IRC logs for #openttd on OFTC at 2023-09-11
⏴ go to previous day
00:31:07 <locosage> truebrain: What you indicated is that you were never interested in merging 11274 in favour of writing it you own way. Which is fine, I understand that writing your own code is often simpler than understanding someone's else and explaining how to fix it, especially in that mess of cargo code. But it would be much simpler for everyone if you just stated that clearly from the start istead of talking
00:31:07 <locosage> riddles and showering me with merge conflicts.
01:18:50 <locosage> truebrain: For all the lolz about direction not a single person suggested a better name or how to better explain the idea behind that calculation. What you suggested is to make yet another payment system. That's no review, that's just asking someone to implement your own idea.
01:30:21 *** NGC3982 has quit IRC (Ping timeout: 480 seconds)
01:33:32 *** NGC3982 has joined #openttd
01:43:45 <locosage> Imo what you proved with your split is that it's completely illogical. There are many ways to fix staion sign manipulation, what makes 11281 the right one? 11274 makes, that's what.
02:00:38 *** Wormnest has quit IRC (Quit: Leaving)
02:55:24 *** debdog has quit IRC (Ping timeout: 480 seconds)
02:57:32 *** D-HUND is now known as debdog
03:35:27 *** Smedles_ has joined #openttd
03:38:34 *** Smedles has quit IRC (Ping timeout: 480 seconds)
04:19:08 *** tokai|noir has joined #openttd
04:19:09 *** ChanServ sets mode: +v tokai|noir
04:26:08 *** tokai has quit IRC (Ping timeout: 480 seconds)
05:29:08 *** keikoz has quit IRC (Ping timeout: 480 seconds)
06:01:08 <locosage> Lol, 11283 reintroduces sign manipulation.
06:14:15 <locosage> Ah, no, nvm, "between source and destination" probably means between load/unload tiles in that context.
06:28:07 <truebrain> Hmmm .. now to bisect the crashing coop game .. this will take a lot of compiling 😦
06:28:29 <truebrain> I keep wondering if ccache would help
06:28:47 <truebrain> Did anyone try that by any chance?
06:45:30 <peter1139> ccache and distcc right? :D
06:48:21 <osswix> Anywho, I guess I'll also just do an ask here.
06:48:21 <osswix> How (un)likely are we to see the ability to do ballast (as newgrf) as separate thing from rail in the same way tram tracks are separate from rail? It would significantly simplify working with rail types decoratively (rather than needing the same rail type multiple times to account for different ground / ballast features)
06:48:21 <osswix> (And likely even more remote, the same but for electrification styles)
06:49:09 <peter1139> As in, very unlikely.
06:49:23 <peter1139> I mean, if you made a PR for it, it would be more likely.
06:51:30 <peter1139> Probably the most appropriate solution is UI changes to make it appear that separate rail types are related.
06:53:17 <jfs> the rail type (and road type) selection UI is still just the toolbar menu, right?
06:53:43 <peter1139> Using map data bits for ballast or electrification type actually reduces flexibility, as combinatorily, you end up reducing the space for actual distinct railtypes.
06:53:46 <jfs> because that's been pretty bad since railtypes was introduced as more than just standard/electric/monorail/maglev
06:54:29 <peter1139> Not really, when the limit was 16 is was fine, as the limited space meant that the types implemented were fairly distinct.
06:56:55 <truebrain> osswix: basically, because we are build on top of a design from 1994, we are kinda limited in how we can extend information on a tile. Which really sucks for so many good ideas 🙂
06:57:14 <peter1139> It's not even that, it's wasteful.
06:59:03 <osswix> Hmm, better ui would probably do the same job yeah. To me having them as separate options sounds a lot easier for making newgrf / compatibility between newgrf.
06:59:20 <truebrain> also I guess means people need to agree how to use the bits for railtype?
06:59:43 <truebrain> from what I read, that is always a hot debate with no actual outcome
06:59:50 <truebrain> not really involved in that, just mentioning as I remember something like that
06:59:54 <peter1139> They'd have to agree on what "ballast type 1" means..
07:00:02 <peter1139> What does "electrification type 2" mean?
07:00:33 <peter1139> If you reserve map bits for those features, you end up have 4 ballast types and 6 electrification types for ... monorail and maglev.
07:00:56 <peter1139> Nobody puts 600V DC on maglev.
07:00:57 <osswix> I guess moreso from the prospect of those editing the game rather than newgrf.
07:01:22 <osswix> Yh, it'd be more of a waste of space in that case
07:01:33 <osswix> Hmm, could you fake it?
07:02:01 <peter1139> This has been discussed before, by the way :) The most efficient use of railtype space is the current way, even though it seems like it is wasteful becuase you have to define different types.
07:02:11 <truebrain> I think the problem you now run into is more of a social problem, to get consensus. We, as developer, can enforce a consensus, but it might result in a lot of pitchforks 😛
07:02:13 <osswix> Have the back end work as is, but the user experience the way tram + road combine?
07:02:41 <peter1139> Yes, that's the point of UI changes, to make it look like you are selecting a railtype, then the electrification type (or vice versa... would NewGRF authors even agree on a standard for that...)
07:03:05 <truebrain> moddable UI you say? 😄
07:04:02 <peter1139> There are 6 bits for the rail type. If you say 2 bits are electrification type (that's only 4 types, including none!), then you have reduced the railtype space down to only 16 types. Add on 2 bits for ballast type, and your railtype space is now 4 types...
07:04:37 <truebrain> stupid map limitation 😦
07:04:47 <peter1139> Add more properties to rail types to allow the game to organise the list better is definitely a better idea.
07:05:35 <truebrain> hmm ... this coop game has been crashing for a long time ... I am back 1000 commits, and it still crashes
07:05:39 <truebrain> do I even bother at that stage?
07:05:52 <peter1139> Does it crash on the version it was saved with?
