IRC logs for #openttd on OFTC at 2020-02-26
⏴ go to previous day
00:09:25 *** gnu_jj_ has joined #openttd
03:28:20 *** Wormnest_ has joined #openttd
04:50:19 *** tokai|noir has joined #openttd
04:50:19 *** ChanServ sets mode: +v tokai|noir
05:49:58 <DorpsGek_III> [OpenTTD/OpenTTD] spauka opened pull request #8023: Fix #7664: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu4O
05:51:04 <DorpsGek_III> [OpenTTD/OpenTTD] spauka updated pull request #8023: Fix #7664: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu4O
05:53:37 <DorpsGek_III> [OpenTTD/OpenTTD] spauka updated pull request #8023: Fix #7664: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu4O
05:55:30 <DorpsGek_III> [OpenTTD/OpenTTD] spauka updated pull request #8023: Fix #7644: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu4O
07:49:10 *** Smedles_ has joined #openttd
08:00:38 *** ChanServ sets mode: +v tokai
08:41:04 *** Smedles has joined #openttd
09:27:03 *** tokai|noir has joined #openttd
09:27:03 *** ChanServ sets mode: +v tokai|noir
09:34:19 *** y2kboy23 has joined #openttd
09:48:05 *** y2kboy23 has joined #openttd
10:15:29 *** Smedles has joined #openttd
10:38:43 *** Smedles has joined #openttd
10:44:55 *** Smedles has joined #openttd
11:03:08 *** ChanServ sets mode: +v tokai
11:18:27 *** gelignite has joined #openttd
11:44:34 <TrueBrain> so ... I need a savegame with a NewGRF or something in it .. hmmm
11:44:55 *** D-HUND is now known as debdog
11:46:43 <TrueBrain> ha, found one on my disk, w00p
11:49:56 <TrueBrain> okay ... next question .. how the hell do you trigger a CLIENT_CONTENT_INFO_EXTID .. hmmmm
11:51:37 *** gelignite_ has joined #openttd
11:53:01 <TrueBrain> via the console, lol, okay
11:58:09 *** tokai|noir has joined #openttd
11:58:09 *** ChanServ sets mode: +v tokai|noir
11:59:59 <TrueBrain> OpenTTD keeps finding an OpenGFX, and I have no clue where it is on my disk :D Sweet!
12:01:18 <TrueBrain> ah, "Public". URL in docs is wrong, funny :)
12:10:35 <TrueBrain> so the code has this function: "RequestContentList(ContentVector *cv, bool send_md5sum)"
12:10:42 <TrueBrain> which is never called with md5sum=false
12:10:54 <TrueBrain> we have a whole protocol function that is never used :P funny
12:12:17 <TrueBrain> I would expected there was a function that only checks for the downloaded content if there are updates .. but what really happens, is that the whole content list is downloaded, and client-side it is checked if there are updates
12:12:20 <TrueBrain> not what I expected :D
12:19:21 <TrueBrain> what I like about looking back at those protocols (that I and Rb designed, basically, so I am laughing at myself) ... 2 out of the 5 near-identical functions have a count of 16bit, the others of 8bit .. wtf really :P CONSISTENCY PLEASE
12:36:13 *** ChanServ sets mode: +v tokai
12:37:32 <TrueBrain> lol .. the download mechanism is a bit funky ... client: request A and B
12:37:35 <TrueBrain> server: here you have C and D
12:37:48 <TrueBrain> client: sure, I just told the user I got A and B, but I really stored C and D, no problem
12:42:10 <TrueBrain> something I did not know (and clearly once knew): we store all content as gzip-files .. that saves a lot of bandwidth :D
12:44:01 <TrueBrain> anyway, basics of ottd_content are done. Userflow is implemented. Now all it needs it a database and filestorage :)
13:36:28 *** gelignite has joined #openttd
13:37:32 <TrueBrain> Endianess is a bit tricky with content I noticed ...
13:38:10 <TrueBrain> the ID OGFX becomes 0x5846474f and not 4f474658 what you would expect when reading OGFX
13:38:32 <TrueBrain> happens when you abuse char for uint32s :)
13:56:41 <peter1138> Just use IDs in the "reserved" range.
14:00:27 * peter1138 copy & pastes code... except when MonoDevelop crashes repeatedly.
14:03:46 *** ChanServ sets mode: +v tokai
14:22:18 <TrueBrain> hmm .. sometimes it is "Base Graphics", sometimes it is "Base Graphic"
14:22:51 <TrueBrain> same with "Base Sound(s)"
14:22:56 <TrueBrain> all others are singular
14:25:40 * LordAro creates a 477G swap partition
14:26:27 <LordAro> Perfectly Legitimate Reasons
14:26:54 <TrueBrain> I hope it is stored on SSDs :P
14:27:49 <LordAro> TrueBrain: I'd say they should all be plural - "Music" is both singular and plural
14:28:11 <LordAro> peter1138: has MonoDevelop ever stopped crashing repeatedly?
14:28:13 <TrueBrain> NewGRF? AI? AI Library?
14:28:22 <LordAro> NewGRFs, AIs, AI Libraries
14:28:36 <LordAro> unless... what are we referring to here?
14:28:43 <LordAro> i should probably know that first :)
14:29:32 <TrueBrain> the last one has all entries fully singular
14:29:35 <TrueBrain> the first one has a mix
14:30:15 <LordAro> i guess it comes from "Base *" being collections, rather than a single item
14:30:37 <TrueBrain> but the person who wrote the one is the same as who wrote the other (and not me), and did not remember that :D
14:30:51 <peter1138> LordAro, and that is one reason why I'm copying _out_ of MonoDevelop, heh
14:31:15 <LordAro> "Base Graphic" and "Base Sound" is sounds weird to me though ("Music" is already both singular and plural)
14:31:31 <TrueBrain> I dont care how things sound; I hate inconsistencies :P
14:31:43 <TrueBrain> existing musa ini files will have "Base Graphic"
14:31:50 <LordAro> peter1138: i don't think i've tried to use MonoDevelop in ...8 years?
