IRC logs for #openttd on OFTC at 2021-06-13
⏴ go to previous day
00:03:59 *** outtatime has quit IRC (Remote host closed the connection)
00:15:07 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
00:48:29 *** Wormnest has quit IRC (Quit: Leaving)
00:58:01 *** Progman has quit IRC (Remote host closed the connection)
01:20:02 *** ChanServ sets mode: +v tokai
01:25:30 *** gelignite has quit IRC (Quit: Stay safe!)
01:26:43 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
01:29:53 *** tokai|noir has joined #openttd
01:29:53 *** ChanServ sets mode: +v tokai|noir
01:36:26 *** tokai has quit IRC (Ping timeout: 480 seconds)
02:44:34 *** ChanServ sets mode: +v tokai
02:51:36 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
03:22:26 *** tokai|noir has joined #openttd
03:22:26 *** ChanServ sets mode: +v tokai|noir
03:29:26 *** tokai has quit IRC (Ping timeout: 480 seconds)
03:30:20 *** ChanServ sets mode: +v tokai
03:37:14 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
03:55:19 *** snail_UES_ has joined #openttd
05:07:50 <DorpsGek> [OpenTTD/OpenTTD] telk5093 opened issue #9362: [UI] Background hover colour slightly overlaps the end of client list https://git.io/JZ5Tb
05:14:35 *** roadt__ has joined #openttd
05:21:40 *** roadt_ has quit IRC (Ping timeout: 480 seconds)
05:30:01 <DorpsGek> [OpenTTD/OpenTTD] telk5093 opened issue #9363: [UI] Online players window's buttons is not resized when use double/quad interface size https://git.io/JZ5sO
05:39:13 <DorpsGek> [OpenTTD/OpenTTD] telk5093 opened pull request #9364: Fix #9362: Background hover colour slightly bigger than entry in the online players window https://git.io/JZ5Cc
06:05:00 *** Ttech has quit IRC (Quit: Este é o fim.)
06:27:23 *** andythenorth has joined #openttd
06:58:34 <andythenorth> so I could do a 3rd company colour as grf parameter eh
07:13:37 <peter1138> Do you want to provide all 4096 remaps?
07:15:16 *** Progman has joined #openttd
07:25:49 *** sla_ro|master has joined #openttd
07:26:46 <andythenorth> but it's just a generator...
07:36:15 *** tokai|noir has joined #openttd
07:36:15 *** ChanServ sets mode: +v tokai|noir
07:43:18 *** tokai has quit IRC (Ping timeout: 480 seconds)
08:04:13 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9364: Fix #9362: Background hover colour slightly bigger than entry in the online players window https://git.io/JZd0C
08:05:37 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9364: Fix #9362: Background hover colour slightly bigger than entry in the online players window https://git.io/JZdEv
08:06:54 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain closed issue #9362: [UI] Background hover colour slightly overlaps the bottom line of client list window https://git.io/JZ5Tb
08:06:57 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #9364: Fix #9362: Background hover colour slightly bigger than entry in the online players window https://git.io/JZ5Cc
08:07:35 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9360: Codechange: [Actions] Improve MSYS2 setup time https://git.io/JZduf
08:16:48 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9314: Codechange: let the setting name be std::string https://git.io/JZdaZ
08:17:20 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 opened pull request #9365: Fix #9361, a2051ba: [Network] Off by one in CanWriteToPacket https://git.io/JZdaw
08:20:15 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9365: Fix #9361, a2051ba: [Network] Off by one in CanWriteToPacket https://git.io/JZdws
08:24:35 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9314: Codechange: let the setting name be std::string https://git.io/JZdoG
08:27:02 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 merged pull request #9314: Codechange: let the setting name be std::string https://git.io/JGCaz
08:31:09 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 merged pull request #9365: Fix #9361, a2051ba: [Network] Off by one in CanWriteToPacket https://git.io/JZdaw
08:36:14 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 updated pull request #9359: Consolidate IConsolePrint functions and improve some of the messages https://git.io/JZ9ZR
08:39:51 <_dp_> Rubidium, why remove IConsoleError and co? looked like a nice shortcut to me
08:40:00 <_dp_> and formatting could be added, especially with fmt::
08:41:20 <Rubidium> because it's a shortcut with side effects
08:42:02 <_dp_> you mean gui.developer check? that could be removed I guess
08:42:10 <Rubidium> and IConsoleError would then be not much more than #define IConsoleError(...) IConsolePrint(CC_ERROR, __VA_ARGS__)
08:42:28 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9359: Consolidate IConsolePrint functions and improve some of the messages https://git.io/JZdyC
08:42:48 <_dp_> Rubidium, and that's fine imo
08:43:06 <TrueBrain> Rubidium: no clue if that update during my review had any effects :P
08:43:34 <TrueBrain> by far the best error we have is "insertion aborted", while registering an alias :D
08:43:43 <TrueBrain> it reads funny, and it is so out-of-context to the user :D
08:43:44 <Rubidium> which would mean you get shown the "doxygen comments" for the define in (some of the) IDEs, but there aren't any parameters to it... so you're a bit out of luck there
08:46:35 <TrueBrain> Rubidium: your PR also closes #8894 right? (as yours supersedes / invalidates it). You can add "Close #8894" somewhere in your PR description, and GitHub will auto-close it for you
08:46:45 <Rubidium> and using parameters is a bit annoying as well, as then you have to add format *and* have a __VA_ARGS__ with potentially nothing in it, which breaks annoyingly... unless C++20 which adds (better) support for empty ...
