it, you should always double-check the result of a rebase with a (graphical) history viewer to ensure everything went right
[20:09] <wjp> (and if doing complicated merges it is also always a good idea to make backup branches)
[20:10] <rsn8887> Can I just checkout a fresh version of vita branch. Then do a git pull --rebase from scummvm/scummvm master? Then rebase everything. Then push to Cpasjuste?
[20:11] <rsn8887> That is my usual workflow, only normally I git pull --rebase from cpasjuste/scummvm:master not scummvm/scummvm:master
[20:11] <wjp> I don't know your setup or the vita branch
[20:12] <rsn8887> I will try it.
[20:12] <rsn8887> On copies of my branches to be sage
[20:12] <rsn8887> safe
[20:12] <wjp> also make sure to abort any in-progress rebases/merges
[20:12] <rsn8887> wjp: yes I learned that the hard way lost a night fiddling with reflog. Quite easy to break everything in git.
[20:13] <rsn8887> Surprisingly easy
[20:13] <wjp> it's also quite easy to avoid breaking everything, fortunately :-)
[20:13] <rsn8887> If it wasn't for reflog, I would have quite this project 3 weeks ago.
[20:13] <wjp> (backup branches, visual checks after steps, abort if strange things happen)
[20:22] <rsn8887> I don't understand. I am doing git rebase -i HEAD~2 and it is showing me dozens of commits. I thought it would only show the last two.
[20:22] <wjp> is one of the last two a merge?
[20:23] <rsn8887> Yes.
[20:23] <wjp> that would do it
[20:23] <wjp> you can try rebase -ip, but YMMV
[20:24] <wjp> (and you'll have to re-do any merge conflict resolutions anyway)
[20:28] --> GitHub158 joined #scummvm.
[20:28] <GitHub158> [scummvm] criezy pushed 1 new commit to master: https://git.io/vytUS
[20:28] <GitHub158> scummvm/master fc0f366 Ben Castricum: I18N: Update translation (Dutch)...
[20:28] GitHub158 (GitHub158@192.30.252.40) left #scummvm.
[20:32] <rsn8887> lots of conflicts. Strange, I thought the branch was already pretty much merged.
[20:32] <rsn8887> Last time I merged in scummvm/scummvm master, I only had 3 conflicts.
[20:33] <rsn8887> Also it doesn't show me AT ALL where those conflicts are from (which commits have caused them)
[20:34] <wjp> what exactly are you trying to rebase/merge into what?
[20:41] <rsn8887> I am taking my local up-to-date copy of the vita branch (I have been working on it with Cpasjuste).
[20:42] <rsn8887> And I am pull-rebasing the scummvm/scummvm master on top.
[20:42] <rsn8887> well I shouldn't say on top.
[20:42] <rsn8887> I am simply pull-rebasing the scummvm master into the vita branch to fix all merge conflicts.
[20:42] <rsn8887> I was hoping the only conflicts would be in files that cpasjuste or me had actually created or modified
[20:43] <rsn8887> But I am getting completely unrelated merge conflicts. It is as if I am getting all the conflicts that have recently happened in other peoples commits that they have already fixed.
[20:47] <snover> rsn8887: is the branch you are trying to rebase onto scummvm/master online somewhere that i can pull it?
[20:47] <rsn8887> Yes one second let me verify/push the latest version.
[20:51] <rsn8887> snover: it is here:
[20:51] <rsn8887> https://github.com/rsn8887/scummvm/
[20:51] <rsn8887> master branch of that repo
[20:52] <rsn8887> I am trying to do git pull --rebase the scummvm master onto my master.
[20:52] <-- kurtwr left irc: Ping timeout: 255 seconds
[20:52] <rsn8887> But it is not working (way too many conflicts).
[20:52] <wjp> that sounds like the wrong direction
[20:53] <wjp> you want to keep scummvm master the same, not put it on top of something else
[20:53] <rsn8887> Hmm I thought it is pretty much the same. But I will try it the other way round.
[20:55] <wjp> you know that cpasjuste's master has already merged your master, right?