14:32:24 <peter1138> LordAro, I'm a bit behind, I've got VS Code open on the other half of the screen, heh.
14:47:28 <TrueBrain> argh, uniqueids are really terrible ... endianess conflicts++
14:48:46 <TrueBrain> char 'OGFX' is stored in memory as 0x"X", 0x"F", 0x"G", 0x"O" ... so 0x5846474f
14:49:05 <TrueBrain> well, yes, but that is not what I meant
14:49:15 <TrueBrain> so the int value is 1481000783
14:49:20 <DorpsGek> TrueBrain: 1481000783
14:49:43 <TrueBrain> so the folder should be named 5846474f
14:50:25 <TrueBrain> so .. struct.unpack("!I", bytes.fromhex("0x5846474f")) .. yes
14:50:37 <TrueBrain> (and yes, ! as it needs to read it in "BE" order ofc)
14:50:46 <TrueBrain> ugh .. this was wayyyy to complicated for something this simple :P
14:51:06 <TrueBrain> so Python tells me b'XFGO'
14:51:12 <TrueBrain> which is correct :) But looks weird :P
15:43:27 <Eddi|zuHause> just make a function for "grfID('XFGO')" that makes a byteswap?
15:44:10 <TrueBrain> I don't really care how Python represents it. I was more wondering how to get it right :P
15:54:13 <TrueBrain> hmm ... minor flaw in current setup ... by only having the most recent GRF in the GitHub repository .. I have no metadata when fetching entries for savegame compatibility .. hmmmm
16:08:58 <TrueBrain> cool ... except for the minor flaw for savegame compatibility, the new content server works :D Well, that was easy .. :P
16:09:11 <TrueBrain> needs cleaning up and some hardening, ofc .. but .. this is not bad :)
16:30:04 <Samu> why's the cost of a corner tile square root of 2 / 2
16:31:42 <Samu> sqrt(2)/2 ~ 0.7071, so for NPF it's 70, for YAPF it's 71? shouldn't it be equal on both?
16:32:09 <TrueBrain> hmmm .... there are 2 NewGRFs with the same uniqueid .. made by 2 authors ... how could this ever been uploaded :o
16:37:38 <TrueBrain> ah .. one author changed its GRFID .. allowing another different package to upload theirs under the same GRFID
16:37:40 <TrueBrain> that is of course wrong
16:37:52 <TrueBrain> so I found a bug in BaNaNaS. Where can I report it? :D
16:39:44 *** Wormnest_ has joined #openttd
16:41:16 <heffer> Yeah. Please just don't use xcf2png :D
16:43:48 <TrueBrain> that comment triggered me :P And not because of the content of the PR :D
16:46:07 <milek7> in case gimp stays decision is needed on that 2.10 differences
16:46:13 <milek7> like whether it needs investigation or new blending is fine
16:46:38 *** sla_ro|master has joined #openttd
16:46:58 <TrueBrain> this double GRFID in BaNaNaS is a bit of a mystery to me ... I don't fully understand how it happened ..
16:47:08 <TrueBrain> maybe not worth my time atm :D
16:49:01 <heffer> TrueBrain: sorry for triggering :P Fedora has this rule that everything needs to be built from source. I do compare the hashes of the results and so far never had an issue with the generated files being different from what was pre-generated.
16:49:21 <TrueBrain> not generating the PNGs is still "from source" :)
16:49:40 <TrueBrain> but it makes me really happy you validate the results :D
16:50:53 <TrueBrain> as I noticed this can very easily go really wrong :P
16:51:00 <TrueBrain> we might just need to look at a better solution for this
16:53:34 <TrueBrain> hmm .. more BaNaNaS odities .. heightmaps and scenarios have no uniqueid .. it is the id of the row in the database .. rriiiigggghhhttttt
16:54:43 <nielsm> I'll test the osx colourspace thing in a mone
17:07:22 <TrueBrain> openttd-supported-version:
17:07:35 <TrueBrain> you run OpenTTD 1.7.0 .. is this GRF compatible?
17:10:27 <LordAro> poorly named, perhaps
17:10:50 <TrueBrain> what would be better names?
17:11:05 <TrueBrain> especially as no longer have a "svn version", this at least becomes a bit easier
17:13:57 *** tokai|noir has joined #openttd
17:13:57 *** ChanServ sets mode: +v tokai|noir
17:19:58 *** arikover has joined #openttd
17:20:02 <TrueBrain> the comment in the musa example ini seems wrong .. but okay .. not that many people used "svn version", so that shouldn't be a real issue tbh
17:21:30 <TrueBrain> openttd-supported-version:
17:21:30 <TrueBrain> supported-from: 1.3.0
17:21:39 <TrueBrain> openttd-supported-version:
17:21:39 <TrueBrain> not-supported-from: 1.6.0
17:21:43 <TrueBrain> guess that is most clear
17:22:24 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on pull request #8023: Fix #7644: (Cocoa) Manually set colorspace to sRGB https://git.io/JvuDw
17:22:44 <TrueBrain> question is ... is 1.6.0 including beta and RC?
17:23:06 <LordAro> i guess it should do really...