07:06:11 <peter1139> (Or 13.4 for the sake of a simple test)
07:06:14 <truebrain> it is the 5th coop game .. that was a very long time ago, and requires me to use the old config system .. I am not going to find out 😛
07:06:24 <truebrain> it doesn't crash on release builds, as asserts are disabled there 😄
07:06:50 <truebrain> it was more that I was scared that our recent changes of modernizing OpenTTD caused it
07:06:56 <truebrain> but it seems to be an existing issue ..
07:08:34 <truebrain> crashed even 2000 commits ago. I am fine with this 🙂
07:09:44 <jfs> introduce a tagging system for rail and road types. keep the current fourcc codes and count limitations, but allow these tags to be used both for UI filtering and for vehicle requirements testing
07:09:44 <jfs> (and enjoy all existing railtype GRFs now missing all that info)
07:10:56 <peter1139> truebrain, is it an openttdcoop one then?
07:11:35 <truebrain> I dunno, didn't check that far back
07:11:42 <truebrain> again, was mostly worried we caused it recently 🙂
07:12:39 <truebrain> the coop site works? 😮
07:13:02 <peter1139> wiki does apparently. Just the devzone was deaded?
07:13:07 <jfs> tags like, "iron wheels", "tyre wheels", "magnetic levitation", "two rails", "one rail", "overhead electrification", "thirdrail electrification", "DC electrification", "AC 50 Hz electrification", "AC 16 1/3 Hz electrification", "600 V electrification", "1500 V electrification", "15 kV electrification", "25 kV electrification", "axle load max 10 t", "axle load max 30 t", "axle load max 70 t", etc
07:13:11 <truebrain> when cmake reruns on my WSL, it calls vcpkg, which starts windows executable to test for various of things .. very slow, very pointless 😛
07:14:01 <truebrain> jfs: from a BaNaNaS perspective, one thing we have learnt, that (some? most?) authors refuse to look at tags others have declared, and just do their thing
07:14:09 <peter1139> jfs, exactly. Apply one of those tags to a railtype, then the UI can group / sort / filter better.
07:14:23 <peter1139> [10] ./openttd(_ZNK9OrderList16DebugCheckSanityEv+0x445) [0x55db49327685]
07:14:24 <truebrain> not saying your idea is bad, just mentioning it might become very annoying to combine NewGRFs. Not saying that is bad either 🙂
07:15:26 <peter1139> truebrain, true, unless we predefine all possible electrification types and say "yeah, that's it"
07:15:28 <jfs> the problem is just that there are so many parameters differentiating railway tracks, and some vehicle sets might be more demanding than others
07:15:56 <truebrain> let's put it this way .. the amount of misspelling done to "electrification" will alone be hilarious 🙂
07:16:22 <truebrain> so what might help, is to settle on the units, and the method of writing, or something. Just as guide. Doesn't have to be enforced
07:17:02 <peter1139> Fun: Tag Labels, but also Tag Type Labels.
07:17:14 <jfs> any set of tags would probably need to be defined by the game, and not be completely free form, to also allow hierarchy in them, so e.g. vehicles can declare "I just want electrification and two rails", and the game will let it run on anything having any tags covering that
07:17:44 <truebrain> omg, did I actually found another issue with "next_station" in CargoPacket?
07:17:49 <truebrain> this is getting annoying 😛
07:17:57 <jfs> and at that point maybe it's better to not really make it tags, but just some parameters
07:18:36 <truebrain> ah, no, it is a bug in my diff
07:19:17 <jfs> and also things like axle load can be determined automatically, if the game knows how many axles a vehicle has
07:19:51 <peter1139> Handy, Game 5 doesn't crash immediately.
07:20:03 <osswix> Couldn't/wouldn't you limit those types of tags for newgrf making?
07:20:05 <peter1139> I'm assuming they didn't use NewGRFs.
07:20:37 <osswix> Like in the same way signals are limited already
07:20:38 <peter1139> osswix, what's the point limiting tags that are meant to the author defined and of which we can't possible provide all possible values out of the box?
07:22:01 <osswix> But I'd welcome UI changes even if as an optional thing newgrf makers could but don't have to tailor their railtypes to?
07:22:06 <truebrain> peter1139: yeah, it takes a few ticks to crash .. and it is CPU bound, so you can't actually FF .. also why I said: fuck it 😛
07:23:22 <peter1139> It's in the OnNewDay thing where it verifies order lists are valid.
07:23:43 <truebrain> yeah, the crash itself didn't give me much info honestly
07:24:34 <peter1139> Okay, the assert checks that order-shared vehicles are actually using the same order list. Which you'd think...
07:25:44 <truebrain> so this most likely was already broken in that nightly, but just not visible? 😄
07:26:15 <peter1139> I'm not sure, I would've thought that the list is built on loading, not saved.
07:34:42 <truebrain> We are a weird bunch of people .. I give up, the next person jumps on it 😄 I honestly like (and appreciate) it 🙂
07:34:44 <peter1139> Ah no, next_shared is saved.
07:35:29 <peter1139> Though that makes sense -- it could perhaps be rebuilt based on shared orders/orderlists but making it consistent for network saves might be awkward.
07:37:02 <peter1139> Probably the DebugCheckSanity() function should checked on load though, and not just on new day when asserts are enabled...
07:37:21 <truebrain> talltyler: the reason regression fails, is because you cannot have a static on TimerGameCommon, inherit is in TimerGameCalendar and TimerGameEconomy, and expect them to have different values 🙂
07:37:31 <truebrain> peter1139: sounds like a good idea, yes
07:39:27 <truebrain> talltyler: so basically, you were sharing the same variable for both timers .. which meant things could go twice as fast or something 😛 Dunno exactly, but things break 😄
07:45:28 <truebrain> will make a commit for you to see what I mean, and also hopefully resolve the StrongTypeDef issue .. testing that currently, see if it actually works 🙂
07:45:58 <truebrain> mainly, getting the linker to like templates and functions is always a bit of a challenge 🙂
07:49:25 <truebrain> hmm .. even then it doesn't complain about types being different .. oh-oh .... let's see ..
