IRC logs for #openttd on OFTC at 2019-01-30
⏴ go to previous day
00:06:51 *** Lejving_ has joined #openttd
00:23:52 <Samu> hmm, looks like company_num is a COMMA
00:28:35 <Samu> How do i change window priority of AI Settings
00:28:52 <Samu> it even gets in front of console window
00:29:07 <Samu> it doesn't need this much priority
01:02:36 <peter1138> On branch nrt-block-rebased
01:02:40 <peter1138> Your branch and 'origin/nrt-block-rebased' have diverged,
01:16:32 * peter1138 performs splittages
01:34:31 <Samu> why is AI config window WC_GAME_OPTIONS ?
01:34:51 <peter1138> History, I imagine.
01:36:42 <Samu> by being WC_GAME_OPTIONS it was always prioritized :(
01:37:03 <Samu> always in front of every others
01:37:41 <Samu> must see if i can solve this "always in front of others" problem for good
01:42:20 <Samu> WN_GAME_OPTIONS_GAME_OPTIONS, ///< Game options.
01:42:27 <Samu> well that wasn't so obvious
01:43:13 <Samu> I see this is a bit messy
01:45:38 <Samu> shouldn't they be their own class?
01:45:46 <Samu> they're a window number :|
01:46:24 <peter1138> And you can close by WindowClass.
01:47:12 <Samu> it's WC_GAME_OPTIONS, WN_GAME_OPTIONS_GAME_OPTIONS to refer to Game options. this looks messy
01:47:48 <Samu> WC_GAME_OPTIONS, WN_GAME_OPTIONS_AI for the AI/GS Config window
01:48:37 <Samu> WN_GAME_OPTIONS_NEWGRF_STATE
01:50:12 <peter1138> anyway, yeah, settings were meant to be on top
01:50:19 <peter1138> because they're kinda modal
01:53:00 <nnyby> is there a key command to pause the game?
01:55:27 <ST2> just tried, F1 doesn't work :|
01:56:16 <nnyby> it works for me, on 1.8.0
01:56:35 <nnyby> on many laptop keyboards, you need to press the Fn key to get to F1 though. not that easy
01:56:43 <ST2> maybe you forgot to say: in SP games ^^
01:57:16 <nnyby> i was thinking it would be nice to map the space bar to the pause operation. that works well for Baldur's Gate
01:57:29 <ST2> oh yeah, people still run MP games that way
02:22:37 <Samu> static uint GetWindowZPriority(WindowClass wc)
02:26:11 <Samu> WC_GAME_OPTIONS is quite high in priority
02:26:41 <Samu> must detach the AI Config window from WC_GAME_OPTIONS
02:26:49 <Samu> there seems to be no other way
02:45:33 <peter1138> Hmm, how do I compile a nightly-compatible server, then?
02:50:45 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
02:53:49 <peter1138> No, I mean get the right revision.
02:54:22 <peter1138> I think it's a mistake to include the branch in the network revision.
02:54:56 <peter1138> When you check out the specific commit that the nightly was built on, unless it's still at the tip of master, then it is no longer master but a detached head.
02:55:09 <peter1138> You can do git reset --hard xxx but that's kinda shit.
02:55:56 <peter1138> Just the sha hash should be enough for compatibility.
02:57:54 <peter1138> As it is, my revision is now "20190129-maste" which is terrible :p
03:09:34 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7121: Fix #7021: Better revision strings for network and gamelog https://git.io/fh1w4
03:13:21 <Eddi|zuHause> peter1138: which is why i suggested using only the hash for network comparison
03:13:56 <peter1138> But I don't see your suggestion in #7121 :)
03:13:59 <Eddi|zuHause> peter1138: as for "which revision", there used to be finger.openttd.org
03:14:10 <peter1138> Which revision is not a problem.
03:14:16 <peter1138> Having to use git reset --hard IS a problem.
03:14:38 <peter1138> Or maybe it's not. I dunno.
03:15:02 <Eddi|zuHause> unless you got weird local changes
03:15:39 <peter1138> Yes, git checkout will work when we stop including the branch name.
03:20:30 <Eddi|zuHause> yeah, that seems to be a problem
03:20:51 <peter1138> 02:13 < Eddi|zuHause> peter1138: which is why i suggested using only the hash for network comparison
03:20:56 <peter1138> ^ From that I thought you knew :)
03:21:43 <Eddi|zuHause> no, the problem i had in mind was that the string is too long, so the distinguishing bit (the hash) at the end is cut off
03:22:23 <peter1138> Yeah that too. Of course, it would then have been "compatible" with any revision, except for the that that 'master' gets lost.
03:22:49 <peter1138> Anyway, up way too late again.
03:23:18 <Eddi|zuHause> so that additional problem with the branch name is that git doesn't remember which branch the commit was originally from, "git branch" is only the head commit
03:24:20 <peter1138> It's a little difference between svn and git branches.
03:24:29 <Eddi|zuHause> also git and hg branches
03:24:48 <peter1138> And there is of course the fact that git branches don't even have to have the same name, yet be the same.
03:25:23 <peter1138> It's really JUST the sha hash (and modified) that matters.
03:25:41 <peter1138> Date might help manually identify it I suppose.
03:26:03 <Eddi|zuHause> peter1138: the branch name is useful in case of patchpacks and stuff
03:26:33 <Eddi|zuHause> for the display name, not the online compatibility
03:26:39 <peter1138> branch name is really a local identifier.
03:27:21 <Eddi|zuHause> peter1138: most notably for people trying to send bug reports
03:27:26 <peter1138> Right, but the network protocol doesn't have a separate displayname version.
03:27:49 <peter1138> Yeah, sure, I'm not arguing against including it everywhere. Just not in network compability.
03:27:57 <peter1138> Anyway, night night :)
03:30:35 <Eddi|zuHause> hm, some steam update seems to have silently disabled wine for non-official games
03:31:00 <Eddi|zuHause> now i ticked that again, and it wants to redownload every game that used it before
03:34:32 <Samu> there is the svn counter yet
03:34:38 <Samu> why did you give up on it?
03:59:37 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 commented on pull request #7121: Fix #7021: Better revision strings for network and gamelog https://git.io/fh1rD
04:18:14 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick updated pull request #7094: Fix #7088: Retrieve an appropriate name for a non-existant AI/GS when… https://git.io/fhrqY
04:30:59 *** snail_UES_ has joined #openttd
05:44:26 *** HerzogDeXtEr has joined #openttd
06:18:24 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas opened pull request #7141: Fix: unicode characters in Romanian town names https://git.io/fh1id
06:24:07 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7141: Fix: unicode characters in Romanian town names https://git.io/fh1id
07:15:49 *** andythenorth has joined #openttd
07:39:16 <andythenorth> so I should turn breakdowns on for my games?
07:41:25 <Pikka> breakdowns are "fun", inasmuch as predicable gameplay is "unfun".
07:52:23 <andythenorth> probably have to set reliability in my newgrfs?
07:52:32 <andythenorth> or not, just let OpenTTD do it :)
08:24:50 *** Vasatko has joined #openttd
08:44:55 * Pikka is banging his head over code which is supposed to stop houses being built near industries not working
08:45:15 <Pikka> until I realise the map generator places the towns before the industries
08:45:43 <Pikka> so of course it won't work on a newly-generated map
08:53:23 <andythenorth> I have nothing :P
09:06:18 <peter1138> I wonder how maps would look if it alternated between towns and industries?
09:08:30 <Pikka> well, it associates the industries with the nearest town when it places them, right?
09:08:57 <peter1138> Yeah so it'll still be after.
09:09:23 <peter1138> But then another town comes along, and that one will be after.
09:09:48 <peter1138> But still, most likely it wouldn't affect your issue much at all, perhaps just make it even more confusing as it "sometimes works".
09:09:58 <Pikka> yes, how it is now is still best
09:10:25 <Pikka> too many issues with doing anything else. and I can easily force-remove the houses if I really care about it.
09:10:28 <andythenorth> trying to keep towns away from industries? o_O
09:11:20 <peter1138> Hmm, I should enable TaI again. The sprawlyness was nice.
09:11:23 <Pikka> just stopping nice residential buildings right up against industries
09:11:52 <andythenorth> yeah TaI is winning town set
09:12:09 <andythenorth> TaI, Av9.8 and that bridge grf
09:12:23 *** WWacko1976-work has joined #openttd
09:12:26 <peter1138> (Which bridge grf?)