17:23:41 <TrueBrain> so that means versions should always be major.minor.patch
17:23:46 <TrueBrain> as we sometimes introduce new stuff in patches, ofc
17:23:47 <LordAro> i mean, we basically do semver, could just say <1.6.0
17:24:00 <TrueBrain> semver is horrible with betas :)
17:24:04 <TrueBrain> 1.6.0-beta1 < 1.6.0
17:24:04 <LordAro> we *shouldn't* introduce new stuff in patches
17:24:24 <TrueBrain> so if you say >=1.6.0, it is excluding betas
17:24:47 <LordAro> well you'd do >1.5 :p
17:24:54 <TrueBrain> which includes 1.5.1
17:25:03 <TrueBrain> (at least, in pip it does)
17:25:17 <LordAro> maybe i'm forgetting the correct syntax
17:25:43 <TrueBrain> seems semver doesn't define this at least
17:26:31 *** Progman has joined #openttd
17:27:10 <TrueBrain> " >1.7 will allow 1.7.1"
17:27:23 <TrueBrain> so yeah ... welcome to the confusion-land, where nobody is sane, etc
17:27:39 <TrueBrain> all I remember when I had to understand this shit, that it is horrible
17:27:44 <LordAro> ok, so you'd do >=1.6 then :p
17:27:51 <TrueBrain> which is excluding 1.6-beta1
17:28:04 <LordAro> yes, but 1.6-beta1 isn't a thing
17:28:17 <TrueBrain> >=1.6 is excluding 1.6.0-beta1, fine :P
17:28:36 <TrueBrain> but okay, we don't have to do what-ever it was that Python tried to do
17:28:42 <TrueBrain> we need to do something people understand :D
17:28:58 <TrueBrain> so we can make that >= 1.6.0 means including betas
17:29:09 <TrueBrain> as that is most likely what people mean
17:29:33 <TrueBrain> as if I look at it from the use case: 1.6.1 introduces a fix for NewGRF, and my GRF is only compatible from there
17:29:43 <TrueBrain> I mean, including the betas and RCs ofc
17:29:47 <DorpsGek_III> [OpenTTD/OpenGFX] planetmaker commented on pull request #26: Change: Use xcf2png instead of GIMP https://git.io/JvuDH
17:30:08 <TrueBrain> as I would say it is unlikely that something is introduced between beta and RC that breaks my GRF .. it has happened, but it is unlikely
17:30:47 <nielsm> I wonder if we can delete/simplify the blocks that support ancient macos versions?
17:31:02 <nielsm> e.g. checks for 10.5 and 10.7
17:31:03 <TrueBrain> we only support N-2, N-1 and N, right? :)
17:31:19 <TrueBrain> I tried to remove them a few months ago .. someone told me "it works, so why remove it"
17:31:24 <TrueBrain> just as a heads-up :)
17:31:44 <FLHerne> nielsm: 10.5 was (IIRC) the last version that worked on PowerPC hardware, so someone might be upset :P
17:31:55 <TrueBrain> openttd-compatibility:
17:32:05 * FLHerne [finally] doesn't have any working anymore
17:32:12 <nielsm> we also removed support for dos builds and various other weird OS nobody seriously use
17:32:26 <nielsm> and I'm not sure if windows 9x builds are really possible any longer either
17:32:32 <heffer> So I just built 1.10.0 RC1 on Fedora 31. Very smooth. Unfortunately SDL_VIDEODRIVER=wayland didn't work. But I'm probably asking too much :P
17:33:13 <FLHerne> Hm, both those things are older, to the point where it's really impossible to use them for normal stuff
17:33:14 <planetmaker> hm... PRs welcome :P
17:33:34 <FLHerne> But no, I'm not certain it's worth caring
17:33:54 <nielsm> nobody cares whether we support linux 2.2 either
17:34:21 <TrueBrain> lol, some people did this for their NewGRF:
17:34:22 <TrueBrain> openttd-compatibility:
17:34:30 <TrueBrain> the content-service only exist from 0.6 and upwards :D
17:34:30 <nielsm> having a ton of ifdef blocks makes code harder to maintain and bugfixes harder to verify
17:34:36 <TrueBrain> basically, making it not available :)
17:34:48 <planetmaker> TrueBrain, that's means to hide it
17:34:58 <planetmaker> and make it available for savegames which need it
17:35:02 <planetmaker> at the very same time
17:35:02 <TrueBrain> others used "< 0.0.0"
17:35:06 <TrueBrain> which is easier to understand :)
17:36:18 <TrueBrain> openttd-compatibility:
17:36:19 <milek7> does "not-supported-from" is sensible at all? it shouldn't be all backwards compatible with old newgrfs?
17:46:24 <BOB> are there any plans for new Vehicles to openttd. i have the idea for a airlander like vehicle
17:47:45 <planetmaker> @BOB, I'm sure that can be implemented via NewGRF. And anyone could do that :) Or anything particular which cannot be mimiced by heli or plane behaviour?
17:49:59 <BOB> so i need to draw a new texture and put that in the NewGRF library?
17:51:53 <nielsm> if you mean something that's neither road vehicle, train, ship, or aircraft, then no that's probably never happening
17:52:13 <nielsm> but as long as you can place the vehicle you imagine under one of those categories, go hog wild
17:52:32 <TrueBrain> hmm .. either the default font cannot render Korean chars, or something is wrong with UTF-8 :)
17:55:35 <TrueBrain> holy crap, pressing "Check Online Content" is hella slow, for the next 10+ seconds OpenTTD is not behaving nicely :)
17:55:47 <TrueBrain> maybe it shouldn't update the UI for every new entry :D
17:55:47 <LordAro> TrueBrain: list has gotten quite big
17:55:55 <LordAro> though yes, that doesn't help :p
17:56:04 <TrueBrain> or maybe it should only fetch per category
17:56:09 <TrueBrain> or only "what can be update"
17:56:29 <BOB> @nielsm, a airland is a variant on a airplane with a couple of diffent attribute than a aircraft. so you can mamic a aircraft but with full features i think that you must not only you another texture.