08:47:03 <TrueBrain> owh, you do mention that
08:47:07 <TrueBrain> but .. it doesn't pick it up
08:47:23 <TrueBrain> ah, nevermind me ... GitHub only shows the issues
08:48:17 <Rubidium> the update in between was mostly solving a conflict with the setting name -> std::string conversion
08:48:51 <TrueBrain> your solution brings IConsolePrint on-par with DEBUG() btw, which is nice
08:49:11 <TrueBrain> next we should replace DEBUG() to have FATAL / ERROR / WARNING / .. instead of numbers :P
08:50:51 <_dp_> haven't checked doxygen or IDEs but pretty sure they would be fine with it
08:51:41 *** HerzogDeXtEr has joined #openttd
08:52:53 <Rubidium> maybe, but it's also a bit chicken-egg-ish as we do not have fmt yet, and adding it and splitting things up is also doing two things at once which makes reviewing more difficult
08:53:34 <TrueBrain> over time we should look to become more like other logging platforms, so ConsolePrint.info() etc :P
08:53:58 <_dp_> well, fmt is optional there, you can still use variadic templates instead of defines without fmt
08:54:05 <TrueBrain> similar, DEBUG() should become Logging.info() etc
08:54:31 <Rubidium> and now you have 5 different functions, although you need even a couple more because there are more "levels"
08:55:12 <TrueBrain> and if we are ballsy, we use an existing library to give a logging facility
08:55:51 <_dp_> Rubidium, well, for logging it seems pretty common to have functions for each level for better readability and stuff
08:58:03 <Rubidium> yeah, but this not logging per-se. This is just printing, and the CC_XXXX is just the color to use
08:59:13 <_dp_> "printing" that only goes to console and log :p
08:59:19 <Rubidium> and the goal here is primarily to get everything consolidated so everything uses a single base, to base future improvements on. Instead of some things doing things differently
08:59:29 <_dp_> or do you suggest using DEBUG() for logging? :p
09:00:45 <Rubidium> okay, how should I call the IConsolePrint variant with IConsolePrint(TC_SILVER, ..) or TC_GREEN, or TC_LIGHT_BLUE
09:01:43 <_dp_> I don't even know why should there be such variants
09:02:12 <_dp_> You think of IConsolePrint as ingame console only while I mostly deal with it in server log, rcon and system console
09:02:23 <_dp_> none of which have color
09:02:32 <_dp_> well, mb rcon has but who cares
09:05:11 <TrueBrain> holy crap Rubidium , you did #9352 completely already, nice :D
09:05:14 <TrueBrain> that are a lot of files ...
09:06:10 <TrueBrain> also, we have too many debug statements :P
09:06:57 <Rubidium> The CC_XXX values are just "helpers" for the color of the text, which might bring some "meaning" in the call but it's more "this is the color that should be used for X, Y or Z", but sadly that also meant people started using them for other things which is a PITA. Things using CC_ERROR that are not an error at all, or CC_WARNING that should be an error, etc.
09:07:20 <Rubidium> so the patch is mostly to clean up all the misuses and inconsistencies between everything than anything else
09:08:25 <Rubidium> TrueBrain: C++20 so we can add unlikely() to the debug define ;)
09:08:30 <_dp_> Rubidium, yeah, there should probably be two separate functions (or sets of functions), one for cli, one for logging
09:08:46 <_dp_> CC_HELP has no business with logging
09:09:41 *** snail_UES_ has quit IRC (Quit: snail_UES_)
09:09:42 <_dp_> btw, iirc because both are the same admin port rcon sometimes gets random logs as result of a command
09:10:07 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9352: Codechange: use the fmt library for simpler debug formats https://git.io/JZdhf
09:10:12 <_dp_> since it just redirects the output
09:10:57 <Rubidium> it would definitely make sense to have separate "logging" functions for things outside of the console commands, but... outside of the scope for my PR
09:12:28 <TrueBrain> #9352 is seriously cool :)
09:12:48 <TrueBrain> can't wait for C++ to support f-strings
09:13:00 <TrueBrain> would be fun to listen to frosch, what his opinion is on that :D
09:13:43 <Rubidium> yeah, the description is slightly out-of-date now
09:14:15 <TrueBrain> "I did 2 files only to see what it does" vs "I DID EVERYTHING"
09:19:27 <peter1138> michael blunck rcon?
09:20:47 <TrueBrain> peter1138: I also still have the same impulse there :D
09:20:58 <TrueBrain> I am so happy I am not alone :)
09:23:22 <_dp_> Rubidium, wait, there kind of are separate functions already, they may not just always be used appropriately but that's different issue
09:23:31 <_dp_> IConsolePrintF is for logging and IConsolePrint is cli
09:23:33 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #9352: Codechange: use the fmt library for simpler debug formats https://git.io/JZFTU
09:24:03 <TrueBrain> michi_cc: you are a smart cookie! :D Much better solution
09:24:04 <_dp_> only downside is that they use same enum for levels
09:24:24 <_dp_> so the answer to CC_HELP is that IConsolePrintF shouldn't even accept that
09:24:39 <_dp_> and IConsolePrint shouldn't use CC_WARNING/ERROR
09:27:19 <_dp_> and the current set of IConsoleDebug/IConsoleWarning/IConsoleError makes perfect sense as they're logging functions, not cli
09:29:29 <Rubidium> yeah... IConsoleDebug is *only* used by DEBUG, nothing else. IConsoleWarning is primarily used by console_cmds telling the user the command they entered was wrong in some manner and IConsoleError was primarily used for the same. So for the latter two they are primarily used for cli.
09:29:38 <andythenorth> didn't I already do 3rd Company Colour for wagons in Horse?