[20:59] <rsn8887> Yes. I am trying to get a local branch ready that I can use to do the rebasing that _sev requested.
[20:59] <rsn8887> I could pull cpasjuste's, but our branches are the same apart from 1 merge commit.
[20:59] <rsn8887> where he merged my branch into his.
[20:59] <snover> this 8c832ff commit should not be here
[21:00] <snover> its some manual merge of code without an actual merge commit
[21:00] <snover> of code from scummvm/master*
[21:00] <wjp> yes, that's quite bizarre, although the next commit presumably mostly takes care of that
[21:01] <snover> the conflicts in `PSP2: add psvita support` are legitimate conflicts
[21:01] <wjp> the very first commit of the branch is also a bit weird, being a merge commit of two commits in master
[21:02] <snover> git cant know that 8c832ff is the same as code from master, so can only assume everything is a conflict because it looks like you changed all the code on your branch manually
[21:02] <snover> looks to git*
[21:02] <rsn8887> But in https://github.com/scummvm/scummvm/pull/901 there are only 4 conflicts listed!
[21:02] <wjp> a quick graft + filter-branch should take care of it (but that's relatively black magic)
[21:02] <rsn8887> why is it now showing so many
[21:03] <wjp> because any conflicts there have already been resolved by later merges
[21:03] <wjp> but rebasing triggers everything again
[21:03] <rsn8887> Oh man, is there a way to not trigger everything again?
[21:04] <rsn8887> The last thing I want to do is re-resolve conflicts that cpasjuste has already resolved
[21:05] <wjp> fixing up the oddities might be nice, but I'm not sure if rebasing the whole thing is worth it
[21:05] <rsn8887> So I think I might understand now: A merge does not show all those conflicts, but a rebase does.
[21:05] <wjp> 'rebase' is basically replaying all commits from your current branch onto another branch
[21:05] <rsn8887> That makes sense, becaue the PR is a merge, but what I am trying to do is pull --rebase
[21:05] <rsn8887> instead of just pull (normal merge)
[21:06] <rsn8887> Is there a "rebase + apply same merge resolutions"? I mean the resolutions have already been made
[21:07] <wjp> not really, as they have been made after a group of commits, not after individual commits
[21:07] <rsn8887> So what if:
[21:07] --> kurtwr joined #scummvm.
[21:07] <rsn8887> I rebase the branch locally (without merging or pulling anything)
[21:07] <rsn8887> And then merge afterwards?
[21:08] <rsn8887> Hmm I guess it might show the same problem of many merge conflicts again then.
[21:08] <snover> git pull --rebase is just git fetch + git rebase
[21:08] <rsn8887> and git pull is git fetch + git merge. The merge is much simpler (almost zero conflicts).
[21:08] <rsn8887> The rebase is a pain.
[21:09] <rsn8887> It also shows a lot of "deleted" file merges and some files with >100 merge conflict lines.
[21:10] <wjp> that broken merge from earlier and duplicate commits that have already been merged into master in separate PR's will make this interesting
[21:10] <snover> a merge commit gives you an extra place to fix the conflicts (in the merge commit itself); a rebase requires you to fix whatever conflicts exist for each commit because it replays each commit one by one against the head of the parent branch
[21:11] <wjp> if we want to rebase this, I'd first fix up that broken merge, then remove those duplicate commits if at all possible, and then try
[21:13] <rsn8887> snover: the head as it was back at the time of that commit?
[21:13] <snover> no
[21:13] <rsn8887> snover: or the *current* head?
[21:14] <snover> the head of whatever branch you are rebasing onto
[21:14] <snover> e.g. `git rebase scummvm/master`
[21:14] <rsn8887> snover: I guess that is why there are so many conflicts.
[21:14] <snover> that would rebase against the head of your local copy of the scummvm forks master branch
[21:15] <rsn8887> snover: but shouldn't it at least skip commits that are identical in both branches?
[21:15] <wjp> it will try to
[21:15] <snover> sure, but these 'merge' conflicts are not identical to any commit in the other branch
[21:15] <rsn8887> snover: or is it replaying all commits all the way from when cpasjuste first forked?