09:12:31 <andythenorth> dunno there are loads
09:13:12 <andythenorth> expensive, short and slow
09:13:28 <peter1138> Oh yeah, 1.8.0 still crashes due to ICU.
09:13:48 <andythenorth> nobody went back and fixed all the 1.8.0 binaries? :o
09:14:03 <andythenorth> probably better do 1.9 soon then
09:14:55 <peter1138> Okay I guess I'll figure out that regression failure.
09:16:18 <peter1138> It was much easier to sort out the sprites after I split them into new files. No image conflicts!
09:23:06 <peter1138> ai/regression/run.sh -g < starts it in gdb
09:35:32 <peter1138> Ok, when did we reduce the mapsize by 1 in each direction?
09:54:54 <peter1138> I believe we added MP_VOID tiles around the top of the map.
10:07:20 *** andythenorth has joined #openttd
10:11:05 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on issue #7142: Assertion when building long road at map border https://git.io/fh1Qa
10:12:03 <peter1138> Seems we have missed tile bounds checking since most maps have MP_VOID at the edges these days.
10:12:13 <peter1138> But old maps don't, and you can still disable that setting.
10:13:09 <peter1138> So why did we add MP_VOID around the top edges? If it was to avoid crashes like this, then it's not right :p
10:13:35 <peter1138> If it is, we need to update the regression test savegame.
10:14:53 <peter1138> (svn r15190) -Feature: Allow terraforming of the tiles at the edges of the map.
10:16:32 <peter1138> On the other hand, it should be simple to fix this CmdBuildLongRoad
10:18:03 <peter1138> The actual bug was introduced in 2012.
10:18:10 <peter1138> 55dccee02fc03118929df2133fcd9140a947024c
10:18:22 <peter1138> CanConnectToRoad() doesn't check for bounds.
10:22:57 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN opened pull request #7143: Fix #7142: Missing map bounds check when building long roads. https://git.io/fh174
10:26:46 <peter1138> (Probably took me longer to fill in the template than to fix it ;))
10:26:53 <peter1138> But, gotta set a good example, eh?
10:29:28 <peter1138> Okay, regression test fails but doesn't crash now.
10:30:02 * peter1138 quickly runs regression tests on work stuff so it's ambiguous
10:58:59 *** Smedles has joined #openttd
11:01:16 <LordAro> peter1138: if it helps, i've occasionally been using ottd to test our own tools at work as well :p
11:02:12 <LordAro> peter1138: i really feel that crash should've been noticed earlier
11:05:15 <planetmaker> people don't build roads at the map edge usually... I guess. But ... 7 years for that to not surface?
11:10:10 <planetmaker> though... it only triggers on ancient savegames
11:10:50 <planetmaker> when you change a setting on them
11:12:30 <planetmaker> in essence: I'm not exactly able to produce a map which triggers it :P
11:16:56 <andythenorth> I use OpenTTD to test work stuff
11:17:29 <andythenorth> I built a couple of UI libraries which use the same python compile as andy newgrfs
11:20:15 <peter1138> planetmaker, needs freeform_edges to be false, which is rare. It's not setting in the GUI.
11:21:03 <andythenorth> settings explosion :)
11:22:12 <peter1138> Oops, I accidentally committed an openttd.grf!
11:23:46 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7141: Fix: unicode characters in Romanian town names https://git.io/fh1Ne
11:24:55 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN updated pull request #7143: Fix #7142: Missing map bounds check when building long roads. https://git.io/fh174
11:26:24 <peter1138> Files which the build process modifies shouldn't be in git ;(
11:38:31 <LordAro> it's not *very* different to having an autotools configure in your repo
11:46:14 <peter1138> Who even uses that any more? :D
11:54:49 <peter1138> LordAro, and actually, those projects that do use autotools still don't put those generated files in git.
12:16:16 *** WWacko1976-work has quit IRC
12:24:52 *** gelignite has joined #openttd
12:37:53 * andythenorth deletes a chunk of UI design
12:37:59 <andythenorth> made, it refined it, deleted it
12:48:01 <peter1138> Is it lunch time yet?
12:48:18 <peter1138> I've been trying to make some error handling work. But it errors.
12:49:59 * peter1138 eats a 'Teaser Celebration instead.
13:01:14 <andythenorth> such very trains
13:02:27 <andythenorth> wagons moving by gravity? o_O
13:02:32 <andythenorth> what railtype is that :P
13:23:14 <Pikka> how much TE does gravity have?
13:31:00 *** WWacko1976-work has joined #openttd
13:54:53 <DorpsGek_II> [OpenTTD/OpenTTD] planetmaker approved pull request #7141: Fix: unicode characters in Romanian town names https://git.io/fhMTN
13:56:20 <planetmaker> at least for 2 out of 2 names I checked, the fix was correct :P
13:57:04 <peter1138> Needs the commit message to be fixed for the CI.
13:57:09 <peter1138> Then it'll need approving again.
13:57:54 <peter1138> Our commit hooks reject non-ASCII commit messages.
13:58:30 <planetmaker> "Fix: unicode characters in Romanian town names" <<-- non ascii?
13:58:53 <planetmaker> Unable to find image 'openttd/compile-farm-ci:commit-checker' locally
13:58:53 <planetmaker> commit-checker: Pulling from openttd/compile-farm-ci <-- that's the failure :| stupid MS
13:59:18 <peter1138> Many of these town names were using the 'ã' character, which should
13:59:18 <peter1138> actually by 'ă'. There were other missing accents as well which I've
13:59:55 <planetmaker> ah... bah, stupid way the errors are shown. Very stupid
14:01:01 <peter1138> Jenkins was a little better.
14:01:19 <planetmaker> we had Atlassian's bamboo
14:01:23 <peter1138> Loads of clicks with Azura Pipelines.
14:02:58 <planetmaker> but yes, Jenkins is a bit easier there, too
14:04:25 <DorpsGek_II> [OpenTTD/OpenTTD] planetmaker requested changes for pull request #7141: Fix: unicode characters in Romanian town names https://git.io/fhMkX
14:04:38 <planetmaker> dunno how I could have dismissed my review before :P
14:11:26 <Eddi|zuHause> <planetmaker> Unable to find image 'openttd/compile-farm-ci:commit-checker' locally <-- that bit is because it's parsing the output in a wrong way, i think... TrueBrain said "it can be fixed, but somebody has to do it" and i have no clue where to even begin looking for what to change
14:11:41 <LordAro> not sure you can disamiss an approving review
14:15:21 <Eddi|zuHause> LordAro: not helpful :p
14:16:27 <Eddi|zuHause> also, there's a second issue where the azure errors (assuming they have been reduced to the relevant bit) don't get propagated to github, where TrueBrain said it only does that for the windows builds
14:33:37 *** snail_UES_ has joined #openttd
14:37:20 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7141: Fix: unicode characters in Romanian town names https://git.io/fh1id
14:38:31 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas commented on pull request #7141: Fix: unicode characters in Romanian town names https://git.io/fhMtI
14:52:52 <DorpsGek_II> [OpenTTD/OpenTTD] planetmaker approved pull request #7141: Fix: unicode characters in Romanian town names https://git.io/fhMq8
14:52:58 <DorpsGek_II> [OpenTTD/OpenTTD] planetmaker merged pull request #7141: Fix: unicode characters in Romanian town names https://git.io/fh1id
15:14:42 <Samu> i still don't know how to fix it without touching too many code
15:15:05 <peter1138> If you still intend to do it, you don't need to close the PR.
15:15:27 <peter1138> Oh, it was closed 3 days ago. Never mind.