17:56:55 <BOB> i have found the NewGRF make documantion
17:57:19 <TrueBrain> LordAro: something for you to fix in 1.11? :D
17:57:43 <TrueBrain> the server needs 0.03 seconds to tansmit the whole list
17:58:00 <LordAro> though not updating the UI for every new entry could reasonably go in 1.10
17:58:06 <nielsm> BOB: also the term is sprite here, OTTD does not use 3D models so there are no textures
17:58:20 <TrueBrain> LordAro: yeah, that sounds like a simple fix. Well, if you have the time, it would make my testing at least a lot easier :D
17:59:12 <LordAro> TrueBrain: you should file a bug
17:59:16 <BOB> @nielsm, with texture i mean the image of the Vehicle i see on screen when playing the game
17:59:23 <TrueBrain> fun thing about fetching the list btw .. when it hits a dependency,it will ask the content server for the info .. which might already being transmitted because it asked for everything ..
17:59:58 <nielsm> BOB yes you're going to confuse everyone else as long as you use the wrong word :)
18:00:32 <BOB> than i will use sprite for it
18:02:29 <TrueBrain> LordAro: as requested ^^. Surprised nobody else had :)
18:09:33 <BOB> @nielsm: i have take a look in the wiki and i wen i will take a look a the TTDviewer the link give my a 502 Bad Gateway error
18:24:52 <Samu> the closer it was getting to 100, the worse it would become, but it's funny that if it's 100, it suddenly becomes good, but 100 means it's equal to the curve45 :(
18:25:06 <glx> weird the cost is higher than 100/600
18:25:39 <Samu> that's not the cost, it's the number of search nodes
18:25:40 <glx> I guess with lower cost it tries more paths
18:26:11 <glx> when the cost is high, paths are rejected earlier
18:27:19 <Samu> but for savegame 1 it meant no difference
18:27:41 <Samu> considering my fix is in
18:27:49 *** ChanServ sets mode: +v tokai
18:30:17 <glx> ignoring the docking cost until close enough makes sense I think, it's similar to roadstops balancing where the cache is trashed in the final approach IIRC
18:41:28 <Samu> i tried to look at how that was fixed, it doesn't mess with penalties, it seems it just stops caching when it's getting closer to roadstops
18:42:14 <Samu> maybe i should look better
18:45:35 *** HerzogDeXtEr has joined #openttd
18:49:47 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on pull request #8023: Fix #7644: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu9Z
18:59:14 <LordAro> cast it into the fire!
19:01:03 *** frosch123 has joined #openttd
19:01:15 <Samu> oh, right, yapf_roads works with segments, not tiles
19:01:51 <Samu> entire different beast than ships
19:04:02 <LordAro> nielsm: and yet you seem to have left support for PPC OSX in?
19:04:41 <LordAro> was PPC OSX even a thing?
19:05:43 <nielsm> I guess I should remove that part too :P
19:10:57 <frosch123> TrueBrain: the performance of the content gui also depends on how much you have downloaded yourself. opening the gui for the first time after launch triggers computing all md5sums of all local scenarios, which can take quite a while if you have a debug build and rb made you download all scenarios for testing somewhen
19:12:39 <glx> also add the time of newgrf scanning before even get access to the main GUI ;)
19:13:18 <frosch123> grf scan also differs a lot between optimised and non-optimised builds
19:14:37 <nielsm> can some of the local content md5 calculations be offloaded to a thread if they aren't already?
19:14:38 <glx> and some of them are even unusable in openttd
19:15:13 <frosch123> nielsm: at some point i wanted to store the md5sum together with file size/date in a ini file
19:15:28 <frosch123> but i dropped that idea when i noticed that optimised build is way faster
19:15:55 <TrueBrain> frosch123: this slowness is without any GRFs ;) So yeah .. lot of things to improve :)
19:16:26 <glx> BTW I think some repos need more people with write access
19:16:51 <TrueBrain> frosch123: a puzzle ... my BaNaNaS idea has a flaw .. when you open a savegame with a GRF you don't have, you can download it .. this can be a really old GRF .. my idea initially was to not put everything in the GitHub repository, but only the recent content .. so how do I get things like "name" in that case? :D
19:16:53 <frosch123> glx: you are owner, you can add people
19:16:59 <frosch123> LordAro: also, get 2fa
19:17:12 <TrueBrain> glx: I think nobody can assign such permissions .. I think it needs to be a bit more specific, which people, what access, to where :D :D :)
19:18:20 <frosch123> glx: just don't add individual/external contributors, use the teams please
19:19:35 <frosch123> TrueBrain: does the name differ between versions?
19:19:43 <frosch123> i thought you can't rename stuff?
19:20:11 <TrueBrain> in the old system it is difficult, but it happens. But I was more thinking about the new GitHub repo, where that will be more common I am sure :)
19:20:40 <TrueBrain> there are also people who will want their GRF removed .. now we can just remove the "current.yaml" in the GitHub repository, and it is gone from listing
19:20:50 <TrueBrain> but there too .. the savegame download will be missing metadata
19:20:55 <TrueBrain> not sure that is a bad thing tbh ..
19:21:07 <frosch123> well, i would suggest to resolve the name only using the 4-byte id. so always show the name of the latest content, even when you download a old version
19:21:29 <TrueBrain> but what if we have that ID also not on record?
19:21:49 <Eddi|zuHause> TrueBrain: a thing that was occasionally missing is, if you release a (beta) newgrf for a new feature, keep the older grf available for the latest (or earlier) release builds
19:22:12 <TrueBrain> Eddi|zuHause: frosch123 mentioned that usecase; something for future extensions
19:22:27 <frosch123> the repo contains meta data per 4-byte id. that is only removed if the content is banned
19:22:43 <TrueBrain> frosch123: I just mentioned the usecase ;) We have people who want to remove their content :)
19:23:00 <TrueBrain> which is fine by me honestly, if they just want to remove their content .. it should be delisted etc
19:23:04 <frosch123> yes, but in that case all is removed, so you do not need to report the name
19:23:06 <TrueBrain> just .. the savegame should still be able to get it :)
19:23:15 <TrueBrain> so when you open such savegame, what name should I report?