09:29:43 <andythenorth> hmm yes I kind of did
09:29:54 <andythenorth> but it's a remap of 2CC for specific vehicles
09:29:58 <andythenorth> not the same as 3CC
09:30:19 <Rubidium> also IConsolePrintF being only for logging? Under which rock was that documented? As plenty of people have missed that as the cli code is full of those
09:30:36 <TrueBrain> (hint: as that was never the case :P)
09:31:20 <TrueBrain> hmm .. after just 9 ticks I have 2 savegames that should be identical differ in roadveh pathing
09:32:04 <TrueBrain> this is the "old savegame vs save/load of that savegame"
09:33:04 <_dp_> Rubidium, yeah, documentation sucks and TextColour should be split in two, but otherwise code is pretty solid
09:33:34 <_dp_> I mean console.cpp code, not the one that uses it xD
09:35:09 <Rubidium> well, people misappropriated IConsole* for logging purposes in (primarily) the network code. As the console and commands were there before the IConsole* calls in the networking code, so the code is not pretty solid. It's just using something that does what they need it to do, but it's conceptually not the right thing to do
09:37:29 <_dp_> misappropriated? but IConsolePrintF does exactly what logging function should, it outputs to both log and console
09:38:17 <_dp_> also I just noticed you removed it completely and started logging in IConsolePrint, wouldn't that spam cli in server log?
09:38:18 <Rubidium> IConsolePrint does exactly what IConsolePrintF does, except the formatting
09:39:08 <_dp_> oh, nvm, I read it wrong
09:39:36 <_dp_> for some reason I thought vseprintf does logging xD
09:39:46 <_dp_> wtf does logging then %)
09:40:22 <Rubidium> IConsolePrint does the logging
09:47:04 <TrueBrain> okay, this savegame has no NewGRFs, so that is not the issue .. and when using NPF there is also no issue
09:47:12 <TrueBrain> so some state in YAPF is different
09:49:14 <_dp_> Rubidium, what a mess...
09:53:04 <_dp_> hm.. does openttd dedicated server not even have a log file?
09:53:43 <_dp_> I'm so used to log files but it seems that it's something that my runner just creates from stdout/stderr
09:57:43 <TrueBrain> something about the roadstops that behave differently ..
10:00:19 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 commented on pull request #9352: Codechange: use the fmt library for simpler debug formats https://git.io/JZFBT
10:00:28 <TrueBrain> bus_station is weird when loaded from an old game
10:01:47 <TrueBrain> yeah .. that value is only updated for savegames after version 27
10:03:01 <TrueBrain> lol, if you would reload NewGRFs, it does correct the issue
10:04:50 <Rubidium> _dp_: it does *if* you configure it, but it gets the same you see in the in-game GUI console
10:05:00 <TrueBrain> Rubidium: fair point, Debug() should not call any functions if not executed. Good reason for the macro indeed :)
10:05:25 <_dp_> Rubidium, configure how? I see there is console command "script" that suddenly makes a log file, but, like wtfd
10:05:25 <Rubidium> and the console log is not only for the dedicated server, you can also use it in other situations
10:06:39 <_dp_> I mean, it kinda explains why it has no logging functions, who needs that when there is no log :p
10:06:56 <_dp_> even "script" just redirects cli output
10:09:15 <_dp_> ok, not even redirects, copies
10:09:43 <DorpsGek> [OpenTTD/OpenTTD] PeterN updated pull request #9289: Change: Shortcut varaction chains for callbacks. https://git.io/JsHuK
10:10:44 <peter1138> You seemed so certain that Console* stuff was for logging, not the in-game console...
10:11:07 <_dp_> yeah, because that's the only thing I use it for xD
10:11:25 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 updated pull request #9352: Codechange: use the fmt library for simpler debug formats https://git.io/JZKfv
10:11:43 <_dp_> well, ok, may be not the only, I did some vanilla patches but majority of the time for sure
10:12:52 <TrueBrain> anyone against the "fmt" inclusion? Going once?
10:13:30 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9352: Codechange: use the fmt library for simpler debug formats https://git.io/JZFg4
10:13:55 <peter1138> You've absolutely, never, opened the in-game console? Hmm.
10:14:39 <_dp_> and I don't use ingame console often indeed
10:14:54 <_dp_> kinda hard to do that on a dedicated server :p
10:16:39 <_dp_> though I guess I use it as chat history...
10:16:47 <_dp_> this freaking thing does everything xDDD
10:18:52 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #9366: Fix 81062163: for (really) old games, station bus/truck station cache was not updated https://git.io/JZFaN
10:19:08 <TrueBrain> Rubidium: I found a bug in your commit from 12 years ago :P
10:19:48 <peter1138> Is this why the cached path is different?
10:20:30 <peter1138> Phew, not my bug then :D
10:21:07 <TrueBrain> funny that it wasn't noticed back when it was introduces .. as those old savegames were more likely back then :)
10:21:43 <TrueBrain> well, I guess not, 27 is really old
10:22:01 <TrueBrain> 14000 commits earlier, basically
10:22:19 <peter1138> Yeah, not very likely, except of course the intro game? :D
10:23:52 <TrueBrain> which clearly doesn't have multiple bus-stops in a single station or something :)
10:23:58 <TrueBrain> or YAPF not enabled
10:23:59 *** frosch123 has joined #openttd
10:24:00 <TrueBrain> that is more likely :P
10:27:32 <TrueBrain> Rubidium also noticed this earlier today, but the MacOS runners are now sometimes a lot quicker. Now MSYS2 often is the slowest target :P
10:30:25 <frosch123> TrueBrain: c++ had f-strings first (stream <<). but because interleaving code and strings is terrible, they added the fmt. python did the reverse
10:30:26 <TrueBrain> as we are adding so many patch-killers to master, maybe a good moment to also move around files into folders ..