[21:16] <wjp> it is doing that, yes
[21:16] <snover> rebase looks back through the history to find the first common commit and rebases from that point
[21:17] <wjp> (technically that would be the root of the whole repo, but nitpicking :-) )
[21:17] <rsn8887> snover: So say Criezy made a commit and resolved a merge conflict manually then. When I rebase, I shouldn't have to re-resolve that conflict again
[21:17] <wjp> unfair, isn't it? :-)
[21:18] <wjp> the thing is that it's a completely different thing
[21:18] <wjp> the merge resolved only the end results
[21:18] <wjp> a rebase replays individual commits
[21:18] --> Begas_VBox joined #scummvm.
[21:18] <rsn8887> snover: But I never had any problem doing rebase's of my local copy of Cpasjuste's branch before.
[21:18] <wjp> those probably only involved a few commits
[21:19] <rsn8887> wjp: True... I guess the problem appears when going back 60 commits to the beginning of the fork
[21:19] <wjp> yes
[21:19] <wjp> with a few complications in there
[21:19] <rsn8887> wjp: I feel merge is a more natural solutions, although it displays a "less-finegrained" history.
[21:19] <wjp> the merge-gone-wrong halfway, and the chain of SDL commits in the branch that are almost-but-not-quite in master
[21:20] <snover> these not-a-merge merge commits really ought to go though, since thats just duplicate code of other commits that should not exist in the history
[21:20] <rsn8887> How can a merge go wrong halfway? I would think git should protect against that
[21:20] <wjp> you keep saying "should" :-)
[21:21] <wjp> git will happily let you shoot yourself in the foot if you tell it to
[21:21] <wjp> i.e., always double-check the results
[21:21] <-- ajax16384 left irc: Quit: Leaving
[21:21] <rsn8887> what is an SDL commit, and where is the merge gone wrong?
[21:22] --> Vampire0 joined #scummvm.
[21:22] <wjp> 21:59 <@snover> this 8c832ff commit should not be here
[21:22] <wjp> that's the merge-gone-wrong
[21:22] <wjp> the SDL commits are the kbdmouse stuff
[21:23] <-- Vampire0_ left irc: Ping timeout: 268 seconds
[21:23] <rsn8887> Oh the SDL stuff, yes but there were changes to PSP2 files in there.
[21:24] <wjp> that fact doesn't really reduce the conflicts
[21:25] <wjp> 308de528802e7c06f17932e5dc4d85634af57364 is probably also an issue
[21:25] <wjp> cleaning this up will be fun
[21:26] <snover> FWIW, this is the knowledge cliff of git that everyone runs into eventually. once you get to the other side you will have much fewer problems :)
[21:27] <snover> its a painful but rewarding process.
[21:28] <wjp> I think part of this will have to be postponed until the pending PRs are merged
[21:29] <rsn8887> Is it possible to squash all the vita commits into one big commit. Diff the result with current master branch, resolve the conflicts and be done with it.
[21:29] <wjp> 911 and 912 in particular
[21:29] <wjp> depends on if the history is at all valuable
[21:29] <rsn8887> E.g. a merge commit with all commits squashed as just a single commit "add vita support" I don't think the history is important.
[21:30] <wjp> that would not be a merge commit
[21:30] <wjp> but doing that would certainly be an option
[21:30] <rsn8887> What I don't like about a merge commit currently is that somehow it "opens up" as many separate commits on subsequent "rebase" I would like it to be just one commit that shows as one on a rebase
[21:31] <rsn8887> If you look at the history and the commit messages there's really not much in there but a standard "initial Vita port"
[21:31] <-- Begas_VBox left irc: Quit: Vision[0.9.8]: i've been blurred!
[21:32] <rsn8887> The problem is currently I cannot even squash all psp2 commits into one, because rebase gives me a ton of manual merges. I wonder if a giant "patch" would be a solution
[21:33] <wjp> that's pretty easy. Do a merge with master, then a soft reset to master, then a commit
[21:33] <rsn8887> E.g. a patch that takes current master and brings it to resemble the vita branch after merge with master.