07:51:11 <truebrain> `error: no viable overloaded '='`
07:52:23 <truebrain> let's see how often it is broken now ....... 😛
07:56:47 <truebrain> happy with this StrongTypeDef 😄
07:58:39 <peter1139> How many seconds...
07:58:48 <truebrain> my PRs take longer! 😛
08:02:13 <truebrain> `no viable conversion from 'Typedef<[...], TimerGame<Calendar>::DateTag<Calendar>, [2 * ...]>' to 'Typedef<[...], DateTag<Economy>, [2 * ...]>'`
08:02:17 <truebrain> the error even is rather clear
08:02:36 <peter1139> Do I attempt to respond or just go to ignore?
08:02:48 <peter1139> Oh dear, $work o'clock :(
08:03:53 <peter1139> Oh wow, MS Teams is actually working this morning. After a force-reload anyway.
08:10:53 <peter1139> We stack up asserts quite a lot. Should we care?
08:11:54 <peter1139> e.g. RailClearCost() asserts that railtype is valid, and then calls RailBuildCost(), which asserts that railtype is valid, which then calls GetRailTypeInfo(), which asserts that railtype is valid...
08:12:11 <truebrain> ah, yes, those stacking 🙂
08:12:18 <truebrain> we can be rather pedantic about that 😄
08:13:29 <truebrain> such template, such type-checking
08:14:30 <truebrain> ```(helpwindow %)$ git rebase upstream/master
08:14:30 <truebrain> Successfully rebased and updated refs/heads/helpwindow.```
08:19:14 <truebrain> small fix; otherwise, just rebased ^^
08:19:36 <andythenorth> oof I only stepped away for a minute 😛
08:19:45 <truebrain> step away more often!
08:20:56 <andythenorth> I might not read back 😛
08:21:00 <andythenorth> did I miss anything?
08:22:07 <peter1139> You probably missed lunch.
08:23:14 <andythenorth> ach changing newgrfs has done a railtypes
08:23:24 <andythenorth> Monorail and Maglev are not supposed to be there 🙂
08:25:39 <truebrain> andythenorth: mostly that peter1139 is far slower in reviews than I am 😛 Not measuring or keeping score, but almost an hour and counting 😛
08:26:12 <andythenorth> he's probably working to pay rent
08:26:20 <truebrain> he actually has a job
08:27:14 <peter1139> Well, after I didn't review the other cargo packet change very well, it needs a bit of time, and yes, $work :/
08:27:41 <truebrain> I am just pulling your leg really 🙂
08:27:56 <peter1139> They need it after yesterday.
08:28:36 <truebrain> and yes, I lovely fucked up the loaded_at_xy PRs 😛 Stupid code ... stupid human ... 😄
08:29:02 <andythenorth> "if we're making mistakes, at least some choices are getting made"
08:29:23 <andythenorth> did the CI catch it?
08:29:38 <truebrain> unless you consider JGR a CI
08:30:10 <andythenorth> I missed the obiwan in engine lifetime, I saw it, but assumed the bug was in Horse 🙂
08:30:26 <truebrain> it isn't even an obiwan .. it is our code being weird
08:30:37 <truebrain> sometimes years are 365 days, sometimes 366, sometimes only on leapyear
08:30:51 <andythenorth> reality intrudes 🙂
08:31:05 <truebrain> so the obiwan becomes a obitwo when you have an age of like 400
08:31:28 <peter1139> Heh, the original code deliberately used the wrong longer length so that it would always look right.
08:31:31 <truebrain> depends what function was actually used 😄
08:31:36 <peter1139> It just didn't explain why it did that.
08:31:44 <truebrain> peter1139: that got lost in translation, yes 😄
08:31:44 <peter1139> Or at least, didn't explain well enough.
08:33:23 <peter1139> So people are 'abusing' tram-types as road graphic overlays or something...
08:33:47 <andythenorth> I had many ideas 😛
08:33:56 <peter1139> What did you do... :(
08:34:22 <andythenorth> like...how many Xtypes does a player need really?
08:34:26 <andythenorth> but then multiplayer 😦
08:34:44 <truebrain> the final boss appeared?
08:35:55 <peter1139> Final-boss: merge conflicts as an indicator of PR worthiness.
08:36:33 <truebrain> in that case nielsm won the lottery 🙂
08:39:13 <truebrain> a .cpp file without safeguard, lol. You don't see that a lot anymore 🙂
08:39:38 <truebrain> that is a terrible way to define yourself
08:40:50 <truebrain> the help stuff looks really good .. I can't believe the links actually work, also the anchors ..
08:41:16 <peter1139> What's a safeguard in the context of a .cpp file?