19:23:24 <frosch123> none, because you can't download it
19:23:41 <TrueBrain> you can ALWAYS download GRFs based on ID+MD5 if it was once uploaded (and not banned)
19:23:47 <frosch123> i think you mixed up some things
19:23:48 <Eddi|zuHause> TrueBrain: so what's the problem if fetching older versions doesn't have a name?
19:23:49 <TrueBrain> that is the whole point of the savegame content download thingy :)
19:24:08 <TrueBrain> frosch123: I think it is more that we are not talking about the same thing :D :)
19:24:36 <frosch123> a 4-byte content id is either banned or available. if it is available, you know the name. if it is available it may also have a md5sum that is compatible with your version of ottd
19:25:06 <Eddi|zuHause> something isn't right, it's reporting 98% traffic flow...
19:25:08 <TrueBrain> that is not the part I am talking about :)
19:25:28 <frosch123> so, you can always get the name of a 4-byte id, unless the whole thing is blacklisted, and in that case bananas pretends to not know anything about it
19:25:30 <TrueBrain> the normal flow is working just fine :) Let me make a few screenshots ..
19:26:25 *** arikover has joined #openttd
19:26:39 <TrueBrain> that is what I am talking about
19:26:58 <TrueBrain> if you load a savegame which was saved with an old GRF, it allows you to retrieve it
19:27:14 <TrueBrain> this can be an uniqueid (your 4-byte id), which is not publicly listed anymore
19:27:37 <TrueBrain> as example of how the repo will look
19:27:56 <TrueBrain> so I have an ID+MD5, I have the tarball I can serve to the client, that is all working fine
19:28:09 <TrueBrain> but what name etc am I going to give it .. where do I get that metadata from
19:28:46 <frosch123> TrueBrain: you still insist that listing the name publicly implies that something can be downloaded
19:29:12 *** arikover has joined #openttd
19:29:17 <nielsm> one of the promises of banana is that an author can delist an item so it's completely invisible
19:29:30 <frosch123> TrueBrain: the case you make "uniqueid (your 4-byte id), which is not publicly listed anymore" does not exist
19:29:31 <nielsm> unless a would-be downloader knows its id and md
19:29:44 <TrueBrain> frosch123: it does? Where do you get that information from?
19:30:04 <TrueBrain> it is already the case, that we have uniqueids that are not publicly listed, yet do exist
19:30:08 <frosch123> one of use has switched the universe
19:30:10 <TrueBrain> (and are not blacklisted)
19:30:44 <frosch123> TrueBrain: no, there are only cases where the latest version is set to "not compatible with your openttd"
19:31:22 <TrueBrain> this is not completely true; we also delisted items
19:31:42 <TrueBrain> but okay, that is not really important tbh
19:31:46 <TrueBrain> the question is: what do we want to do
19:31:57 <TrueBrain> on that GitHub I linked, I am pretty sure people will make PRs to remove their GRFID completely
19:32:02 <TrueBrain> because they got mad, or what-ever
19:32:09 <TrueBrain> that is a use-case I fully understand and I completely see happening
19:32:22 <TrueBrain> so you say: we are not going to support that usecase, if I understand you correctly?
19:32:48 <frosch123> no, banned stuff is removed completely
19:32:57 <TrueBrain> we are not talking about banned stuff
19:33:02 <frosch123> or are you now trying to do stuff with a public repo?
19:33:26 <frosch123> TrueBrain: "banned" and "deleted" are the same thing
19:33:40 <TrueBrain> they are? Okay .. there is a difference in my opinion, but sure
19:33:40 <Eddi|zuHause> i'm fairly convinced you two are still not talking about the same thing
19:33:43 <frosch123> both are 4-byte ids that can neither be uploaded nor downloaded
19:33:45 <TrueBrain> Eddi|zuHause: I am too :)
19:34:22 <TrueBrain> frosch123: the above usecase I just gave, shows that won't be the case
19:34:28 <TrueBrain> or are you against that usecase?
19:35:06 <TrueBrain> (which is fine, but clearly we are talking cross eachother, so I am just trying to sync up :D)
19:35:12 <frosch123> i understand your example as "auithor want to make new content unavailable because they are mad, but want to keep old stuff available". and i think that usecase is nonsense
19:35:18 <frosch123> either you leave, your you dont
19:35:26 <frosch123> there is no middle "somewhat mad"
19:35:35 <TrueBrain> okay, now I get where you are coming from :)
19:35:50 <TrueBrain> how ever mad an author will be, the ToS allows us to always distribute their content for savegame compatibility
19:35:55 <TrueBrain> so there is where our paths split :)
19:36:12 <TrueBrain> there is banned content, content that was never allowed to be uploaded because of license issues
19:36:16 <Eddi|zuHause> frosch123: the idea is: however mad they got, we can still distribute it to people who have a savegame
19:36:21 <TrueBrain> and there is deleted content, content that was delisted because someone got mad
19:36:25 <TrueBrain> the first case, data got removed
19:36:30 <TrueBrain> the last case, data exists, and is retrievable
19:37:01 <frosch123> Eddi|zuHause: that is wrong, that is betraying them and distributing the content beind their back
19:37:04 <frosch123> i hope we dont do that
19:37:20 <TrueBrain> our ToS is explicitly written to allow that
19:37:24 <TrueBrain> unless your content is banned
19:37:42 <TrueBrain> so that "promise", is not what BaNaNaS v1 does