10:30:44 <TrueBrain> frosch123: I somehow just knew it would trigger you :P
10:31:15 <frosch123> [12:30] <TrueBrain> as we are adding so many patch-killers to master, maybe a good moment to also move around files into folders .. <- heh, i expected the second sentence to be "rename it to main" :p
10:31:40 <TrueBrain> I did a few now, and I am giving that some time to indicate any issue :P
10:31:45 <TrueBrain> but that is not really a patch-killer
10:31:51 <frosch123> what files do you want to move?
10:32:05 <TrueBrain> src/ is just too big, navigation is incredibly annoying
10:32:13 <frosch123> i thought we are at the "delete files, mostly from misc" step
10:32:15 <TrueBrain> so I was thinking stuff like: "newgrf" folder for all newgrf files
10:32:44 <frosch123> i don't consider that useful
10:33:00 <frosch123> when you do industires, i need industy_cmd, industry_gui, and newgrf_industries
10:33:03 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 approved pull request #9366: Fix 81062163: for (really) old games, station bus/truck station cache was not updated https://git.io/JZFP0
10:33:07 <TrueBrain> you might not, but I would :P
10:33:19 <TrueBrain> frosch123: those are not next to each other anyway
10:33:35 <TrueBrain> similar now that industry_sl is not in the same folder
10:35:15 <TrueBrain> anyway, a folder for "game" stuff, like vehicles, stations, ships, etc
10:35:31 <TrueBrain> one for more gui-like stuff, like window, widget, viewport
10:35:47 <frosch123> do you want the main src folder to be empty?
10:35:48 <TrueBrain> string/text related stuff in a folder ..
10:35:59 <TrueBrain> I don't mind if it would be empty, no, but I wouldn't say it is a goal
10:36:09 <TrueBrain> just 450 files in 1 folder is annoying as ...
10:36:51 <TrueBrain> I scroll more on a day working on OpenTTD than in a week otherwise :P
10:36:57 <frosch123> i always type the name
10:37:08 <TrueBrain> different people different workflows
10:37:13 <frosch123> and my editor is set to only show opened files, not the whole folder
10:37:31 <frosch123> i prefer typing over clicking :)
10:37:33 <TrueBrain> I hate that workflow :P
10:37:37 <TrueBrain> I have a mouse for a reason :)
10:37:49 <TrueBrain> anyway, it is now really focused on such a workflow
10:37:59 <TrueBrain> and we somewhat did create some folders over the years
10:38:04 <TrueBrain> but it is now in this weird in-between state
10:38:25 <frosch123> "widgets" is my favorite folder :)
10:38:41 <TrueBrain> for me, that kind of foldering is just pointless
10:38:45 <Rubidium> I guess the first consensus that has to be reached is over which lines do we want the folders to be split up
10:38:53 <frosch123> dropdown.cpp is completely out of place between all the enum-headers
10:39:07 <TrueBrain> I also have a real issue with "game" as foldername
10:39:17 <TrueBrain> Possibly I created it, but it is a shitty name :P
10:39:35 <TrueBrain> core vs misc .. never understood the difference :D
10:39:36 <andythenorth> this is controvs
10:39:47 <frosch123> TrueBrain: "misc" is stuff for deletion
10:39:54 <andythenorth> but could we have both options for industry funding in the 'fund' window?
10:40:06 <andythenorth> at the appropriate costs
10:40:18 <andythenorth> I know that it will cause server-owner outrage and shit
10:40:26 <Rubidium> do we do company/industry/house/town/... maps with all code related to those objects in there, or do you get a map/gui/sl/... folder with everything of that type in a map
10:40:29 <andythenorth> but I am bored of toggling the game setting
10:40:41 <TrueBrain> anyway, I am not that eager on changing everything or something .. but I think we can find groups of files we all agree on we can move to a folder .. that alone would help a lot :)
10:40:53 <andythenorth> I use 'prospect' when I am short of money OR I want a new random area of map
10:41:07 <TrueBrain> Rubidium: currently the second is done ... kinda. Personally I like the first more
10:41:07 <andythenorth> I use 'fund' when I need more of X in a certain area
10:41:23 <TrueBrain> (as what frosch123 pointed out, when you work on industries, you most likely only need those files)
10:42:28 <TrueBrain> as example, we have a "sound" folder and sound.cpp in the root :P
10:42:40 <TrueBrain> totally makes sense if you check the "sound" folder
10:42:57 <TrueBrain> but it sometimes takes a bit long to find stuff this way :D
10:44:03 <_dp_> andythenorth, why would server owners care?
10:44:18 <frosch123> i think (1) is impossible to achieve. there are too many overlapping things.
10:44:21 <TrueBrain> btw, frosch123 , in your workflow, does it change anything? As the search happens in subfolders too, not?
10:44:34 <TrueBrain> (honest question, to be clear, don't want to be snappy)
10:44:46 <frosch123> i also like (3) move stuff with shared speical rules to a separate folder
10:44:52 <frosch123> i think that applies to saveload code
10:44:57 <frosch123> and it may also apply to gui code
10:45:05 <frosch123> and game mechaincs/command handler code
10:45:13 <frosch123> since those may only talk through commands with each other
10:45:17 <andythenorth> _dp_ maybe they won't
10:45:20 <_dp_> andythenorth, unless you want to enable funding for prospect-only industries prospect is just somewhat inferior option
10:45:38 <_dp_> andythenorth, so it's mostly a ui nuisance, not gameplay change
10:45:49 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 merged pull request #9352: Codechange: use the fmt library for simpler debug formats https://git.io/JZKfv
10:45:58 <andythenorth> isn't there some big server owner drama about industries?
10:46:08 * andythenorth might be making something out of nothing
10:46:11 <andythenorth> goes it stop doing that?