15:16:11 <Samu> I mean, I have it fixed on my ai gui topic
15:16:54 <Samu> I have different rules to what is enabled or disabled
15:17:07 <Samu> than those in the master
15:17:45 <Samu> and those in the master are quite with buggy behaviour in scenario editor
15:18:00 <Samu> so, I'm unsure how to apply the fix
15:19:15 <Samu> besides, there's yet the other underlying issue I opened yesterday, that GS's become automatically active upon loading a save
15:19:39 <Samu> it would also impact what is enabled or disabled
15:20:16 <Samu> this is not fixed on my ai gui
15:21:55 <Samu> something that was supposedly an easy fix, turns out to be complex
15:25:13 <Samu> don't know how to approach this
15:27:47 <Samu> if i change the master rules for the reset button, i'd also have to change the rules applied everywhere else in the gui. This touches many windows
15:28:10 <Samu> for consistency purposes
15:28:54 <Samu> okay, first things first
15:40:31 *** andythenorth has left #openttd
15:41:30 <Samu> oh, that's about the ability to select the GS in the config window
15:41:53 <Samu> yeah, Zuu was on the right direction
15:42:19 <Samu> I went further and made every slot selectable
15:43:24 <supermop_work_> Pikka: might try getting back to northern bits of aus this coming summer/your winter
15:44:10 <Pikka> it might even have cooled down a bit by then, supermop_work_ :D
15:46:03 <supermop_work_> going to a wedding in melbourne in august, so want escape the rain down there for a few days
15:46:14 <supermop_work_> might go up to the pointy bit
15:47:38 *** sla_ro|master has joined #openttd
15:48:28 <Eddi|zuHause> the northern part is where nobody lives?
15:48:52 <supermop_work_> that's the middle
15:49:02 <supermop_work_> people live around the edges
15:49:16 <supermop_work_> trees live in the pointy part i guess
15:49:17 <Eddi|zuHause> yeah, but most of the edge people are in the south?
15:52:49 <Eddi|zuHause> i've seen some video about flat earth theories, and one thing they apparently like to do there is picking out some random connection and taking there being no direct flights on that connection as "proof" that the earth is flat
15:53:06 <Eddi|zuHause> and one of these connections was something like "Johannesburg to Perth"
15:53:54 <Eddi|zuHause> completely ignoring that a) there actually are (rare, but regular) direct connections, and b) demand might not warrant that kind of connection
15:56:32 <supermop_work_> mostly commuting surfers i imagine
15:58:13 *** gelignite has joined #openttd
15:58:31 <Eddi|zuHause> (the point they're trying to make is that for typical flat earth projections, longitudinal distances in the southern hemisphere appear much larger than on "globe" projections, and lack of these direct connections is being seen as "proof" that the real distances are actually longer)
16:16:32 <Samu> if a GS has died, i save the game, and load, should the gs still be dead?
16:16:43 <Samu> it currently makes it alive
16:18:01 <Samu> the information about the instance being dead or alive should be saved imo
16:18:39 <supermop_work_> why would i want it to stay dead?
16:18:53 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7121: Fix #7021: Better revision strings for network and gamelog https://git.io/fhMCX
16:18:55 <Samu> because it was saved as dead
16:19:39 <nielsm> if it's dead it either crashed, or it completed
16:19:46 <nielsm> if it has completed it probably shouldn't restart
16:20:07 <nielsm> if it has crashed it may be in an unrecoverable state and not be safe to restart
16:28:02 <Samu> I really had to write this down
16:28:32 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7121: Fix #7021: Better revision strings for network and gamelog https://git.io/fhM88
16:32:15 <peter1138> Do I need to edit the template?
16:32:30 <peter1138> Because you have not actually stated what the problem is in normal words.
16:32:38 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7121: Fix #7021: Better revision strings for network and gamelog https://git.io/fhM8Q
16:33:04 <Samu> i'm still writting to it
16:33:29 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7121: Fix #7021: Better revision strings for network and gamelog https://git.io/fhM4e
16:33:50 <nielsm> but rushed submitting the ticket to be sure you got that exact issue number?
16:35:29 <nielsm> peter1138, I meant are two modified versions from the same base revision considered equal, i.e. is g12345678M equal to g12345678M (the modifications may be different in each, you can't actually know)
16:35:55 <peter1138> nielsm, ah, well you can't do much about that.
16:37:00 <peter1138> We never attempted with svn versions either :)
16:37:23 <peter1138> Heh, disable all network play with a modified client.
16:37:35 <peter1138> All changes must be committed :D
16:42:43 <nielsm> then, if including just git hash and modified flag, transform the git hash to base64 to fit it in 27 bytes instead of the 40 for hex?
16:42:52 <nielsm> (so it can be used in full instead of truncated)
16:44:01 <Samu> in the scenario editor, only slot 1 is human player, even though i loaded a scenario with more human players
16:44:42 <nielsm> scenarios assume singleplayer
16:45:05 <nielsm> and the scenario editor is a singleplayer game mode
16:47:00 <Samu> i can move ai config slots up and down with ais on it oh, gosh so many issues to report
16:49:17 <Samu> that new issue i created is a tip of an iceberg, a big chunk at least
16:49:32 <Samu> i just fail at explaining...
16:50:45 <nielsm> when you load a saved game that has already been played in normal/network game mode, into the scenario editor, you get tools to violate all kinds of assumptions the game usually makes
16:50:50 <nielsm> so all rules are out the window
16:51:03 <nielsm> and you can't really assume anything about what is sensible or not any more
16:51:55 <nielsm> and the user of the scenario editor may well be able to make changes that break the saved game, but can't be detected as breaking it before loading it back in normal play and letting the simulation run for days, months or years
16:52:44 <nielsm> so trying to prevent someone from doing bad things to savegames loaded in the scenario editor is a futile task imo
16:52:57 <nielsm> and you may as well not attempt at all
16:53:10 *** Wormnest has joined #openttd
16:56:52 <nielsm> also, I'm sure if you attempt to prevent changes to player configuration in a multiplayer game loaded in scenario editor, someone is going to complain that now they can't use their network server to have friends set up a scenario and stuff AIs in it for a singleplayer scenario use
16:57:19 <nielsm> is it guaranteed to work? no. are they allowed to try? yes why not?
16:58:22 <Samu> from my limited understanding, loading a save in scenario editor, also loads saved ai data, which when saved, will be saved exactly as it was loaded
16:58:57 <Samu> changing the settings in the editor might not be a good idea, haven't really attempted to
16:59:21 <Samu> it will not execute a save from the AI script, it's reusing the loaded data
16:59:43 <Samu> it's a "snapshot" of how the AI was at that time
17:00:52 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] closed pull request #6817: PR for Modernizing for C++11: replacing instances of SmallVector https://git.io/vh0VX
17:00:53 <Samu> in a normal game, it won't reuse the just loaded saved data, it's going to request the ai script to execute its save routines
17:01:25 <Samu> it is fine in this case to allow editing the in-game parameters in this situation
17:04:33 <planetmaker> <nielsm> peter1138, I meant are two modified versions from the same base revision considered equal, i.e. is g12345678M equal to g12345678M (the modifications may be different in each, you can't actually know) <-- they might or might not. But we should not care and allow connecting these builds with eachother
17:04:49 <planetmaker> sorry, unintended, double high-light
17:07:19 <Samu> i followed this somewhere into a save stack
17:07:57 <planetmaker> SE is basically like normal game. Except time progression
17:08:13 *** Thedarkb-T60 has joined #openttd
17:10:50 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh updated pull request #7121: Fix #7021: Better revision strings for network and gamelog https://git.io/fh6im
17:12:20 <Samu> Now, the "Reset" button situation will depend heavily on the outcome of how the parameters should be handled
17:12:37 <Samu> that's why it wasn't a simple fix
17:12:39 <nielsm> so now we'll just need a database to look up network revision strings and convert to "which build do I download" :)
17:13:02 <planetmaker> or ... what for do we need it?
17:13:15 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7130: Codechange: Remove value mangling and field misuse in SLE_WRITEBYTE. https://git.io/fhMuG
17:13:26 <planetmaker> it contains the hash, so git will tell us which version to check, doesn't it?
17:13:45 <nielsm> yes, but won't tell a player looking at the server list which version they need
17:14:15 <nielsm> it's great for the programmatic check, not good for displaying in UI
17:14:33 <planetmaker> do we know the displayed version string of the remote?
17:14:46 <planetmaker> as that's what would cut it
17:15:19 <nielsm> I use the _openttd_revision string when IsReleaseVersion() is true, since then it's probably "1.9.0" or whatever
17:15:46 <nielsm> which is probably good enough?
17:16:06 <planetmaker> so all releases have no issue
17:16:15 <planetmaker> yes, that's good enough. At least for now
17:16:27 <planetmaker> if it turns out not good enough, it's time to act then.
17:16:32 <nielsm> yeah it's nightlies and custom branches that are troublesome
17:16:56 <nielsm> I don't know what JGR or other patchpack authors do
17:17:02 <planetmaker> nightlies have the hash in the name, as does the network revision. No issue either
17:17:07 <nielsm> if they set a custom release version string
17:17:25 <nielsm> are nightlies marked as "release" versions?