19:37:44 <Eddi|zuHause> frosch123: you can't revoke licenses just becaue you got mad
19:37:47 <frosch123> TrueBrain: i very much hope that delisted content is not downloadable
19:37:52 <frosch123> otherwise i think we have a big issue
19:38:06 <TrueBrain> it is, that is what I am saying :) And no, we have no issue, it is explicitly written to be like that
19:38:30 <TrueBrain> basically, the ToS says: players come first, unless a license has been broken
19:38:34 <frosch123> i was convinced that it is explicitly written to not do that, and i still hope you are wrong
19:38:53 <Eddi|zuHause> i'm gonna side with TrueBrain on that
19:39:16 <TrueBrain> basically, in 2007 or 2008, the issue was people wanted their content gone, as they were mad
19:39:24 <TrueBrain> and that would invalidate 20+% of the people using NewGRF savegames
19:39:31 <TrueBrain> a LONG discussion followed
19:39:34 <TrueBrain> and the ToS got changed
19:39:43 <TrueBrain> to allow content, once uploaded, to always distribute, under conditions
19:39:51 <TrueBrain> basically, OpenTTD get a distribution license
19:40:07 <TrueBrain> the conditions are simple: only based on uniqueid+md5, the content can be made available
19:40:19 <TrueBrain> in other words: it only works for existing savegames
19:40:27 <TrueBrain> so if you get a savegame for 10 years ago, it will load in OpenTTD
19:40:28 <frosch123> the tos was never changed, it was extended for gs at some point, but there is only one version of the tos
19:40:42 <frosch123> you may remember drafts before bananas was started
19:40:45 *** tokai|noir has joined #openttd
19:40:45 *** ChanServ sets mode: +v tokai|noir
19:40:45 <glx> except for banned content that should not have been uploaded
19:41:44 <TrueBrain> frosch123: not really a relevant discussion; the initial ToS did not allow for this behaviour, and now it does :P The timeline is not really relevant tbh
19:42:34 <frosch123> i think we live in different universes
19:42:50 <TrueBrain> well, what I described is what is currently implemented :)
19:43:06 <TrueBrain> I am fine changing it etc etc
19:43:13 <TrueBrain> I am just moddeling around the current system
19:44:21 <TrueBrain> styling got lost :(
19:45:28 <TrueBrain> anyway, I get now why we didn't understand each other, we work from a different set of rules :D
19:45:54 <TrueBrain> - You grant the OpenTTD team to retain older versions of your content for the purpose of loading save games with said older version.
19:45:54 <TrueBrain> - You grant the OpenTTD team the rights to distribute your content from a central server when specifically asked for it by its unique identifier and MD5 checksum.
19:46:01 <TrueBrain> these two rules is were the difference come from, I guess
19:46:46 <frosch123> i checked svn, in my universe the tos have been unmodified since the start of bananas in 2009
19:46:56 <TrueBrain> frosch123: again, timeline is really not that relevant
19:46:57 <frosch123> if your universe has a different svn, i cant help
19:47:05 <TrueBrain> if this is the only ToS you have ever seen, even better :)
19:47:07 *** arikover has joined #openttd
19:47:14 <TrueBrain> I am not sure why that became a focus point :)
19:47:46 <frosch123> because we disagreed on what was decided how in the past, and i see my understanding backed up by vcs
19:47:52 <TrueBrain> so you say: if a user wants his content removed, savegames that used the GRF won't be able to download the content anymore, right?
19:48:05 <frosch123> so, unless you hacked the vcs, i have no idea where you are coming from
19:48:23 <frosch123> TrueBrain: yes, keep all, or remove all. no middle ground
19:48:43 <TrueBrain> okay. As that is not what is currently implemented, so the question is there: do we want to change the current implementation?
19:48:48 <frosch123> the author can always set stuff "not available for ottd < 0.3", but that does not delete stuff
19:48:53 <frosch123> it just makes the gui hide it
19:49:08 <frosch123> TrueBrain: i am convinced it is implemented like that
19:49:12 <TrueBrain> it is not the GUI that hides it; it is the ottd_content that stops telling you about it
19:54:00 <frosch123> in case you forgot, the website lists more stuff than ottd_content
19:54:23 <TrueBrain> I know :) But you cannot download from there, only the most recent
19:54:33 <TrueBrain> which is a whole different can of worms
19:55:05 <TrueBrain> yeah, we have a GRF of 20 that are simply delisted in the database
19:56:27 <TrueBrain> okay, so I guess the question is: the ToS allows us to keep the content of "deleted" GRFs .. if we really delete the content, it means we break savegames. Besides if we think that is currently done or not (as they are rare events far apart), do we agree that is a good idea?
19:57:09 <frosch123> the original question was "how to get the name", right?
19:57:27 <TrueBrain> yeah, and that is split in two cases: what if there is no metadata, and what if we have metadata of a newer version
19:57:40 <TrueBrain> the latter we agree'd on I believe, just use the metadata of the newer
19:57:51 <TrueBrain> so that leaves a few entries where I need an answer for the first :)
19:57:57 <frosch123> i suggest that "no metadata" only applies to "never uploaded" or "banned"
19:58:21 <TrueBrain> so we are not going to accept any PR that removes a folder, basically?