08:41:32 <truebrain> we include safeguard.h at the end of the include block
08:41:38 <truebrain> making sure you don't use functions you are not suppose to use
08:41:50 <peter1139> This must be old ;)
08:42:02 <truebrain> cleaning it up as I go 🙂
08:42:27 <truebrain> lol @ dotted line .. IT IS TOO LONG FOR MY SCREEN 😄
08:43:11 <truebrain> owh, there are even navigation buttons on the top of the window 😮
08:43:16 <truebrain> jfs: you outdid yourself here
08:45:53 <truebrain> the help window stuff, you really did so much more then I expected it to do 🙂
08:46:07 <truebrain> that it even has navigation buttons
08:46:38 <jfs> I also recently considered making major headings (like level 1 and/or 2) use the large font instead of normal, but thought that would be tmwftlb
08:46:52 <jfs> yes and forward/back buttons for hyperlinks
08:47:08 <truebrain> this is already amazing. So much stuff we can improve and thinker with .. but .. Rome isn't build in one day
08:47:42 <jfs> loading and displaying PNGs in the GUI is something that would be really neat to have as well
08:48:06 <jfs> instead of having all graphics go through a GRF format
08:48:37 <truebrain> we have a similar request for heightmaps and scenarios .. but it kinda requires us to ditch the 8bpp as option
08:48:49 <truebrain> all OSes already default to 32bpp, so we might be ready for that .. survey with 14.0 will tell 🙂
08:49:19 <jfs> an idea I've had is allowing basesets to include a logo graphic, and display those logo graphics on the title screen in the corners or something like that, similar to TTD having Chris Sawyer and Microprose logos in the corners
08:49:40 <truebrain> ideas we have plenty 😄
08:49:52 <truebrain> time to execute .... is the issue 😛
08:50:18 <truebrain> as I would like basesets to show a preview before downloading .. so you have an idea what you are actually getting
08:57:58 <andythenorth> that was killing me 🙂
08:58:06 <andythenorth> I think docs and grf stuff also do it in places
08:58:16 <andythenorth> and track_type vs railtype
09:02:41 <truebrain> there is even a Table of Content that is functional 😮 Did not spot that earlier ...
09:02:57 <truebrain> dude ... there is so much in this PR ...
09:08:46 <truebrain> `std::vector<Line> lines;`
09:08:47 <peter1139> When adding one little feature requires a major refactor... means I got it wrong before, and potentially again... :/
09:08:49 <truebrain> `std::vector<Line> lines{};`
09:09:17 <truebrain> peter1139: what little feature are you adding? 😄
09:11:59 <peter1139> Minimum order value, along with delivery charges. The refactor is because too many things are pulled together at display (view) time, whereas I now need them during controller/model time.
09:12:29 <truebrain> rewrite from scratch?
09:13:15 <peter1139> Not entirely, just moving it from "where it's used".
09:14:16 <andythenorth> peter1139: possibly over-dramatised
09:14:17 <peter1139> {} is needed to make std::array initialize, but std::vector is fine without. But putting {} in does not harm it.
09:14:22 <peter1139> andythenorth, it's my job eh?
09:14:24 <truebrain> oof, lot of doxygen missing in that Help window PR .. let's fix that ..
09:14:49 <truebrain> peter1139: yeah, the `array` keeps confusing me. Tnx. So PODs only, basically. Works for me.
09:17:21 <truebrain> it even remembers where you were scrolled to when navigating, so when you go back you end up there again
09:17:32 <truebrain> this PR ... it is one of the most polished I have seen in ... years? decades?
09:34:30 <truebrain> hmm, nightlies haven't run for a week 😛
09:35:04 <truebrain> so my loaded_at_xy fuckups never made it into a nightly \o/
09:40:49 *** Deep3D has quit IRC (Quit: znc)
09:44:05 <peter1139> Of course, although Teams was loaded, it wasn't actually working...
09:44:21 <peter1139> truebrain, cunning.
09:44:22 <truebrain> 😄 The sweet silence of working-from-home? 😛
09:45:39 <peter1139> It doesn't "ring" unless the browser tab is focused... I think that's probably a Firefox issue.
09:46:03 <peter1139> Anyway, the call was "why was this product incorrectly charged" and the answer was, of course, "because you put the wrong price in"
12:00:26 <locosage> weird, why does github place comments not in the code order
12:02:51 <peter1139> AI-powered, 'nuff said.
12:08:57 <locosage> apperently c++ allows to call static functions via objects, how "fun"
12:27:44 <andythenorth> pyflakes has a surprisingly low level of anger about FIRS etc
12:31:53 <_pruple> what's wrong with FIRS?
12:32:00 <_pruple> (apart from the obvious)
12:32:31 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
12:33:54 <andythenorth> specifically the pyflakes linter finds a bunch of minor code quality issues 😛
13:35:47 *** virtualrandomnumber has joined #openttd
13:36:27 *** virtualrandomnumber has quit IRC ()
13:42:29 *** ufo-piloot has quit IRC (Remote host closed the connection)
13:42:42 *** ufo-piloot has joined #openttd
14:17:52 <truebrain> talltyler: saw the commit I prepared for you?
14:26:14 <talltyler> Yes, I'm working on it right now 🥰
14:26:54 <talltyler> News items could use a redesign someday to not be a linked list, but that's probably out of scope for now
14:26:58 <truebrain> Good. Just checking it didn't get lost in the chat of the day 😄
14:28:49 <talltyler> I'm adding separate economy and calendar dates to each news item, because they are cleaned up by economy date, but the date on the newspaper needs to be calendar date... Not ideal, but it'll do for now I guess. News items don't get stored in the savegame (as far as I can tell) so it shouldn't be a big deal to change later
14:32:09 <truebrain> Savegames are my new nemesis
14:32:22 <truebrain> For some reason I keep fucking it up 😛
14:32:39 <peter1139> Hmm, I think I had already worked on a patch for that. Probably got it, lost somewhere...
15:16:43 *** HerzogDeXtEr has joined #openttd
15:55:54 *** Wormnest has joined #openttd
16:29:02 <andythenorth> I really thought that was going to be a Harvester pub thing
16:29:09 <andythenorth> dissappointed TBH 😛
16:29:37 <LordAro> which has nothing to do with cubicals either
16:31:23 <andythenorth> hmm how much map array rewrite would we need to do biomes?
16:31:52 <locosage> in a simple form probably not much
16:32:09 <locosage> just need to unstuff ground types, trees are already universal
16:33:19 <locosage> it's probably significantly more work to sort out graphics though
17:20:59 *** gelignite has joined #openttd
17:45:35 <truebrain> LordAro: Cool, never heard of the first, looks like a nice solution! And k8s ... if you have a problem which you solve with k8s, now you have two problems 😛
17:45:58 <truebrain> andythenorth: Find some free map bits!