17:17:26 <planetmaker> JGR seems to set JGRPP-x.yy.z
17:17:32 <planetmaker> no, nightlies are no releases
17:17:47 <nielsm> okay then they just get the opaque git hash as network revision
17:18:30 <planetmaker> hm, ok... could it include the commit's date?
17:18:53 <planetmaker> then it would be solved for nightlies, too
17:19:06 <planetmaker> like yyyymmdd (8byte)
17:19:54 <planetmaker> Honestly not sure whether it needs solving, though
17:20:04 <planetmaker> A nightly server might profit from such, though
17:22:36 <nielsm> uhh... just discovered, determineversion.vbs never sets modified=0 (not modified), only ever 1 (unknown) or 2 (yes modified)
17:26:23 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7121: Fix #7021: Better revision strings for network and gamelog https://git.io/fhMzh
17:28:08 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh approved pull request #7143: Fix #7142: Missing map bounds check when building long roads. https://git.io/fhMgL
17:32:48 <peter1138> You can't tell if it's a nightly or not, sadly.
17:33:06 <peter1138> Maybe include the date to help that? Hmm.
17:33:22 * LordAro mumbles something about just using yyyymmdd for nightlies
17:33:34 <nielsm> or maybe have the web-based server listing have its own lookup table for known common versions
17:33:44 <nielsm> which is automatically updated when a new nightly is released
17:34:11 <peter1138> Then you need lookups too :/
17:46:20 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on issue #7144: Editable AI/GS Parameters in the 3 game modes https://git.io/fhM2p
17:48:22 *** HerzogDeXtEr has joined #openttd
18:08:17 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on issue #7144: Editable AI/GS Parameters in the 3 game modes https://git.io/fhMVS
18:08:39 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN merged pull request #7143: Fix #7142: Missing map bounds check when building long roads. https://git.io/fh174
18:10:01 *** Progman has joined #openttd
18:11:03 <Samu> I'm sorry Peter. it's not easy to make this into just a single issue, it has to have everything else considered
18:11:23 <peter1138> You can reference multiple issues in a single PR.
18:11:42 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN merged pull request #7130: Codechange: Remove value mangling and field misuse in SLE_WRITEBYTE. https://git.io/fhPQ2
18:12:36 <Samu> that would be quite too sparse, considering the number of issues
18:12:56 <peter1138> You can't track them individually then!
18:13:05 <peter1138> Say some get fixes, others don't.
18:13:19 <peter1138> Separate issues makes that much easier to track.
18:13:55 <Samu> but... i need some centralized topic somehow
18:14:04 <peter1138> Use your own notepad then.
18:14:24 <peter1138> Don't dump everything onto github.
18:15:45 <Samu> when I say, I, I mean, everyone needs to see the whole of the problems being dealt with
18:16:38 <nielsm> Samu, or open a thread on tt-forums to discuss it and reach some consensus on what correct behavior should be
18:17:15 <nielsm> because I think that's the first thing to do with scenario editor versus loaded savegames, figure out what should even be considered "correct"
18:17:44 <peter1138> That issue was NOT coherent.
18:17:58 <glx> personnaly I would just removed any player stuff ;)
18:18:02 <peter1138> Just a dump of data and some points after it.
18:18:36 <glx> as a scenario should be an empty map with only towns and industries
18:18:41 <peter1138> nielsm, yeah, that assertion in nrt was actually present in master, just happened to not trigger.
18:18:56 <nielsm> glx ah yeah, convert all roads and channels to public, delete all other transport and delete all companies
18:19:03 <peter1138> nielsm, regression still fails now, but it doesn't crash
18:19:19 <peter1138> I think people like using scenario editor to "touch up" their creations
18:19:24 <glx> but people use SE to hack the savegame
18:19:24 <peter1138> sandbox mode style.
18:19:47 <nielsm> well it might still be useful to have a "clean out players" button somewhere
18:22:47 <peter1138> Thought we did. Hmm.
18:29:11 <DorpsGek_II> [OpenTTD/OpenTTD] flitzpiepe opened pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMov
18:30:27 <peter1138> That's a lot of commits that don't comply with our commit message style :D
18:31:23 <nielsm> and a bunch of merge hmm
18:31:48 <peter1138> As long as it's not merging master...
18:31:59 <planetmaker> if that is all we will have to criticise, that will be good :)
18:32:00 *** Thedarkb-T60 has joined #openttd
18:32:04 <planetmaker> (I haven't looked yet)
18:33:49 <Samu> nielsm there is a clean out players button already
18:34:30 <Samu> removes all companies (but not the gs)
18:40:31 <nielsm> is anyone making a review for the TBTR PR?
18:41:48 *** gelignite has joined #openttd
18:45:58 *** Thedarkb1-T60 has joined #openttd
18:48:28 <peter1138> Format it yourself?
18:48:37 <peter1138> nielsm, way too soon!
18:50:08 <TrueBrain> planetmaker: awesome that you merged in OpenTTD/website; minor request for next time, please check the commit message of people :)
18:50:22 <TrueBrain> personally I don't like the 'enforcing' of the commit message, but I tend to fix it when squashing the commit :)
18:50:37 <TrueBrain> I mean .. "update about.html" ... that is not descriptive :D
18:55:23 <Eddi|zuHause> is that better than dinzip?
18:56:07 <TrueBrain> basically, someone has to ask Azure what's up :D
18:56:09 <peter1138> Trying and failing a way to get 7zip in there. As clearly that is superior.
18:56:15 <planetmaker> TrueBrain, hm... you're right
18:56:46 <TrueBrain> planetmaker: I made it more pretty now :D
18:56:57 <Eddi|zuHause> TrueBrain: no, i need some more basic understanding of which file does what, and how. and possibly a way to test locally
18:57:02 <TrueBrain> I did the horrible thing of force pushing to a public master branch .. people will hate me :P
18:57:17 <TrueBrain> Eddi|zuHause: you mean Dockerfile?
18:57:31 <planetmaker> I was basically like "all checks green. Change looks valid. --> approve & merge"
18:57:49 <TrueBrain> planetmaker: yeah :D I am still on the fence if I should add the commit message checker ..
18:57:53 <TrueBrain> I find it so rude ..
18:58:04 <TrueBrain> to tell someone they failed, while they only made a bad commit message
18:58:55 <Eddi|zuHause> TrueBrain: but why would they write a better commit message next time if you don't tell them?
18:59:14 <planetmaker> TrueBrain, the problem is a bit: we handle repos differently... so it needs different thinking for each :) I agree with your sentiment though
18:59:15 <TrueBrain> Eddi|zuHause: first time contributors? Possibly they are drive-by
18:59:25 <TrueBrain> Eddi|zuHause: so that is only an issue if they make a 2nd or 3rd PR
18:59:39 <Samu> One thing I'd like to be able to change in the scenario editor is the #op codes
18:59:43 <Eddi|zuHause> TrueBrain: or they might drive-by other projects
18:59:46 <TrueBrain> planetmaker: we should make it a single way of doing, I guess :)
19:00:08 <TrueBrain> Eddi|zuHause: pros and cons; often I tell them after the merge: tnx, but your commit messages need improving, so I did that for you
19:00:10 <planetmaker> well, not necessarily. OpenTTD is our main. So extra care there is ok
19:00:11 <TrueBrain> works for me at least
19:00:15 <Samu> but without knowing if it would be problematic, I'm not sure
19:00:30 <planetmaker> And being a bit more relaxed with the rest (at the cost that oversights like mine happen at times)
19:00:49 <TrueBrain> planetmaker: fair point; I agree :)
19:01:23 <Eddi|zuHause> TrueBrain: maybe an inbetween mode "fail" "warning" "ok"?
19:01:34 <TrueBrain> not sure that is possible, but worth a check
19:01:39 <planetmaker> good idea... is that even possible?
19:01:50 <TrueBrain> I guess you can always make a bot write a message if commit message fails
19:01:53 <peter1138> TrueBrain, meanwhile there's a PR #7145 in OpenTTD with 207 commits. All of them fail commit message check :D
19:02:06 <Samu> where does this gist have any visibility?
19:02:19 <peter1138> I suspect they need to squash it a bit.