19:58:29 <frosch123> and i suggest to just always report the name for the latest content
19:59:08 <frosch123> TrueBrain: for me removing metadata implies removing all data
19:59:17 <frosch123> it makes no sense to only delete some data
19:59:46 <frosch123> keeping the metadata available does not imply that there is a "latest" for download
20:00:01 <TrueBrain> the times I processed a request to remove entries from BaNaNaS, with concent of the authors btw, we always just flipped a bit to "not list it" .. this is of course often considered "deleting", so that is a bit tricky wording :D
20:00:20 <TrueBrain> okay, that is a bit of a change of the current setup, but I can work with that
20:00:35 *** arikover has joined #openttd
20:00:42 <frosch123> TrueBrain: if you did that, i am quite sure that was not the authors intention
20:00:57 <TrueBrain> for the cases I handled, and I think I did all, it was
20:01:07 <TrueBrain> as I delibrately ask for that (pointing to the ToS)
20:01:33 <TrueBrain> I don't like listing files if that is really against the author's wishes; but most of the time they understand what the consequences are, and allow "delisted" content
20:01:43 <TrueBrain> the only exception ever was of course mb, but his content is banned
20:02:05 <frosch123> still, for me "no metadata"="banned"
20:02:38 <Eddi|zuHause> from a client's perspective, yes
20:02:56 <TrueBrain> and there is a minor change in the new idea .. as currently we have all the metadata, just not public
20:02:59 <TrueBrain> and I want to make that public
20:03:15 <TrueBrain> so basically, in every folder there will be an "info.yaml" and a "currentl.yaml" or what-ever
20:03:21 <TrueBrain> the first with the metadata, the second with download information
20:03:27 <TrueBrain> you can remove the second
20:03:31 <TrueBrain> never the first (unless banned)
20:03:45 <frosch123> yes, i agree with that
20:03:49 <Eddi|zuHause> do you have a GDPR form of everyone who ever uploaded anything to bananas? :p
20:04:06 <frosch123> info.yaml remains available forever, unless banned
20:04:37 <TrueBrain> in essence that is the same as we do now, so that makes things easier :)
20:04:39 <frosch123> Eddi|zuHause: both tb and i had plenty of training on gdpr
20:05:06 <TrueBrain> I already had to process a GDPR request for OpenTTD .. it was horrible, as the user could not identify himself ..
20:05:08 <frosch123> (and not for ottd :p)
20:05:30 <TrueBrain> (forgot password, email was no longer working, spelled the name wrong, ...)
20:05:57 <frosch123> well, if they can't spell their name :)
20:06:17 <Eddi|zuHause> TrueBrain: request as in "please delete all my data"?
20:06:47 *** Wormnest_ has joined #openttd
20:07:36 <TrueBrain> okay, frosch123 , that was a very long conversation to figure out where we diverged :D But I think in the end we meant the same, but the wording made a disconnect :P
20:08:01 <TrueBrain> I suggest we also make this in the HTML like that: either you request your shit to be banned, or you simply "deactivate" your stuff
20:08:10 <frosch123> LordAro: you are the first one who is in every team except the bot team :p
20:08:47 <TrueBrain> and all that still without 2FA .. he is just asking to being target, isn't he? :D
20:09:12 <frosch123> TrueBrain: as said a few days ago, it would be fancy if the "available for new users" would be a checkbox in front of every version, instead of just "latest"
20:09:22 <frosch123> but can also be added later
20:10:40 <TrueBrain> as I guess that is now somewhat easier to do
20:11:01 <TrueBrain> as you have a uniqueID, an in there you have a few YAML files which state the files available for download ("listed" in the UI)
20:11:16 <TrueBrain> but they will all have the same description etc
20:11:28 <TrueBrain> well, we can make it so you can overrule that
20:11:45 <frosch123> yes, every content has a readme and changelog in the package. info.yaml is global for all
20:11:46 <TrueBrain> hmm ... that would make a fun export, tbh
20:12:33 <TrueBrain> currently I only exported the "latest"
20:12:41 <TrueBrain> but I can also export all versions, just marked as "delisted"
20:13:00 <TrueBrain> not sure authors would appreciate that
20:13:33 <TrueBrain> as it would also show all their typos :)
20:13:33 <frosch123> well, unless you give me examples in private, i still think that people expected "delisted"="banned"
20:13:58 <TrueBrain> haha, sorry, I am trying to find a good word for it, but I was refering to: you have v1, v2, v3, v4
20:14:04 <TrueBrain> only v4 shows up in the UI ("listed")
20:14:09 <TrueBrain> v1, v2 and v3 are "delisted"
20:14:18 <TrueBrain> bad word, hence the quotes .. looking for a better one
20:14:29 <TrueBrain> "available to new users" is what you meant with that too I guess, but "new users" sounds wrong :)
20:14:36 <frosch123> "available for savegames", "available for new games"
20:15:09 <TrueBrain> and I now found 2 usecases to have more than 1 available for new games
20:15:15 <TrueBrain> 1) 32bpp and 8bpp packages
20:15:29 <TrueBrain> 2) v1 for >= 0.6 < 1.10, and v2 for >= 1.10
20:16:02 <frosch123> yep, and v3.alpha for >= 1.11 (nightly)
20:16:14 <TrueBrain> can we see nightlies atm?
20:16:29 <TrueBrain> does their version differ? (honest question)
20:16:33 <frosch123> TrueBrain: nightly is 1.11, stable is 1.10
20:17:00 <milek7> uh, gimp loads all plugins on every startup
20:17:02 <TrueBrain> okay, so that means I will export all versions that ever existed for all GRFs on file
20:17:15 <TrueBrain> and put them on "available for savegames" for all but latest
20:17:21 <frosch123> current bananas allows entering a svn revision, but that is deprecated :)
20:17:28 <TrueBrain> yeah, I found out today :P
20:17:41 <TrueBrain> I also found out we have 2 packages with the same uniqueid written by 2 different authors ..
20:17:44 <TrueBrain> really no clue how that happened
20:17:56 <milek7> and it turns out one of them crashes when launching parallerly with -j12..
20:17:57 <TrueBrain> one of the 2 authors ragequit, so it is fine
20:18:33 <TrueBrain> I was hoping this GitHub repository would only have a single file per uniqueid .. having this whole history might be hard to read ..
20:18:43 <Eddi|zuHause> can't be the first time we have a GrfID clash
20:18:45 <TrueBrain> well, I can put it in one file, I guess
20:18:53 <TrueBrain> Eddi|zuHause: the upload should refuse it
20:18:56 <TrueBrain> that is what confuses me
20:19:34 <frosch123> TrueBrain: so, if content is banned in the future, the metadata remains in the git history?
20:19:50 <TrueBrain> hadn't thought about that, tbh
20:20:17 <TrueBrain> the MD5s are not readable, so I guess that isn't really an issue?