10:46:34 <_dp_> andythenorth, I can think of a lot of idustry-related issues but no drama
10:46:50 <frosch123> TrueBrain: newgrf is a bad example. because you only look at the filename, while i think of all the functions which would be misplaced in the wrong file when applying a strong separation
10:47:05 <frosch123> and then i am back to: is it even possible to separate newgrf from the game mechanics
10:47:21 <TrueBrain> ah, you want to put the line like that
10:47:26 <TrueBrain> yeah, I am fine with that, honestly
10:47:32 <TrueBrain> anything to reduce this huge list :P
10:47:52 <TrueBrain> I think font/string could be a category too, not sure where it starts and ends
10:48:03 <TrueBrain> but that is such an isolated part of our game nobody really touches anymore
10:48:50 <peter1138> dropdown was in widgets before all the other stuff :-)
10:49:15 <peter1138> I have no objection to it moving out.
10:49:38 <TrueBrain> rename folder to "window" and move more stuff in :P
10:49:43 <peter1138> At the time I figured there would be other widget stuff that would be there, but, uh, of course dropdown is really a window...
10:50:32 <frosch123> so, i am fine with separation like: (1) saveload is special, because it calling any functions from game-mechanics causes dependencies in savegame-conversion. (2) gui is speical, because it may access the gamestate only readonly (3) rendering may be special if we put it into seprate threads, which may only access the gamestate special sync places (4) gamelogic may be special because it must be determinsitic and single-threaded and everything
10:50:55 <TrueBrain> fileio might also be a candidate
10:51:30 <frosch123> string.cpp is core, while strings.cpp is something else :p
10:51:58 <TrueBrain> yeah .. that never confused anyone evah :P
10:53:15 <TrueBrain> fileio could also contain stuff like ini-reader/writer
10:54:06 <TrueBrain> could stuff like cpu.cpp/h just move to core folder, or is that weird?
10:54:54 <frosch123> core is more like stdlib
10:55:02 <frosch123> cpu/filio would be /system ?
10:55:17 <TrueBrain> cpu.cpp has rdtsc :P
10:55:32 <frosch123> at both my workplaces, those functions would be in "os abstraction" :p
10:56:16 <TrueBrain> but system works .. maybe also stuff like crashlog could be in there
10:56:31 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on issue #9358: Bug Report - tar loader ignores empty files. An AI that requires such file fails to start https://git.io/JZ93n
10:56:34 <TrueBrain> lol @ dedicated.cpp .. I think that can be merged in the dedicated video driver :P
10:56:57 <frosch123> i guess you can add /system without controversy
10:57:14 <frosch123> while moving all gui files would probably be seriously upsetting for jgr and others :)
10:57:25 <frosch123> but noone touches that /system stuff
10:57:35 <TrueBrain> mostly moving files isn't a real issue with git, but for sure we have to check in with him :P
10:58:03 <frosch123> hmm, i don't think i have ever done a rebase involving renamed files
10:58:08 <frosch123> so, no idea what it means
10:58:26 <TrueBrain> rebases go surprisingly well, if done properly
10:58:45 <TrueBrain> merges is the one I am not fully sure about :)
10:59:11 <TrueBrain> fontcache and fontdetection .. those are basically too os-abstractions :P
11:01:18 <frosch123> hmm, jgrpp looks like it does neither merge nor rebase, but rather cherry-pick everything from master
11:01:44 <frosch123> all the master commits are "authored by <rb/tb>, committed by jgr"
11:02:36 <TrueBrain> so we should script this, and he can apply it too, if he likes
11:04:27 <peter1138> Rebase works very well with renamed files.
11:04:29 <TrueBrain> what btw always annoyed me .. MSVC users silently already did this a bit
11:05:13 <TrueBrain> for example, they have a virtual folder called "NewGRF" with all the newgrf files in there (sorry frosch123 , not meant to change your mind, as I think you are right)
11:05:24 <TrueBrain> and a "GUI" folder with all the _gui.cpp files in there
11:05:59 <frosch123> virtual folders are hell of confusing to me
11:06:24 <frosch123> everytime i add a new file, i have to check two folders, where to place it
11:06:29 <TrueBrain> I also really do not like them; but the fact that they can have them and I cannot annoys me :P
11:06:39 <frosch123> and only one matters for #include, while the other one is arbitrary
11:07:19 *** LordAro has quit IRC (Ping timeout: 480 seconds)
11:07:37 <TrueBrain> anyway, guess we should draft up what folders we want to have, and what should go in them
11:07:39 <TrueBrain> and take it from there
11:07:50 <frosch123> TrueBrain: look, there is a "Game Core" virtual folder :p
11:08:03 <TrueBrain> yeah ... wrong name for that folder :P
11:08:17 <TrueBrain> I think we should rename ai/game to script-ai and script-game
11:08:23 <TrueBrain> or put them under script/
11:08:42 <TrueBrain> I am always a bit lost what to find where anyway :D
11:08:49 <peter1138> Does MSVC still have that issue where you can't have two files named the same, in different paths?
11:09:02 <TrueBrain> I always wonder that myself
11:09:13 <TrueBrain> glx is not here yet? :D
11:09:26 <peter1138> That's why we have the _v, _s and _m suffixes, IIRC...