19:02:20 <planetmaker> hehe, yeah. And... I wonder how to review such gigantic queue
19:02:25 <TrueBrain> peter1138: holy crap
19:03:01 <LordAro> (i am joking, of course)
19:03:11 <peter1138> Apparently it's a reimplementation, so I'm not sure why it includes fixes.
19:03:29 <TrueBrain> just a working banch
19:03:47 <planetmaker> I wonder about that, too. It looks like organically grown over a year or more. And everything stacked... and not commits amended and squashed
19:03:49 <TrueBrain> that needs a bit love and attention :)
19:03:55 <planetmaker> so the real stuff is very hard to see
19:04:13 <peter1138> Yeah, it's what I do when I'm working on stuff.
19:04:13 <TrueBrain> which reminds me, I am going to try to create binaries for NRT this weekend or so
19:04:17 <TrueBrain> see if we can do that now :)
19:04:22 <peter1138> I then squash small commits together, though...
19:04:35 <peter1138> TrueBrain, it fails regression testing.
19:04:47 <peter1138> TrueBrain, and the PR is out of date, my branch is current.
19:04:48 <TrueBrain> peter1138: owh .. so I will wait till that is fixed :D
19:04:50 <nielsm> This is becoming a very long review for code style...
19:05:11 <peter1138> Apparently I can push to andy's branch but I need to check with him.
19:05:17 <TrueBrain> nielsm: what we used to do, was just ask them to look at the coding style, and try to apply it, without giving specifics
19:05:17 <peter1138> There's some bugs too ;)
19:05:27 <planetmaker> Not sure it's worth individual review on code style... more exemplary on that
19:05:35 <nielsm> how many of us are looking over it right now? :P
19:05:45 <TrueBrain> *launches Minecraft*
19:05:56 <Wolf01> Mmmh, what is steam doing?
19:06:04 <Samu> gist doesn't parse images? what the heck is gist for
19:06:09 <planetmaker> I looked at 207 commits with no apparent order and useful descriptions and though "hell, that needs clean-up before review"
19:06:19 <Wolf01> Almost half of my installed games are updating
19:06:27 <TrueBrain> so tell him exactly that, in nice and friendly words :D
19:06:27 <peter1138> Wolf01, updating everything with 0 byte updates.
19:06:33 <peter1138> Means they changed a load of metadata, I think.
19:06:41 <planetmaker> it's like reading the draft-draft-draft of some student's thesis which assume one has infinite time to go through unreflected collection of thoughts
19:08:53 <nielsm> okay I have a review with 36 individual comments...
19:09:36 <Samu> gist is not what i was looking for
19:09:51 <TrueBrain> so that is the "in between" I guess :P
19:10:02 <Eddi|zuHause> i'm trying to write this message: "
19:10:04 <Eddi|zuHause> Thank you for your patch, but this is completely unreviewable in the current state.
19:10:05 <Eddi|zuHause> Please consider grouping commits into more logical units, squashing a few minor edits, and adhere to the commit message style as explained in contributing.md"
19:10:14 <glx> peter1138, Wolf0: directx install package I heard
19:10:24 <Eddi|zuHause> anyone objecting?
19:10:39 <TrueBrain> Eddi|zuHause: minor rewording:
19:10:54 <TrueBrain> "Thank you for your patch! Before we start to review, we have a few requests for you"
19:11:02 <Samu> into more logical units, more commits!
19:11:07 <TrueBrain> followed by your second line
19:11:26 <TrueBrain> "completely unreviewable" and "but" are kinda negative :)
19:11:38 <peter1138> Might need reminding that push -f is okay :)
19:11:39 <TrueBrain> and link to the CONTRIBUTING.md :)
19:11:54 <TrueBrain> try to guide him to the right path :)
19:12:27 <Eddi|zuHause> how do i put a link to contributing in?
19:12:49 <TrueBrain> Markdown yo! :) Possibly the interface also has a "link" button
19:13:16 <TrueBrain> I am so used to Markdown these days :D
19:13:40 <peter1138> Dinrar completed, food successfully, er, compressed.
19:13:46 <planetmaker> I like TB's suggestion for the reply
19:14:13 <planetmaker> and peter's hint on push -f and link to contributing.md
19:15:17 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z commented on pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMPi
19:17:01 <glx> hmm why mixer.cpp contains mixed EOL ?
19:17:10 <TrueBrain> because someone FAILED :D
19:17:41 <glx> and that's in the git repo
19:18:06 <glx> because if I fix it git diff shows the changed lines
19:18:18 <glx> instead of saying no diff and eol diffs
19:18:18 <TrueBrain> Eddi|zuHause: nice! :)
19:18:21 <peter1138> It's nielsm's changes, he had CRLF issues until just yesterday :)
19:18:50 <peter1138> Not sure how they were allowed in, mind you.
19:18:54 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMPh
19:19:21 <Samu> my communication skills... arg... i hate myself at times :!
19:19:30 <Eddi|zuHause> peter1138: commit-checker could have crlf checks?
19:19:37 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMXf
19:21:22 <Samu> so, 1 issue per issue, i'm sure this is not going anywhere
19:21:48 <peter1138> show us a gist before posting one
19:21:58 <peter1138> i'm sorry but we need to review your issues :p
19:22:54 <peter1138> I mean a gist of the issue you wish to post.
19:23:18 <Samu> - AI Config window in SE: loading a scenario of a savegame with multiple human companies,
19:23:18 <Samu> only one human company is displayed as being human. The other human companies can be selected
19:23:18 <Samu> and editable. It's not what happens on a NG (normal game).
19:24:08 <peter1138> And what's the point of using your own abbreviations when you need to explain them?
19:26:11 <Samu> i can already see it going the opposite direction, you're now gonna fix it by disabling it in the editor
19:26:16 <peter1138> Samu, please follow the template.
19:28:50 *** andythenorth has joined #openttd
19:29:36 <andythenorth> anyone actually tried TBTR?
19:29:37 <Samu> the expected result is also wrong~
19:29:57 <andythenorth> my first thought is that the GUI doesn't even fit in my screen
19:30:34 <nielsm> sunken separator bars look out of place
19:30:39 <andythenorth> I have an entirely counter proposal to templates
19:30:42 <andythenorth> I wanted them for years
19:31:06 <andythenorth> if I have 50 trains to change consist on
19:31:18 <andythenorth> I make one new one, send all the old ones to depot, and clone the new one 49 times
19:31:36 <planetmaker> samu, what is a Hu company?
19:32:03 <TrueBrain> saving 3 letters, SMS style FTW!
19:32:22 <TrueBrain> 1990 called .. they want their letters back
19:32:30 <TrueBrain> T9 texted me, it is okay to use the keyboard
19:32:38 <planetmaker> Hu company reads to me like Hungarian company :)
19:32:46 <TrueBrain> even with swiping you can talk binary all day
19:32:52 <peter1138> andythenorth, can I push my nrt directly to yours?
19:33:23 <Samu> the expected result is not what I'd aim for a fix
19:33:40 <Samu> because, looking at the whole... it's not the better way
19:33:53 <andythenorth> peter1138: I give you rights, just a sec
19:33:56 <planetmaker> Samu, and why do you then call it "expected result" when you don't expect / want it?
19:34:27 <Samu> because that's what everyone would expect atm
19:34:30 <Eddi|zuHause> andythenorth: ok, that GUI is in dire need of a cleanup
19:34:44 <andythenorth> peter1138: invited you to my repo
19:34:58 <peter1138> Samu, write what *you* expected to happen when you did it.
19:35:06 <andythenorth> so one counter for templates would be 'clone n' which has been requested multiple times, but eh
19:35:16 <planetmaker> samu, you should indeed make-up your mind what is a good overall way. What you expect a good interface to do and look like
19:35:21 <andythenorth> another is 'auto-replace n' instead of 'autoreplace 1'
19:35:23 <glx> Samu: only network savegames can have more than one human company
19:35:24 <planetmaker> and compare that to the current one
19:35:27 <Samu> i already made up my mind
19:35:39 <andythenorth> rule based autoreplace is probably not very intuitive, but likely very powerful
19:35:39 <Samu> but you want me to slice it into minor prs
19:35:46 <planetmaker> glx, yes-ish... but what is when you load a network save in single-player? Easy to do
19:35:59 <Samu> they aren't sensible alone, it's part of a whole that they only start making sense
19:36:02 <planetmaker> The difference betwwen "network save" and "single-player save" does not exist
19:36:13 <glx> and SE was never supposed to work well with companies on it
19:36:39 <planetmaker> Indeed it makes IMHO sense to mark company slots as "AI (AI-name)" and others as "human" - irrespective of SP or MP game
19:36:49 <glx> I know planetmaker, that's why SE assumes human as company 0
19:36:56 <andythenorth> so when's the 1.9.0-RC1?