17:46:50 <frosch123> or add more, people would want 64k biomes anyway
17:50:46 <truebrain> also fixes railtypes!
17:51:03 <frosch123> also consiider standardized biome labels to describe temperature, humidity, wind speed, rock structure; each min/max/average over the year
17:51:18 <truebrain> someone read a backlog 🙂
17:51:41 <frosch123> no, i didn't 🙂 but maybe discussions are predictable
17:51:51 <truebrain> if you didn't, it is scary accurate
17:51:59 <truebrain> if you did, you sound sarcastic 😄
17:52:21 <andythenorth> can't we do the biomes in a separate map array? 😛
17:52:24 <andythenorth> and index into it? 😛
17:53:07 <frosch123> i just extrapolated from rail and road type labels
17:53:18 <truebrain> that was the discussion this morning, yes 🙂
17:54:20 <frosch123> i better don't read the backlog then, what ends up on the wiki is funny enough 🙂
17:58:49 <truebrain> frosch123: people update the wiki?!
17:58:56 <truebrain> I am shocked someone found the edit button 🙂
17:59:52 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
18:00:21 <frosch123> the newgrf wiki is still on mediawiki, we can't hide the edit button that easily
18:10:20 <truebrain> frosch123: I am also not going to migrate it 😛 What a hell was migrating the wiki .......
18:10:53 <frosch123> i regret not deleting the "archive" part
18:11:05 <truebrain> also not deleting the rest
18:13:14 <truebrain> hmm .. how to best store a TileIndexDiffC
18:13:18 <truebrain> let's see if someone else did
18:15:12 <locosage> it may even be more compresseable
18:15:18 <truebrain> no, they are actually different structures
18:15:24 <truebrain> I can cast it to any uint32, but .. hmm
18:15:41 <andythenorth> blank page for me
18:15:52 <andythenorth> I'm the last to edit it, have I crashed the wiki?
18:16:02 <andythenorth> maybe the link is wrong, top result on google though
18:16:44 <locosage> yeah, but you don't really need DIffC, calculating vector just Diff will even require one operation instead of two
18:16:56 <locosage> though you'll still need somethnig for distance
18:17:32 <peter1139> truebrain, looks like andy's crashed it all, maybe you should delete it all and start from blank...
18:17:43 <locosage> oh, btw, I had this weird idea that source_xy can be stored as source_xy - cargo_movement to make it closer to 0 and so more compresseable
18:17:47 <truebrain> peter1139: yes, I like that idea! 😄
18:18:12 <locosage> was trying to think of a way to avoid saving source_xy but got this instead xD
18:18:36 <andythenorth> all I wanted to do was show frosch123 that the spec has hardly any 'missing' 😛
18:18:50 <locosage> saving source_xy in some form seems unavoidable
18:18:58 <truebrain> hmm .. I made a mistake somewhere in caching code .. so it cached a file with a "1" in it
18:19:07 <truebrain> removed the weird file, will purge cache now
18:19:17 <truebrain> there might be more pages
18:19:21 <truebrain> will check that tomorrow
18:20:04 <truebrain> hard reload fixes the page at least
18:21:57 <truebrain> okay, this is perfect
18:21:58 <andythenorth> meh, frosch123 found that gif 😦
18:22:09 <truebrain> this will be used so often now
18:22:12 <frosch123> i hope i used it correctly
18:22:31 <frosch123> though i was not aware tb had not seen it before
18:22:38 <andythenorth> seems you know how to meme yes
18:22:45 <andythenorth> someone should make a TB version
18:22:51 <truebrain> I have either not seen it, or forgot about it; both are possible
18:24:33 *** alfagamma7 has joined #openttd
18:24:33 <alfagamma7> Someone should really update nml section of grfspecs
18:24:57 <truebrain> let that someone be you!
18:25:09 *** Wormnest has joined #openttd
18:25:19 <frosch123> instead of "someone" you could use the "anti-royal 'we'"
18:28:12 <truebrain> you do have your party pants on today, don't you? 🙂
18:29:22 <frosch123> it helps, if you don't read the backlog?
18:29:34 <truebrain> so that is what is different today 😛
18:30:07 <andythenorth> alfagamma7: update to...? 😛
18:30:41 <alfagamma7> I will come prepared next time with the answer, 😁
18:34:24 <andythenorth> alfagamma7: I did add a link to 'string codes' yesterday from nml main page 😛
18:34:46 <andythenorth> generally, there are quite a lot of nml authors, and very few of them maintain the nml wiki 😛
18:37:06 <truebrain> I still can't believe how nice the Help Window is .. it really has everything .. can't believe nielshm spend so much time on getting the UI done etc 🙂
18:38:15 <truebrain> let's test out the release blob on different OSes now!
18:38:36 <DorpsGek> - Update: Translations from eints (by translators)
18:39:37 <andythenorth> yes, although I am playing tanks in multiplayer right now 🙂
18:39:40 <andythenorth> also the link is 404 🙂
18:39:47 <truebrain> remove a few more chars
18:40:28 <truebrain> hmm, COMPILING.md seems missing
18:41:00 <truebrain> feels a bit weird to ship that
18:41:04 <truebrain> but it is linked .. hmm
18:42:13 <truebrain> ha, found a bug .. clicking a link to a directory makes it go BOOM
18:42:56 <andythenorth> I did some basic clicking around the help panel and think it's great
18:43:06 <truebrain> but README etc works?
18:43:11 <truebrain> it could find the content of the files?
18:43:19 <truebrain> does it work for both .dmg and .zip? (sorry if I am overasking)
18:43:51 <andythenorth> everything in the .dmg works
18:43:55 <andythenorth> I am between games 🙂
18:44:00 <andythenorth> they last 7 mins 😛
18:44:22 <truebrain> so I clicked the GNU GPL link in CONTRIBUTION.md
18:44:44 <truebrain> so I read more closely .. "The GDPR notice is adapter from rsyslog"
18:44:48 <truebrain> funny, how reading helps
18:45:48 <truebrain> okay, there are a bunch more dead links in the docs, but all of them are not useful to the player .. release procedure .. compiling .. commit message guidelines ..