11:09:43 <TrueBrain> at least, that was the argument
11:10:12 <frosch123> i would keep that independent of msvc
11:10:24 <TrueBrain> we have 2 game_info.cpp now
11:10:24 <frosch123> saves you from naming the fullpath in irc everytime
11:10:26 <TrueBrain> nobody complained yet :P
11:11:09 <TrueBrain> frosch123: you do have a point, but I was wondering when the last time was when I said a filename on IRC .. :P
11:11:14 <TrueBrain> these days I mostly link GitHub :D
11:11:17 <TrueBrain> funny how that changed
11:11:21 <DorpsGek> [OpenTTD/OpenTTD] PeterN commented on pull request #9289: Change: Shortcut varaction chains for callbacks. https://git.io/JZFxP
11:11:42 <frosch123> i usually use the name, when some beginner coder is here
11:11:56 <frosch123> i know them from memory, but i am too lazy to look them up on github
11:12:10 <TrueBrain> I have found out that linking the file helps a lot against confusion
11:12:16 <TrueBrain> especially the right line number and all
11:12:24 <frosch123> i use github links only when i want to link a specific line
11:12:27 <TrueBrain> never realised before that I started doing that :P
11:12:40 <TrueBrain> GitHub is a good boy
11:12:48 <frosch123> though i think sometimes i just type the linenumber here as well
11:13:03 <_dp_> TrueBrain, few dozen messages earlier? :p TrueBrain> lol @ dedicated.cpp .. I think that can be merged in the dedicated video driver :P
11:13:07 <frosch123> hmm, does vscode have a "link this line from github"?
11:13:32 <TrueBrain> frosch123: good question; there are enough GitHub integration thingies .. you would expect so?
11:14:15 <TrueBrain> I only have the help with authentication installed
11:15:33 <michi_cc> TrueBrain: Just make the virtual folders real folders? (Heresy.... :D)
11:15:49 <TrueBrain> michi_cc: that is kinda what we are talking about :P
11:17:13 <TrueBrain> ugh, my test-suite is still not finished running
11:17:18 <TrueBrain> and I need to find a way to publish this ..
11:17:45 <TrueBrain> but, in contrast to yesterday, so far only the max_train_length is different
11:21:50 <TrueBrain> I like how summoning JGR works :P
11:23:21 <JGR> The cherry-picking into my branch is just a temporary thing, at some point I will merge everything and go back to regular merges
11:23:48 <TrueBrain> how do you feel about us moving files around?
11:24:33 <JGR> I'm not worried about that really, git can handle moves fine
11:25:51 <JGR> The main reason that I'm holding back on merging stuff at the moment is the network changes
11:26:06 <TrueBrain> yeah .. and I found a nice rabbit hole to fix first :P
11:26:22 <TrueBrain> SaveLoad -> Settings -> Network-code ..
11:26:26 <TrueBrain> it is not THAT deep :P
11:26:46 <michi_cc> I think it is more like a whole rabbit burrow :)
11:27:12 <TrueBrain> but the new savegame format is so much nicer :D
11:27:25 <TrueBrain> but michi_cc has to finish his newmap branch
11:27:27 <TrueBrain> that would be even better
11:28:15 <michi_cc> It's not "fixing" any m13245s, though.
11:28:36 <TrueBrain> I haven't looked at it, honestly, I am just repeating other people :P
11:29:33 <michi_cc> Tiny bits and pieces of it occasionally surface, like e.g. the land info window refactor.
11:30:01 <TrueBrain> mainly I would like the map-chunks in the savegame to be descriptive
11:30:09 <TrueBrain> so automated tooling can read it much easier
11:30:16 <TrueBrain> but that is not a short-term goal :D
11:32:14 <michi_cc> My newmap branch will not help you one bit with that :)
11:32:24 <TrueBrain> so I pick one of the others :P
11:33:09 <TrueBrain> but okay .. I only have to convert linkgraph and gamescript-translations .. so close
11:33:21 <TrueBrain> after that, clean up all commits, and make proper PRs :D
11:33:40 <TrueBrain> oof, not looking forward to rebasing the STUN PR :P
11:36:20 <Rubidium> just cherry-pick it then ;)
11:37:00 <TrueBrain> .... yes, I call that a rebase :P
11:37:07 *** gelignite has joined #openttd
11:37:49 <TrueBrain> owh, right, aircrafts also have a bug somewhere, in really old savegames ..
11:42:23 *** ChanServ sets mode: +v tokai
11:46:03 *** JGR has quit IRC (Quit: back later)
11:48:44 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #9366: Fix 81062163: for (really) old games, station bus/truck station cache was not updated https://git.io/JZFaN
11:48:52 <TrueBrain> okay, that PR really fixed most of the problems :D
11:49:14 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
12:00:04 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 updated pull request #9359: Consolidate IConsolePrint functions and improve some of the messages https://git.io/JZ9ZR
12:01:18 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 commented on pull request #9359: Consolidate IConsolePrint functions and improve some of the messages https://git.io/JZbZX
12:03:09 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9359: Consolidate IConsolePrint functions and improve some of the messages https://git.io/JZbnR
12:18:44 *** Progman has quit IRC (Remote host closed the connection)
12:21:22 <peter1138> "Game cannot be unpaused manually; disable pause_on_join/min_active_clients." Nice :-)
12:28:42 *** Progman has joined #openttd
12:35:11 *** Eddi|zuHause has quit IRC ()
12:39:32 *** Eddi|zuHause has joined #openttd
13:04:30 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 updated pull request #9359: Consolidate IConsolePrint functions and improve some of the messages https://git.io/JZ9ZR
13:15:58 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9359: Consolidate IConsolePrint functions and improve some of the messages https://git.io/JZbQo
13:25:36 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 closed issue #8853: There are 17 places where "IConsoleError" can be used instead of "IConsolePrintF". https://git.io/JqCmA
13:25:39 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 closed pull request #8894: Replace IConsolePrint*(F)( calls with specific console level calls https://git.io/JYqiY
13:25:42 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 merged pull request #9359: Consolidate IConsolePrint functions and improve some of the messages https://git.io/JZ9ZR
13:28:11 *** sla_ro|master has quit IRC ()
13:33:09 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick opened pull request #9367: Fix #9358: TarScanner no longer ignores empty files https://git.io/JZbpT
14:01:24 *** LordAro has joined #openttd
14:03:26 <Samu> wow, just noticed the new online players window
14:03:44 <Samu> we can join AI companies via it
14:16:43 <Samu> oh, only admin can join AI companies
14:17:05 <Samu> the button is still available for clients but clicking on it has no effect
14:27:41 <peter1138> I wonder if we have a dedicated place to report issues... Hmm.