19:37:00 <planetmaker> though OpenTTD need not offer and facilities to add other human companies
19:37:05 <Samu> i have my stuff done, pretty much, it's just that I now have to turn it into smaller parts, it's not easy to explain it, to expose it this way
19:37:06 <andythenorth> clock is ticking, I think we'll have bugs arising
19:37:14 <peter1138> andythenorth, merge nrt! do it!
19:37:22 <TrueBrain> lets do NRT after 1.9 :D
19:37:27 <peter1138> Ignore the regression test!
19:37:31 <andythenorth> NRT is blatantly 2.0
19:37:32 <TrueBrain> andythenorth: is the nightly play-tested you think?
19:37:40 <andythenorth> TrueBrain: already it has found bugs
19:37:45 <Eddi|zuHause> planetmaker: maybe i wouldn't call it "human" but "player"?
19:37:45 <andythenorth> how many downloads?
19:37:57 <TrueBrain> will check about that this weekend
19:38:12 <planetmaker> Eddi|zuHause, I don't mind the semantics in that case. Either way is fine for me
19:38:31 <planetmaker> Honestly it's not a bad thing, if OpenTTD would allow also in SP different player companies
19:38:40 <planetmaker> it's not like you cannot move between them anyway
19:38:43 <Eddi|zuHause> "Company Name 1 (Player)" "Company Name 3 (AI)"
19:39:42 <Eddi|zuHause> planetmaker: well, all the interfaces for that are already there, just guarded behind "is multiplayer"
19:39:46 <glx> isn't there an isai mark in companies ?
19:40:00 <Eddi|zuHause> planetmaker: i once played alone on a local server just to have multiple companies
19:40:03 <peter1138> Samu, so your "expected/actual" results are ... completely unrelated to your "steps to reproduce"
19:40:04 <planetmaker> I think there is @ glx
19:40:18 <planetmaker> Eddi|zuHause, exactly. Me, too. And we're certainly not the only ones
19:40:28 <peter1138> Samu, so I still don't know what the actual issue.
19:40:31 <glx> the SE probably ignore it
19:40:38 <peter1138> I can probably guess, but we should not need to guess.
19:41:19 <peter1138> (How do you load a savegame in the scenario editor, anyway? :p)
19:41:40 <peter1138> So it's not even supported anyway.
19:42:08 <Samu> that's how i learn about it too
19:42:23 <glx> it's not recommended to use it that way
19:42:27 <LordAro> "supported" is not the same as "it works"
19:42:46 <peter1138> Anyway, can you try rephrasing your gist to explain what issue you are talking about?
19:43:15 <peter1138> Maybe the conversation in here could help you do that (but don't just paste the conversation)
19:45:02 <glx> indeed there's is_ai in Company
19:45:24 <planetmaker> you can still move yourself to an AI company
19:45:31 <planetmaker> doesn't exactly make it supported either
19:46:09 <glx> true, but as I understand it SE doesn't check is_ai
19:46:35 <planetmaker> why should it check or distinguish companies?
19:46:54 <glx> I mean ai config window in SE
19:47:49 <peter1138> is the AI config window different in SE?
19:48:05 <peter1138> And is this samu's issue?
19:48:25 <glx> same as intro and game, with different rules inside I think
19:48:43 <glx> and I guess SE is identical to intro
19:48:47 <Samu> the AI config window needs different rules in all 3 game modes
19:49:21 <peter1138> presumably it can be more lax in the scenario editor. it is an editor, after all.
19:49:35 <glx> considering human in slot 0 and IA in other slots is very intro to me
19:50:20 <Samu> you enter scenario editor, and slot 1, or 0 as you say, has no company
19:50:39 <Samu> it's not equal to main menu
19:51:54 <glx> yes because it ignores the saved companies
19:53:36 <Samu> what shall my gist be written like? sorry for asking
19:56:32 <Samu> the expected result is...
19:56:37 <Samu> - human companies should not be selectable
19:56:37 <Samu> - every slot should be selectable and editable
19:57:04 <Samu> my suggestion would be the 2nd line
19:57:41 <Samu> but what it's currently assumed is that human companies should not be selectable
19:57:50 *** Thedarkb1-T60 has joined #openttd
19:58:43 <Samu> in a normal game it's not selectable
19:58:46 <glx> for me behaviour should match running game regarding selectable, but intro regarding editable settings
19:59:21 *** supermop_work has joined #openttd
19:59:23 <glx> because it's an existing game, but it's not running
19:59:39 <Samu> human company could also be reseted in multiplayer
20:00:06 <Samu> the slot could be editable, regardless if it's currently occupied by a human
20:00:38 *** supermop_work__ has joined #openttd
20:00:43 <Samu> and move down, move up the config into a non occupied slot for commodity
20:02:37 *** supermop_work_ has quit IRC
20:07:36 <Samu> I can't do this.. i'm stuck
20:08:08 <Samu> this = split into smaller commits
20:12:44 <LordAro> Samu: you're good at testing pathfinders, feel like testing #7114 ?
20:15:12 <LordAro> peter1138's right though, the "overhaul" commit does need splitting up
20:15:21 <LordAro> (and the tt-f link removing from the commit message)
20:17:19 <Samu> ah, that's a fix to something i reported, cool, gonna test
20:34:04 <DorpsGek_II> [OpenTTD/OpenTTD] flitzpiepe commented on pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMH1
20:34:51 <DorpsGek_II> [OpenTTD/OpenTTD] flitzpiepe commented on pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMHy
20:39:29 <nielsm> hmm, not sure how much to put in those words, but the comment about "much more stable" for TBTR feels omnious to me, "it doesn't crash quite as often"
20:39:38 <nielsm> (meaning it still crashes occasionally)
20:43:44 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z commented on pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMQl
20:43:55 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMQ8
20:44:49 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMQg
20:46:43 <LordAro> not a huge fan of the TBTR acronym either
20:47:36 <Eddi|zuHause> a (simplified) version of that might also be useful for other vehicle types
20:48:03 <Eddi|zuHause> but maybe shelve that idea until we cleaned it up :p
20:48:32 <nielsm> looks like it's a swedish dude, by this commit message "normalized the {} for some case:s"
20:48:51 <nielsm> I don't know anyone but swedes using : instead of apostrophe
20:57:13 <nielsm> any chance for another review of fix-networkrevision? :}
20:59:06 <nielsm> ...I suppose the increase in NETWORK_REVISION_LENGTH may be moot now the network revision is only git hash?
20:59:10 <Samu> LordAro tried to trigger the asserts but couldn't
20:59:18 <Samu> it's always chosing a track
20:59:32 <LordAro> sounds like it's working then :)
20:59:44 <LordAro> try other vehicle types as well, to make sure they don't do unexpected things
20:59:53 <Samu> tried ship, road vehicle and train
21:00:19 <LordAro> comment as such on the PR, if you would be so kind :)
21:01:22 <Samu> it was only a ship issue
21:01:39 <Samu> tested the same manner with road vehicle and train, it was behaving fine
21:05:35 <Samu> wanna see what the code is doing different, brb
21:07:36 <Samu> if it ignores the first tile, but there is no tile ahead, what does it do then?