18:45:54 <truebrain> guess we just accept them as "broken" 🙂
18:51:12 <frosch123> i would not have recognised that as a dropdown
18:51:28 <truebrain> no, I did neither, but there is a dropdown icon there
18:51:32 <truebrain> so not sure if we can improve that
18:51:35 <truebrain> nevertheless, it is so cool 😄
18:51:39 <frosch123> yes, at the very end 🙂
18:51:41 <andythenorth> makes me think we should explain mods better somehow also
18:51:48 <peter1139> The buttons on left open in a browser, the buttons on the right open in game.
18:51:48 <andythenorth> but one win at a time eh
18:52:00 <peter1139> Cool, but there is nothing to distinguish the which.
18:52:17 <truebrain> do we have an icon for that?
18:52:29 <truebrain> mostly external links have this nice icon to indicate that they are external
18:52:51 <frosch123> there are "navigation history" buttons 😮
18:53:00 <truebrain> in de markdown, blue is external and the rest is not. You need to figure that out first, but after that, it is clear
18:53:21 <truebrain> frosch123: yes! mister n really want all in here 😄
18:53:41 <truebrain> I said earlier today, it is the most complete and prestine PR I have seen in a .. decade?
18:54:07 <andythenorth> I know we're reviewing the nice thing
18:54:16 <andythenorth> but how could this button read better?
18:54:21 <andythenorth> it looks particularly odd now
18:54:25 <andythenorth> 'Check' is a weird verb
18:54:30 <andythenorth> 'Online Content' for what?
18:54:36 <andythenorth> help? manuals? DLC?
18:55:02 <andythenorth> "Download Mods"?
18:55:20 <truebrain> andythenorth: how was the .zip doing?
18:55:33 <frosch123> why does the "community" button link to the "contact" page instead of the "community" page?
18:55:49 <truebrain> it was what the PR did
18:55:54 <truebrain> is there a community page (not on the wiki)?
18:56:05 <truebrain> but links to the wiki
18:56:06 <frosch123> it links to the wiki :p
18:56:12 <truebrain> I guess that is why 😉
18:56:25 <frosch123> is it not allow to link to the wiki?
18:56:30 <truebrain> I will make a community.openttd.org 🙂
18:56:39 <truebrain> I would prefer not to have a deep-link into a wiki, no
18:56:51 <truebrain> OpenTTD cannot change existing releases 😛
18:57:17 <frosch123> either way, please don't add more links to the contact page, people will write more mails 🙂
18:57:34 <truebrain> yeah, I will just put community.openttd.org to the wiki
18:58:01 <frosch123> while you adding more redirects, also add one for bugs?
18:58:04 <truebrain> ```751 char *buf = MallocT<char>(filesize);
18:58:04 <truebrain> (gdb) print filesize
18:58:04 <truebrain> $1 = 9223372036854775807```
18:58:06 <truebrain> weird that it fails ..
18:58:15 <truebrain> frosch123: bugs.openttd.org ?
18:58:20 <frosch123> the "bugs" link goes to the main repo page, not the issue tracker
18:58:36 <truebrain> owh, yes, it is github.openttd.org
18:58:42 <frosch123> ah, so "bugs" already exists, but the PR doe snot use it
18:58:52 <peter1139> bugs. goes to the right place.
18:59:14 <andythenorth> truebrain: macOS zip is fine
18:59:19 <truebrain> andythenorth: thank you!
19:00:04 <truebrain> so for some reason `FioCheckFileExists` says "yup, exists" when opening a folder ... that is a bit odd
19:00:53 <frosch123> looks good, the rest does not pass the bikeshed filter 🙂
19:01:23 <truebrain> will fix both links; leaves the question of: do we need to make it more clear that it is a dropdown, and that the left side are links, the right side is a viewer?
19:02:22 <frosch123> ok, i put the dropdown in the bikeshed category. but if you want to change stuff, i would try putting all the buttons on one row: back, forward, dropdown, wrap
19:02:36 <frosch123> only the real window decorations in the title bar
19:02:44 <andythenorth> I missed the dropdowns
19:02:52 <truebrain> wrap is already part of the title bar in all other windows we have that have that button
19:02:56 <andythenorth> they work though, when they were pointed out
19:03:04 <frosch123> i think the dropdown looks mostly weird because it has the same bounds as the text viewer, so adding stuff on left/right should help
19:03:05 <truebrain> but how does moving the others make more clear that the dropdown is a dropdown?
19:03:27 <truebrain> okay, so `fopen(<folder>)` actually works
19:03:31 <truebrain> that is a bit annoying 😄
19:03:39 <andythenorth> it will always be weird having the dropdown down arrow above a scroll page
19:04:07 <andythenorth> we have an established convention to solve that
19:04:11 <andythenorth> it just uses more space
19:04:15 <frosch123> hmm, oh right... on what files does this browser actually work?
19:04:18 <andythenorth> might not be worth it for bookmarks
19:04:26 <frosch123> does it also resolve links inside readmes of content download?
19:04:46 <frosch123> then validation may be slightly more important
19:05:06 <truebrain> I did not see if that is the case or not
19:05:10 <truebrain> mind taking a look for that?
19:05:42 <truebrain> so `fopen` indeed returns a filepointer when used on a folder .. that is .. shit
19:05:49 <truebrain> we have one link to a folder in our docs
19:06:01 <truebrain> we can just remove that link?