14:32:33 *** LordAro has joined #openttd
14:40:26 *** andythenorth has quit IRC (Quit: andythenorth)
14:51:22 <Samu> crap, spectators need a button
14:55:11 <glx> are you looking at RebuildListCompany() ?
15:01:53 *** sla_ro|master has joined #openttd
15:07:11 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick opened pull request #9369: Fix: don't show join button in online players window when the company is an AI https://git.io/JZNiJ
15:10:48 *** Wormnest has joined #openttd
15:11:07 <Samu> hmm, I am assuming the company_id exists at that point
15:11:57 <Samu> if the company doesn't exist, Company::Get(company_id) is gonna crash
15:14:09 <glx> if company doesn't exist the function is not called
15:14:33 *** andythenorth has joined #openttd
15:15:35 <Samu> it is called for spectator
15:15:42 <Samu> which isn't a valid company
15:16:02 <glx> yes and you already protect this case
15:18:07 <glx> so it's called for valid company (Company::Iterate() ensures that) or COMPANY_SPECTATOR, and you check that already
15:49:34 <peter1138> Samu will one day learn how to write a commit title that makes sense.
15:50:06 <peter1138> This once sounds that TarScanner used to ignore empty files, but doesn't any more, and you are fixing it so that it does ignore empty files agains.
16:15:20 <peter1138> master doesN'T start
16:16:04 <peter1138> Crashes doinbg RoadStop stuff in AfterLoadGame. Hah!
16:18:01 <peter1138> 137: assert(this->checked != 0);
16:20:35 <peter1138> pool_func.hpp:137, that was meant to be.
16:23:47 <Rubidium> master starts for me
16:25:26 <Rubidium> c++ ... -DWITH_ASSERT ...
16:29:15 <Rubidium> do you not have the default start save game maybe?
16:29:26 <Rubidium> or might something really fishy be going on?!?
16:45:33 <peter1138> I can't really guess why that would make pool stuff not work.
16:46:51 <Rubidium> that's not HEAD of master though
16:47:55 <peter1138> Not sure what the point is, but it is correct.
16:48:56 <Rubidium> TB's change to AfterLoadGame is not even in 55a11710, so you got some weird partial version or so?
16:51:11 <peter1138> Also correct, by which I can assume it's nothing to do with TB's change.
16:53:15 <peter1138> When I wrote 55as11710, I meant, that is the first version it occurs in. a9o9ac62c starts fine.
17:01:22 <peter1138> No diff. Already, full on debug builds.
17:02:25 <peter1138> Keyboard dyslexic :(
17:10:59 <peter1138> Okay, debug works fine.
17:11:25 <TrueBrain> a while ago I pulled my hair out of my head, as not calling CanAllocateItem crashed the game
17:11:30 *** snail_UES_ has joined #openttd
17:11:33 <peter1138> Just deleted my build folder and starting fresh.
17:11:34 <TrueBrain> but the function itself doesn't anything useful to be a reason why
17:11:51 *** tokai|noir has joined #openttd
17:11:51 *** ChanServ sets mode: +v tokai|noir
17:11:55 <TrueBrain> never figured out what was going on ... at a certain point it "just worked" again
17:12:23 <TrueBrain> can't say the Pool gives readable errors :P
17:12:52 <peter1138> Yeah, value was optimised out... and full debug works fine. Nice.
17:17:14 <TrueBrain> okay, fmt is making our assert act up a bit
17:17:18 <peter1138> Fresh build, RelWithDebInfo, still crashes.
17:18:06 <TrueBrain> release build with asserts is not behaving as it should :D
17:18:30 <TrueBrain> at least, I assume it is fmt, as it wasn't an issue yesterday :P
17:18:41 *** tokai has quit IRC (Ping timeout: 480 seconds)
17:18:48 <TrueBrain> peter1138: crashes for me too
17:19:03 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
17:21:28 <TrueBrain> and indeed, the usage of fmt disrupts our assert stuff
17:24:28 <TrueBrain> lets see if a stupid fix helps .. just recompiling takes for-ever
17:28:36 <TrueBrain> indeed, it is the inclusion of cassert in the fmt library
17:28:40 <TrueBrain> disabling that fixes the issue
17:29:19 <LordAro> we already include cassert ourselves, surely?
17:29:21 <TrueBrain> because it uses asserts, and wants to be self-contained ...
17:29:29 <TrueBrain> this is a 3rdparty library, mind you
17:30:28 <TrueBrain> and it seems assert.h has no header-double-include-protection-thing
17:31:29 <TrueBrain> no clue why assert.h does that
17:31:33 <TrueBrain> but.. system-include, so ..
17:31:38 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
17:31:52 <TrueBrain> so either we have to include fmt in stdafx
17:31:56 <TrueBrain> or modify fmt, I guess
17:32:45 <LordAro> almost every file has debug statements, right?