21:07:48 <Samu> it's chosing a track still
21:16:53 <Samu> interesting, this change may affect river generation
21:17:03 <Samu> maybe that's what he was refering to, with thoroughly
21:17:48 <Samu> have to check river gen code again
21:18:03 <Samu> i don't think a river can start and end in the same spot
21:23:10 <Samu> yep, won't cause any difference to river generation
21:24:29 <Samu> it always require a minimum of moving a tile
21:24:45 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #7114: NPF bails out when source tile is also a destination tile https://git.io/fhMdR
21:36:27 *** tokai|noir has joined #openttd
21:36:27 *** ChanServ sets mode: +v tokai|noir
21:52:03 *** Thedarkb1-T60 has joined #openttd
21:56:02 <DorpsGek_II> [OpenTTD/OpenTTD] flitzpiepe commented on pull request #7145: TBTR 2.0 (Template-based Train replacement) https://git.io/fhMNe
21:58:19 *** Thedarkb-T60 has joined #openttd
22:31:42 *** andythenorth has joined #openttd
22:35:01 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 opened pull request #7146: Fix: don't show OS error box for non GUI video drivers https://git.io/fhMp3
22:38:52 <DorpsGek_II> [OpenTTD/OpenTTD] michicc approved pull request #7146: Fix: don't show OS error box for non GUI video drivers https://git.io/fhMpo
22:45:12 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 merged pull request #7146: Fix: don't show OS error box for non GUI video drivers https://git.io/fhMp3
22:47:37 <Samu> jesus, even the simple change of increasing max no competitors to 15 requires deep changes to avoid crashes
22:48:24 <Samu> by allowing 15, move down needs changes
22:48:57 <Samu> bah, i'm working backwards
22:49:36 <Samu> i already have this done
22:50:01 <Samu> i can't just simply split this into small pr's
22:54:12 <LordAro> Samu: almost as if there's an awful lot of stuff that assumes 14
22:54:25 <LordAro> refactors first, i.e. fixing all the things that assume 14
22:54:35 <LordAro> then add individual "features" one at a time
22:54:56 <andythenorth> I don't want to be discouraging, but does TMWFTLB apply here?
22:55:05 <andythenorth> 14 -> 15 is not much gain for a lot of effort
22:55:26 <Samu> it is already done, the entire work
22:55:35 <andythenorth> if it's a large changeset, we're not likely to accept it
22:55:35 <Samu> now i'm trying to split it apart
22:55:45 <Samu> that's where I'm failing
22:55:49 <andythenorth> the risk of side effects vs. the benefit, the review time doesn't pay off
22:56:18 <andythenorth> there's an implicit calculation that is not often rationalised, but summed up by TMWFTLB
22:56:48 <andythenorth> 'but this one goes to 11'
22:56:51 <Samu> yeah, you guys even get confused with my ramblings
22:56:53 * andythenorth should go to bed
22:57:13 <andythenorth> a big change introduces risk
22:57:25 <andythenorth> the risk isn't worth it for a tiny benefit
22:57:42 <andythenorth> risk means reverts, or hot fixes, and having to read bug reports
22:57:50 <andythenorth> risk is a party pooper
22:58:23 <Samu> I'm failing to tear my big PR into small parts, as requested
22:58:42 <andythenorth> big UI changes are very very hard to split across multiple PRs
22:58:54 <andythenorth> they tend not to make sense in fragments
22:58:56 <Samu> the PR itself is finalized
22:59:00 <glx> and 14->15 should not have been part of the GUI work
22:59:24 <Samu> it's deeply integrated into it
22:59:32 <andythenorth> that's a bad sign :)
22:59:36 <LordAro> well that's part of the issue, it shouldn't be
22:59:50 <Samu> the move down button is fine for a max of 14
22:59:51 <LordAro> 14->15 should be a separate commit(s)/PR, GUI changes should follow from that
22:59:57 <andythenorth> I write a lot of big UI changes, and I know how they end up integrated with other changes, but they shouldn't
23:01:11 <Samu> 15 companies is the real max
23:01:17 <andythenorth> or to ask another way, why is it 14 currently?
23:01:25 <LordAro> Samu: ok, first step, is to make the GUI not care about how many AIs there are
23:01:27 <Samu> and there were no real issues with manually starting 15 ais
23:01:45 <andythenorth> do we need a gui to start 15?
23:01:55 <glx> the limitation was due to available company colours
23:01:56 <Samu> yes, if you move yourself to spectator
23:02:29 <andythenorth> no but we could add more company colours :P. There are a few ranges unexploited
23:02:40 <andythenorth> *that* would be nice
23:02:53 <Samu> you can already create 15 companies
23:03:44 <andythenorth> how many AIs can you create?
23:04:15 <milek7> btw. how it was handled in 240 companies patch? or it wasn't handled properly?
23:04:32 <Samu> i didn't try my gui with your patch
23:04:44 <Samu> it would fail, i removed the scrollbar...
23:05:25 <andythenorth> what were we discussing again? o_O
23:05:41 <Samu> the simple fact that changing 14 to 15 requires gui changes
23:06:11 <andythenorth> you say above we can create 15 already
23:06:35 * andythenorth is thoroughly confused and will stop asking
23:06:39 <Samu> the gui imposes a limit of 14
23:06:45 <Samu> but the real limit is 15
23:06:48 <andythenorth> how do you create 15 manually?
23:06:57 <andythenorth> manual = gui, no?
23:07:56 <andythenorth> ok so in summary: you're fixing a perceived defect with the AI GUI, which should support 15 AIs, but only supports 14?
23:08:13 <andythenorth> anyone know why it only supports 14?
23:08:18 <andythenorth> mistake, or design?
23:08:27 <Samu> because it was probably designed for single player in mind
23:08:31 <glx> Samu: your change in src/table/settings.ini is wrong btw
23:08:34 <milek7> or rather, allowing to spawn ai in place reserved for player normally
23:09:58 <andythenorth> is company colour a nibble?
23:10:04 <glx> because companies range from index 0 to index 14, and that's 15 companies
23:10:20 <Samu> MAX_COMPANIES = 15 already
23:11:11 <andythenorth> so...more company colours then?
23:11:17 <Samu> it's used like < MAX_COMPANIES everywhere i see it
23:11:34 <glx> MAX_COMPANIES is not a valid company number
23:11:34 <LordAro> Samu: arrays start from 0
23:11:43 * andythenorth would like black, there's a grf for that
23:11:51 <andythenorth> company colour grfs!
23:11:56 <andythenorth> just recolour sprites :P
23:12:08 <glx> MAX_COMPANIES = 0x0F, ///< Maximum number of companies
23:12:08 <glx> OWNER_TOWN = 0x0F, ///< A town owns the tile, or a town is expanding
23:12:10 <peter1138> There's the RGB company colours patch ;)
23:12:15 <Samu> but MAX_COMPANIES - 1 is what defines it to be 14
23:12:22 <andythenorth> I even saw a picture once
23:12:25 <peter1138> You can have BLACK with that.
23:12:30 <andythenorth> company colour grfs, with date callbacks
23:12:39 <andythenorth> so you encode the full history of British Rail liveries
23:13:07 <andythenorth> also 3rd company colour, it's only a few more bits
23:13:16 <glx> you can't have company 15
23:13:21 <Samu> that setting is used for max_no_companies
23:13:51 <andythenorth> isn't company 15 town?
23:14:05 <Samu> but for counting the num of companies
23:14:08 <andythenorth> didn't I run into company null or something once, for dead companies?
23:14:08 <LordAro> i'm pretty sure we don't want to allow people to start 15 AIs normally
23:14:22 <andythenorth> somewhere in NRT, I'm sure I found there's a dead company value for infra
23:14:26 <glx> but player + max_no_companies = 15 companies
23:14:57 <Samu> ah, you refer to the current way ai_gui is counting them
23:15:11 <glx> that's how the game counts it
23:15:22 <Samu> the changes I've already went through
23:15:54 <glx> would be better to allow company 0 to be an AI without touching this setting
23:16:34 <glx> like you set 14 max competitors + an IA as company 0
23:17:50 <glx> because max_no_companies should still be an addition to the company 0
23:18:26 <Samu> i made the necessary changes regarding that
23:18:31 <peter1138> andythenorth, there's probably a patch for that...
23:18:50 <glx> so when you set it to 15 no Human company can't exist
23:19:12 <glx> but that totally changes the meaning of the setting
23:19:28 <glx> it's a difficulty setting for real players
23:19:38 <Samu> you can still start 15 ais in single player
23:19:58 <glx> as I said better to add a option to replace human company by an AI
23:20:02 <Samu> requires moving to an AI company, via cheat screen, and letting first company bankrupt
23:20:56 <glx> not if it's set before the game start and handled correctly by the engine
23:21:19 <glx> dedicated server has no issue to start a game without players
23:21:35 <glx> you just want a dedicated server with GUI and 15 AIs
23:22:11 <Samu> yes, but it's already possible via openttd interface as well
23:22:33 * andythenorth is thoroughly confused
23:22:43 <andythenorth> this is a lot of work to add 1 more AI to the UI
23:23:41 <andythenorth> so if I start 15 AIs, how do I play?