[21:33] <rsn8887> and that patch is applied as a single commit.
[21:33] <rsn8887> hmm I did the merge with master. How do I do the soft reset.
[21:34] <wjp> git reset --soft scummvm/master
[21:34] <wjp> and then just 'git commit'
[21:34] <rsn8887> ah ok I will try that. Maybe you and sev can decide if that would be the best way to commence with this.
[21:34] <wjp> but I'd still suggest waiting until after #911 and #912 are in
[21:35] <rsn8887> I feel that might be good. We loose all history of vita branch, but I don't think anybody cares about the history of those 30 commits or so.
[21:35] <wjp> if necessary I can also take a few hours later this week to go through the history and fix everything up
[21:36] <rsn8887> Especially since the messages are not very telling anyways and many are just "update readme"
[21:36] <rsn8887> I think If you diff that soft reset commit with current master, you will see the actual changes are not that large.
[21:37] <rsn8887> I guess that is shown by the current vita PR. It would not be an overwhelmingly huge commit. ANd it would fix the SDL problem that I created
[21:37] <wjp> ok
[21:37] --> girafe joined #scummvm.
[21:38] <rsn8887> I agree that our current PR's regarding the control menu should probably be merged first, since that will reduce the changes introduced by the single vita soft reset + commit action.
[21:39] --> kurtwr2 joined #scummvm.
[21:39] <-- kurtwr left irc: Ping timeout: 260 seconds
[21:45] Action: wjp removes that kRandom test script from SQ6 again before he wonders next month why SQ6 is suddenly broken :-)
[21:47] <rsn8887> I tried the single commit. It shows 109 changed files and 6648 additions. But it looks very nice to go through, quite logical additions.
[21:48] <rsn8887> 64 deletions. I think it would be the best option.
[21:49] <-- ny00123 left irc: Quit: Leaving
[21:50] <rsn8887> With the other SDL PR merged into mainline ScummVM, the additions will be even less.
[21:50] <wjp> Ok, so that sounds like a realistic option
[21:51] <rsn8887> I am a bit concerned that PR 901 shows 125 files changed and my single pull request shows 109 files changed ?!?! Those should be the same ?!?!?
[22:14] <-- vinayakvivek left irc: Quit: Connection closed for inactivity
[22:17] --> Deledrius__ joined #scummvm.
[22:19] <-- Deledrius_ left irc: Ping timeout: 255 seconds
[22:21] --> marcos-sb joined #scummvm.
[22:25] <-- waltervn left irc: Quit: Leaving
[22:29] <-- marcos-sb left irc: Quit: marcos-sb
[22:41] <-- m_kiewitz left irc: Quit: technology isn't intrinsically good or evil. It's how it's used. Like the Death Ray.
[23:01] --> pegartillo joined #scummvm.
[23:03] <-- bgK left irc: Ping timeout: 260 seconds
[23:03] --> dreammaster joined #scummvm.
[23:03] #scummvm: mode change '+o dreammaster' by ChanServ!ChanServ@services.
[23:05] --> bgK joined #scummvm.
[23:05] #scummvm: mode change '+o bgK' by ChanServ!ChanServ@services.
[23:17] <-- Farmboy0 left irc: Remote host closed the connection
[23:20] --> Deledrius joined #scummvm.
[23:21] <-- Deledrius__ left irc: Ping timeout: 260 seconds
[23:24] --> Deledrius_ joined #scummvm.
[23:27] <-- Deledrius left irc: Ping timeout: 260 seconds
[23:41] <-- 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.
[23:42] --> Deledrius__ joined #scummvm.
[23:43] Nick change: Deledrius__ -> Deledrius
[23:43] <-- Deledrius left irc: Client Quit
[23:43] --> Deledrius joined #scummvm.
[23:45] <-- Deledrius_ left irc: Ping timeout: 240 seconds
[00:00] --- Tue Feb 28 2017