17:32:51 <LordAro> no great loss to include it in stdafx
17:33:03 <TrueBrain> we don't have to include the debug stuff
17:33:19 <TrueBrain> it is just a bit weird
17:37:16 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
17:39:59 *** jottyfan has joined #openttd
17:40:30 *** WormnestAndroid has joined #openttd
17:48:15 *** ChanServ sets mode: +v tokai
17:53:33 <TrueBrain> maybe Rubidium has some ideas too
17:55:05 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
18:11:49 *** Wormnest has joined #openttd
18:12:53 <frosch123> hmm, so WITH_ASSERT is set, but assert() is skipped
18:13:49 <TrueBrain> yeah, and we fix that in stdafx.h nicely
18:14:04 <TrueBrain> basically, release builds set NDEBUG, which means assert.h skips setting assert()
18:14:12 <TrueBrain> we change that again when WITH_ASSERT it set
18:14:27 <TrueBrain> but ... assert.h resets that again
18:14:47 <TrueBrain> basically, after stdafx.h you are not allowed to include cassert anymore .. but fmt does :P
18:15:53 <LordAro> can we do WITH_ASSERT/NDEBUG stuff 'better' ?
18:15:58 <LordAro> such that this isn't an issue?
18:16:15 <TrueBrain> assert.h assumes a binary world
18:16:17 <TrueBrain> either you use NDEBUG
18:16:24 <TrueBrain> CMake assumes a binary world
18:16:28 <TrueBrain> Debug is without NDEBUG
18:17:02 <TrueBrain> and we .. want an in-between :P
18:18:03 <TrueBrain> normally ofc this is not a problem at all .. unless we use 3rdparty libraries that run a shitton of includes themselves :D
18:18:22 <Rubidium> I'm a bit torn between adding it to stdafx and removing the cassert from 3rdparty/fmt
18:18:35 <TrueBrain> we also modified squirrel heavily
18:18:35 <Rubidium> stdafx is used by about double the amount as debug.h
18:19:18 <TrueBrain> but it makes updating more difficult
18:19:46 <andythenorth> river building in normal game mode...when? :)
18:20:11 <Rubidium> it's 489 "files" vs 235 "files", but yeah that's why I'm a bit torn about it
18:20:49 <TrueBrain> andythenorth: REALISM!
18:24:38 <frosch123> pff. assert.h goes to great extend that include guards should not work in any case
18:25:27 <TrueBrain> NO! IF PEOPLE INCLUDE ME! I! HAVE! TO! WIN!
18:27:04 <frosch123> i vote for "remove cassert from 3rdparty"
18:27:50 <frosch123> i tried "#define cassert cstdint", but apparently macros do not work inside #include :p
18:28:11 <TrueBrain> I was wondering if you could keep the define in a temp
18:28:15 <TrueBrain> and set the define back
18:28:20 <TrueBrain> but .. that might be tricky :P
18:29:40 <Rubidium> I'd go with the removal from 3rdparty too, but not only removing but rather adding a comment at that location so when/if a new version is pasted over it is not just adding "cassert" but also removing the "do not add cassert as it breaks things" comment, so it should be easier to spot during a potential update
18:36:43 <peter1138> Yeah, removing the include works.
18:37:33 <Rubidium> anyone working on a PR for that?
18:40:34 <Rubidium> LordAro: 235 files are recompiled when changing debug.h, 489 when changing stdafx.h
18:40:51 <Rubidium> oops... scrolled back and replied to something I thought was new ;)
18:41:12 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 opened pull request #9370: Fix a99ac62: fmt's include of cassert breaks our assert logic https://git.io/JZxG4
18:55:43 *** tokai|noir has joined #openttd
18:55:43 *** ChanServ sets mode: +v tokai|noir
18:55:59 <DorpsGek> - Update: Translations from eints (by translators)
19:01:29 <DorpsGek> [OpenTTD/OpenTTD] PeterN approved pull request #9370: Fix a99ac62: fmt's include of cassert breaks our assert logic https://git.io/JZxBW
19:02:21 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 merged pull request #9370: Fix a99ac62: fmt's include of cassert breaks our assert logic https://git.io/JZxG4
19:02:30 *** tokai has quit IRC (Ping timeout: 480 seconds)
19:02:43 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9370: Fix a99ac62: fmt's include of cassert breaks our assert logic https://git.io/JZxB7
19:02:49 <TrueBrain> also put our IRC conversation in the ticket .. as this will hurt us sooner or later again :)
19:04:18 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
19:07:41 <TrueBrain> nicely spotted btw peter1138 :D
19:29:10 *** Wormnest has joined #openttd
19:46:01 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
19:47:05 *** WormnestAndroid has joined #openttd
20:02:47 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 opened pull request #9371: Assorted C-string -> std::string conversions in network https://git.io/JZxHY
20:06:02 *** sla_ro|master has quit IRC ()
20:11:14 <TrueBrain> not sure what prompted this message of love, but it is appreciated :D
20:14:56 *** jottyfan has quit IRC (Quit: jottyfan)
20:29:58 *** virtualrandomnumber has joined #openttd
20:30:16 <NGC3982> just thought about it
20:30:28 *** virtualrandomnumber has quit IRC ()
20:37:01 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:03:48 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
21:07:54 *** gelignite has quit IRC (Quit: Stay safe!)
21:21:10 *** Samu has quit IRC (Quit: Leaving)
21:24:24 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 opened pull request #9372: Fix: [Network] Determining GetNetworkRevisionString could overflow and underflow its buffer https://git.io/JZpCf
21:37:45 *** Progman has quit IRC (Remote host closed the connection)
21:41:43 <DorpsGek> [OpenTTD/OpenTTD] rubidium42 updated pull request #9371: Assorted C-string -> std::string conversions in network https://git.io/JZxHY
21:58:20 *** andythenorth has quit IRC (Quit: andythenorth)
22:14:42 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:35:20 *** ChanServ sets mode: +v tokai
22:41:56 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
23:08:20 *** Beer has quit IRC (Quit: Leaving)
23:08:46 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
continue to next day ⏵