23:23:44 <Samu> single player could have a spectator slot
23:24:03 <andythenorth> ok so one of my kids plays with 14 AIs
23:24:12 <andythenorth> he sets it to 15
23:24:15 <andythenorth> and starts a new game
23:24:20 <glx> everything already handles correctly the 15 companies, you really just need some changes to start as spectator and put an AI in slot 0
23:24:53 <Samu> ais already can start in slot 0
23:24:58 <andythenorth> so if I start a game with 15 AIs, I'm a spectator?
23:25:41 <Samu> the ais don't start right
23:25:52 <Samu> it'd be preferible to start a game as a spectator
23:25:53 <glx> yes, but I mean for a non dedicated game, you just need an option so the "player" starts as spectator instead of creating company 0, then you're free to put an AI there
23:26:10 <andythenorth> so we change the game start for all single player?
23:26:18 <andythenorth> you start as spectator, then have to join a company?
23:26:39 <glx> it's a special case to get 15 AIs running
23:27:19 <Samu> gonna see if i can do it
23:27:58 <andythenorth> I'm trying to imagine the pitch
23:28:08 <andythenorth> "Feature: we made it possible for you to spectate an all-AI game"
23:28:27 <peter1138> andythenorth, no "AI Settings Overhaul"
23:28:29 <glx> sure some people would like it
23:28:30 <andythenorth> "Feature: when starting single-player, you will now always start as a spectator, and you will have to join a company"
23:28:32 <Samu> could bring back ai competitions
23:28:33 <peter1138> That totally explains what it does
23:28:51 <andythenorth> "Note: if you don't join a company before the AIs start, you will not be able to play"
23:29:07 * andythenorth must be missing the point
23:29:22 <Samu> well yeah, i added a Stop AI button too
23:29:29 <peter1138> andythenorth, I love that we've spent weeks discussing this ;)
23:29:41 <andythenorth> to add 1 extra AI to the UI
23:29:53 <glx> to finally understand the goal :)
23:29:56 <andythenorth> to support the need of the 1 person in the world who wants to do it
23:29:59 <andythenorth> who can already do it
23:30:05 <andythenorth> but now wants a UI for it
23:30:30 <andythenorth> to be really clear, and not mean, I think samu's goal is fine for Samu
23:30:35 <andythenorth> I just don't think it's shared
23:30:53 <andythenorth> depot ctrl-click counter for vehicles anyone? :P
23:31:19 <glx> I'm quite sure AI devs would like to see their AI running without having to check the money for company 0
23:31:41 <andythenorth> I'm quite sure as newgrf dev, I'd like to always start with £10m money
23:31:47 <andythenorth> and no town ratings
23:31:58 <andythenorth> and industry production cheat enabled
23:32:02 <glx> it's possible, use a game script
23:32:03 <andythenorth> and magic bulldozer
23:32:24 <andythenorth> newgrf deity mode
23:33:00 <glx> anyway it should be an option, not a change affecting all other players
23:33:30 * andythenorth wonders about newsandboxgame
23:33:52 <andythenorth> which ships with a bunch of cheats enabled
23:34:00 <andythenorth> for content development
23:34:33 <andythenorth> possibly even just loads a different config
23:35:04 <Samu> anyway, the gui could use some improvements
23:35:28 <Samu> because setting up scripts in gui is much more natural
23:35:37 <Samu> than doign in console way
23:35:49 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7121: Fix #7021: Better revision strings for network and gamelog https://git.io/fhDfD
23:36:25 <Samu> speaking about when a game is already running
23:38:55 <Samu> okay, i'm alone on this :|
23:39:35 <LordAro> Samu: not saying your gui improvements aren't good, just don't want the 14 -> 15 change :)
23:39:47 <peter1138> andythenorth, i found my stash containing related vehicles changes
23:39:56 <peter1138> (read: doesn't do anything yet)
23:40:44 <andythenorth> few problems with that drawing, but eh
23:41:02 <andythenorth> synchronising intro dates was separate
23:41:18 <peter1138> Yeah, cos that's useful for other things.
23:42:04 <andythenorth> Eddi|zuHause had a proposal for that using the game seed, can't remember what
23:43:09 <Eddi|zuHause> yeah, each time intro date variation is called, init with game seed and date property, so you get reproducible randomisation
23:43:32 <Eddi|zuHause> every vehicle set to enter at the same date will stay at same date, just moved randomly
23:44:14 <Eddi|zuHause> all vehicles with the same introdate will be moved "en bloc"
23:44:38 <peter1138> uint32 r = Random();
23:45:01 <peter1138> if (somepropertysetonengine) uint32 r = Random(gameseed);
23:45:04 <milek7> how likely it is to get merged? ;d
23:45:27 <peter1138> Grey grass? Never!"
23:45:40 <Eddi|zuHause> peter1138: i'd just do it for all existing sets, without new property
23:45:59 <peter1138> Eddi|zuHause, that breaks them because the spec's been changed.
23:46:17 <Eddi|zuHause> peter1138: where in the spec does it specify how the randomness is calculated?
23:46:48 <peter1138> Guaranteed random number.
23:48:02 <Samu> looks like I'm not entirely alone
23:48:02 <Eddi|zuHause> peter1138: intro date has weird corner case specs anyway
23:48:09 <Eddi|zuHause> peter1138: i don't think there's much harm
23:49:16 <andythenorth> it's depot flip all over again!
23:49:22 * andythenorth doesn't much care prop or no prop
23:49:36 * peter1138 flips andythenorth's depots.
23:49:40 <andythenorth> I think it would be interesting to have a separate prop to specify random range per vehicle :P
23:49:40 <Eddi|zuHause> peter1138: but i guess we could add a flag somewhere
23:49:53 <andythenorth> I would like to make some jokers
23:50:09 <Samu> i used to watch 4 ais vs 4 ais in starcraft, warcraft... i don't think i'm the only one in the world that enjoys AIs
23:50:49 <Eddi|zuHause> train prop 27 (misc flags) seems to be already full
23:50:59 <Samu> even made some ai scripts for war3
23:51:36 <Samu> it would always let me use all slots for AIs
23:51:49 <Samu> why must openttd be different
23:53:14 <andythenorth> historical reasons
23:53:36 <Samu> if i recall, even brood war now lets 8 ais start, there's spectator slots
23:54:22 <andythenorth> so we're still winning ;)
23:54:41 <Samu> + some 255 spectator slots or so
23:55:20 <andythenorth> eh what " full git model with branches being merged instead of rebased "
23:55:28 <Samu> ok, suppose I change it back to 14
23:55:32 <andythenorth> I thought rebase was the bees knees?
23:55:46 <peter1138> It is the bees knees.
23:55:48 <glx> it's just slot 0 is currently reserved to human
23:55:54 <Samu> i would still have to manually start the 15th ai :(
23:56:02 <peter1138> But it's not how git is meant to work :D
23:56:12 <andythenorth> I thought I was doing it wrong all these years
23:56:23 <andythenorth> feature branch -> milestone branch -> master
23:56:29 <andythenorth> merges all the way
23:57:11 <Eddi|zuHause> it depends a bit
23:57:12 <andythenorth> but tbh, rebase makes complete sense for working with github and 'everything is a fork'
23:57:13 <LordAro> Samu: last time i checked, AIs don't start all at once anyway, so if you're doing any half decent AI battle, you're going to want to start them all manually at the start anyway
23:57:39 <andythenorth> rebasing on upstream seems a lot cleaner than merging
23:58:31 <andythenorth> 'New Game', 'Play Scenario', 'Watch AI Game'
23:58:43 <Samu> hmm, making ais start right away
23:58:43 <peter1138> LordAro, it's probably all in the AI settings overhaul patch.
23:58:53 <peter1138> LordAro, as in... the one that doesn't explain anything :p
23:58:53 <andythenorth> also 'New newgrf dev deity mode game'
23:58:55 <Samu> that could be a nice idea
23:59:18 <andythenorth> how big are the knees on bees?
23:59:20 <peter1138> Hmm, I should have a shower and then get dressed into pyjamas.
23:59:22 <Samu> min start_date is 1 atm, i could make it 0
23:59:25 <andythenorth> I was going to shower
23:59:29 <andythenorth> I just went climbing
23:59:34 <Samu> and that would make them start immediately
23:59:36 <andythenorth> so I smell of chalk and natural deodorant
23:59:44 <andythenorth> but I cba and am getting in bed
continue to next day ⏵