19:06:04 <andythenorth> can I replace the target with malware 😛
19:15:01 *** gelignite has quit IRC (Read error: Connection reset by peer)
19:15:15 <truebrain> if I read the code right, it parses the file as markdown if the file ends with .md
19:16:02 <truebrain> we force BaNaNaS uploads to be readme.txt and license.txt
19:16:13 *** gelignite has joined #openttd
19:18:19 <truebrain> after this PR, the game also allows readme.md
19:18:23 <truebrain> it now only allows readme.txt
19:18:36 <truebrain> so yeah, we should show a popup: we are about to open your browser and navigate to NNN, I guess?
19:18:50 <truebrain> maybe in general, for any link we open
19:21:40 <truebrain> ugly, but is more visible
19:26:11 <truebrain> that is less ugly, much more visible
19:26:40 <andythenorth> let's go with that
19:26:44 <truebrain> ( I can't move the navigation buttons or wrap-text down, as they are always visible .. the ToC only for markdown content )
19:27:17 <andythenorth> the only problem with "Wrap text" there is that it clips 😛
19:27:26 <andythenorth> which is beautiful
19:27:50 <andythenorth> well, strictly it doesn't clip, but the shadow is touching the border
19:28:09 <frosch123> truebrain: iirc bananas1 allowed uploading .md files, though also iirc you checked that no existing package made use of that
19:28:32 <truebrain> if you say so 😛 But we don't allow it anymore 😄
19:28:53 <andythenorth> can our UI framework give more left/right padding?
19:28:57 <truebrain> how do I hide a component in the GUI again .. I always forget ...
19:28:58 <andythenorth> or do we just put spaces in?
19:29:08 <frosch123> the back/forward buttons are visible for non-markdown?
19:29:13 <frosch123> what links are in txt files?
19:29:16 <truebrain> disabled, but they are there
19:29:22 <truebrain> you can go from markdown to textfile
19:29:45 <truebrain> he really went all-out 🙂
19:29:46 <frosch123> so "back" makes sense for txt
19:29:54 <truebrain> so the location makes sense too
19:30:10 <truebrain> now I just need to remember how to hide components in GUIs ..
19:30:25 <frosch123> hiding works via the selection thingie
19:31:04 <frosch123> SZSP_VERTICAL and _HORIZONTAL
19:31:15 <peter1139> andythenorth, shadows touching the border is by design.
19:31:33 <peter1139> Don't make me first up DOSBox again...
19:32:16 <andythenorth> well can we not shadow it then 😛
19:32:59 <truebrain> okay .. so that is the ToC done
19:33:03 <truebrain> now for links .. popup?
19:33:09 <truebrain> seems like the cleanest solution here
19:33:55 <frosch123> there was some disclaimer-do-not-read-click-away thingie for the "search external stuff" button in the content gui
19:34:44 <truebrain> let me try to trigger it 🙂
19:37:13 <truebrain> yeah, that really is over-the-top
19:38:14 <truebrain> okay, so I added a NWID_SELECTION, but now my window isn't as wide as it used to be
19:40:41 <truebrain> ah, can't resize in the width anymore
19:41:29 <frosch123> TextfileWindow::NavigateToFile looks dangerous to me, we should block that for .md files from DLC
19:41:49 <truebrain> what is your worry?
19:41:57 <frosch123> SZSP_HORIZONTAL should make your resizing work
19:42:47 <truebrain> nah, it doesn't work if I show the entry, so that is not it .. I will fiddle on 🙂
19:43:46 <frosch123> currently NavigateToFile does a "basename with PATHSEPCHAR", then it does a slash conversion, and then it does stuff on the # fragment
19:44:01 <frosch123> in that order, it should be the reverse
19:44:41 <frosch123> much like the recent svg exploit, where a rust library properly validated an url, and the underlying C library did not know about fragments and used them as filenames on disk
19:45:18 <frosch123> if you include / or \ after #,. NavigateToFile probably does scary things
19:45:29 <frosch123> i would much rahter block the whole function for untrusted files
19:45:30 <truebrain> ah; that is what you meant
19:45:34 <truebrain> sorry, you were rather cryptic 😛
19:45:45 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
19:46:16 <truebrain> yeah, I am fine with that honestly
19:46:25 <truebrain> but still, making sure `#` is stripped first, sounds like a plan 🙂
19:46:34 <frosch123> redirecting people to external websited, which steal their login cookies is one thing. opening arbitrary local files is another 🙂
19:47:48 <LordAro> peter1139: haven't thought about bzflag is *ages*
19:47:57 *** kuhnovic has joined #openttd
19:47:57 <kuhnovic> KuhnovicviaGitHub: For a half year old rebase there were surprisingly little conflicts
19:49:53 <kuhnovic> No worries. Thank you guys that you want to take the time to playtest (torture) my pathfinder 🙂
19:50:24 <truebrain> frosch123: okay, address the hash stuff, and will work on preventing it from working on NewGRF content etc
19:50:46 <kuhnovic> I will be moving soon so I will not have a lot of time to fix any issues immediately, but I will address them later
19:50:48 <truebrain> now to figure out why my window is smaller all of a sudden .. its weird ..
19:50:58 <truebrain> kuhnovic: no rush; we just want to playtest so we can give input
19:53:30 <truebrain> frosch123: if we allow .md and .txt content to be opened, is there actually an issue?
19:54:41 <truebrain> added `if (!StrEndsWithIgnoreCase(newfile, ".md") && !StrEndsWithIgnoreCase(newfile, ".txt")) return;` after stripping the anchor
19:54:49 <truebrain> we really shouldn't be trying to render anything else 😛
19:56:34 <truebrain> yippie, fixed my width issue
19:56:58 <truebrain> ieuw! TC_BLACK missing? 😄
19:57:40 <truebrain> no, it was just set to be WHITE
20:01:29 <truebrain> truebrain: hmm, I guess that means we should also check for things like `..`, but that is relative trivial
20:03:06 *** gelignite has quit IRC (Quit: Stay safe!)
20:05:55 <frosch123> is it? it sounds rather difficult to me :p
20:06:03 <truebrain> `if (newfile.find("..") != std::string::npos) return;`
20:06:33 <frosch123> is there a reason to support directory traversal? can we also ban / and \ ?
20:06:42 <truebrain> our docs are in `docs/` 😛
20:06:44 <peter1139> Yeah, is there is a library we can import to handle this? :p
20:07:10 <frosch123> if you can navigate down a directory, you usually also want to get back up?
20:07:22 <frosch123> so supporting / or \ while banning .. is weird
20:07:33 <truebrain> back button does that 😛
20:07:41 <truebrain> I am just wondering if it is worth the risk ... I guess not
20:07:50 <truebrain> from a BaNaNaS perspective we don't allow uploading any other files
20:08:01 <truebrain> so in that sense, why bother supporting
20:08:17 <truebrain> okay, will make that you can't click links for content
20:08:20 <truebrain> only for our own files 🙂
20:08:35 <frosch123> both local files and external websites?
20:08:58 <truebrain> external website can be rather useful; one thing that annoys me currently is that I cannot click on links in those files
20:09:03 <truebrain> like .. I have to type them?! 😛
20:09:18 <truebrain> I have to make a popup anyway, for the Help and Manual window
20:09:26 <truebrain> currently it is a bit surprising the left 4 open a window
20:09:40 <truebrain> by having a popup saying: you are about to open blablabla, you solve that
20:09:50 <truebrain> so I guess that means we can leave it for 3rd party content?
20:10:00 <truebrain> (I am not sure I agree with my own line of thought here)
20:11:55 <frosch123> different idea: add some groupboxes around the buttons: "documents" and "websites" or so
20:12:04 <frosch123> hard to find captions though :p
20:12:29 <truebrain> "Website" / "Manuals"
20:12:46 <truebrain> saves me from making another window 😛
20:12:57 <frosch123> well, the manual is on the web :p
20:13:07 <truebrain> and should have spaces around `/`
20:13:13 <truebrain> so "Website" / "Documents"
20:14:14 <truebrain> and change the first button to just "OpenTTD"
20:15:00 <truebrain> well, for tomorrow, enough painting this bikeshed for one day 🙂 Making this PR even better, one color of paint at the time!
20:15:38 <truebrain> wow, I fucked up the community subdomain nicely ...
20:17:26 <frosch123> can there be spaces between "filename.md" and "#fragment"?
20:18:29 <truebrain> after removing `#...`, it has to end with `.md` or `.txt`
20:19:08 <frosch123> ah, right, it tests the positive case, and rejects everything else
20:19:58 <frosch123> not sure about find_last_of/find_first_of for the # :p
20:20:03 <frosch123> urls are so difficult 🙂
20:20:16 <truebrain> they are now documents we write
20:20:25 <truebrain> so either way, it will be fine 😄
20:20:31 <truebrain> but yeah, should be `find_first_of`
20:21:07 <truebrain> I could use libcurl to do this shit
20:21:35 <truebrain> no, really, let's not allow external documents to open links in any way, and it becomes a lot easier 🙂
20:22:10 <truebrain> will finish these things up tomorrow; will be lovely 🙂
20:23:38 <truebrain> we do however need reviewers .. my "count of open PRs" is getting a bit big 😛
20:24:13 <truebrain> now, zzz time; night guys
20:25:26 *** Wormnest has joined #openttd
20:44:31 <peter1139> Is grfcrawler still active? Its T
20:44:39 <peter1139> Its TLS certificate is expired.
20:44:56 <peter1139> Maybe it's not meant to be TLS, but hey, LetsEncrypt etc...
20:54:01 <peter1139> Huh, how have I never noticed the CargoPaymentPool before. What even is it? Why so many possible items...
20:55:41 <peter1139> Hmm, so it allows 1 cargopayment per vehicle.
21:00:00 <andythenorth> what does it do? 🙂
21:01:15 <truebrain> It isba funny class
21:01:48 <peter1139> Doesn't really matter, there are never normally that many items in it. It's just in case :D
21:01:48 * andythenorth tries to please pyflakes
21:01:59 <andythenorth> pyflakes is judging certain choices I made 😛
21:02:02 <andythenorth> and finding them poor
21:04:21 <andythenorth> oh it has found an undefined var used in a BaseException though 😛
21:04:27 <andythenorth> proving that error handling code has errors
21:06:41 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:09:10 <andythenorth> hmm I have a utils module, which also imports a handful of other things but doesn't use them directly
21:09:26 <andythenorth> so they're exposed in the namespace as `utils.foo` etc
21:09:35 <andythenorth> but they live in a different package
21:09:47 <andythenorth> pyflakes does *not* approve of that
21:15:26 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
21:21:02 <_glx_> jfs: I written one a long time ago (not directly usable in regression flow)
21:47:32 <_glx_> ahah and looking at `convert.c` I can see a bug (most likely never triggered) ```
21:47:32 <_glx_> // retrieve the PE header pointer
21:47:32 <_glx_> offset = ((content[0x3f] << 8 | content[0x3e]) << 8 | content[0x3d]) << 8 | content[0x3c];
21:49:21 <peter1139> That's an odd way of writing it, but is it a bug?
21:50:56 <Eddi|zuHause> maybe if there ever will be big endian PE? :p
21:51:14 <_glx_> 3f and 3e are 0 usually
21:51:34 <_glx_> but still they should be shifted more 🙂
21:52:34 <Eddi|zuHause> either that, or there should be some () placed
21:53:44 <_glx_> oh I missed the parenthesis
21:54:26 <peter1139> It's actually a little bit simpler CPU instructions, or would be if you were hand-writing assembly.
21:55:38 <_glx_> yeah actually the code is fine, just hard to read 🙂
23:50:56 *** ChanServ sets mode: +v tokai
23:55:13 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
continue to next day ⏵