20:22:33 <TrueBrain> that works, I guess
20:23:00 *** andythenorth has joined #openttd
20:23:53 <frosch123> TrueBrain: isn't there some kind of upload timestamp?
20:24:10 *** Smedles has joined #openttd
20:24:13 <TrueBrain> that sounds like sane metadata to add :D
20:25:43 <TrueBrain> question is, would an author want to have a different description or something per version?
20:25:48 <TrueBrain> or is it okay to have the same for all
20:26:17 <TrueBrain> for the 32bpp vs 8bpp I can imagine the names should be different
20:26:24 <TrueBrain> OpenGFX (high-res) / OpenGFX (low-res)
20:26:47 <frosch123> TrueBrain: info.yaml is only for new content. descriptions of old content are inside the package as readme etc
20:27:00 <TrueBrain> there isn't much metadata inside packages, sadly
20:27:06 <TrueBrain> only readme, license and changelog
20:27:18 <TrueBrain> and they aren't shown before you download the package
20:27:21 <frosch123> and content id for heightmaps and scenarios
20:27:58 <frosch123> TrueBrain: thing is, bananas descriptions and readme in the packages differ, already today
20:28:10 <TrueBrain> a readme should have a lot more text :P
20:28:44 <frosch123> TrueBrain: metadata may need "license" per version though. people are upset when they cant see the license in advane before downloading
20:28:50 <frosch123> and uploading new versions can change the license
20:28:55 <TrueBrain> I guess my lowres/highres example would have different versions .. 0.5.5-lowres and 0.5.5-highres .. hmm
20:29:51 <TrueBrain> so a global.yaml, and a yaml per version .. where you can overwrite any value from global.yaml
20:30:24 <andythenorth> did anyone approve glx nml fix yet? o_O
20:30:31 * andythenorth still reading very long logs today
20:31:19 <frosch123> TrueBrain: compatibility: { "master": ["0.7.0", None], "jgrpp": ["1.2", "2.0"] }
20:31:43 <TrueBrain> frosch123: I like that idea, just we currently have no way to tell that to the server or client
20:31:55 <TrueBrain> but it becomes more clear that we need to revise the protocol anyway :D
20:33:06 <TrueBrain> I really think that is a good idea, to allow patchpacks to work in the common flow too .. hmm
20:33:15 <TrueBrain> btw, "min"/"max" is very confusion we found out today
20:33:20 *** supermop_work has joined #openttd
20:33:22 <TrueBrain> hence the >= and <= stuff :)
20:34:09 <andythenorth> wow much to read :)
20:34:29 <andythenorth> are valid DMCA takedowns covered by 'author got mad'?
20:34:54 <TrueBrain> a DMCA takedown is fine, but that only works if we are in violation of a license ;)
20:35:22 <TrueBrain> truth be told, we never cared WHY you wanted to remove content, we always tried to find a dialog what it really is the author wants
20:35:26 <TrueBrain> no matter what the source was
20:46:10 <frosch123> i still miss a timestamp per version, i do not fancy using the commit date for that
20:47:23 <frosch123> content type (baseset, newgrf) is a directory?
20:47:56 <frosch123> license is still complicated
20:48:20 <TrueBrain> especially if they are "custom"
20:48:20 <frosch123> it can be different per version, and i think that licenses were actually changed in the past
20:48:40 <TrueBrain> yeah .. but we support changing now
20:48:53 <TrueBrain> but if you select "custom", you still can't see it in the UI
20:49:18 <frosch123> do you still want to support "depenencies"? iirc they were mostly broken
20:49:38 <frosch123> hmm, though i guess ai/gs really need them for libraries
20:49:47 <TrueBrain> the UI of the webinterface is mostly broken .. but not sure if they are useful
20:50:10 <TrueBrain> I hate the format they are in .. content_type:uniqueid:md5
20:50:20 <TrueBrain> but at least it is a solid reference to something unique
20:51:21 <TrueBrain> okay, in this format it is also reletive simple to make a webpage for to edit this
20:51:26 <TrueBrain> so I am fine with this format tbh
20:52:15 <frosch123> using the "version" and filename is troublesome though
20:52:38 <frosch123> restricts characters, and stuff like version "authors" :p
20:52:49 <frosch123> a single file is better for escaping
20:53:25 <frosch123> ah, though you include the version also inside the file
20:53:32 <frosch123> so maybe just number them consecutively?
20:53:34 <TrueBrain> yeah, filename is irrelevant, basically
20:53:48 <TrueBrain> and put them in a folder to make it more clear
20:53:48 <LordAro> a file per version seems a bit much, perhaps?
20:54:03 <TrueBrain> you know some people have TONS of versions, right? :D
20:54:03 <frosch123> oh, how about using the date, like nightlies?
20:54:18 <LordAro> TrueBrain: that's why i thought a single file might be easier :p
20:54:30 <TrueBrain> frosch123: what do you mean?
20:54:42 <frosch123> using the timestamp as the filename for the version files
20:54:50 <TrueBrain> but you can edit them
20:55:01 <TrueBrain> or you just mean "created" date?
20:55:10 <frosch123> upload date of the content
20:55:24 <frosch123> yes, "created" is "upload date"
20:55:50 <frosch123> i also thought of the "date" in the file as created/upload date. did you think of "last modified"?
20:56:08 <TrueBrain> I hadn't thought about it :P
20:56:11 <frosch123> because i would be fine with "last modified" coming from git
20:56:12 <TrueBrain> now I renamed it to upload-date
20:56:37 <TrueBrain> LordAro: mainly, this is easier to process for a webpage (different files)
20:56:53 <frosch123> upload-date is the date of the content, i do not really care when someone changes the version requirements
20:56:54 <TrueBrain> for the manager part
20:57:15 <TrueBrain> the date the tarball got uploaded, yes
20:57:19 <TrueBrain> suggestions for a better name?
20:58:11 <TrueBrain> hmm .. maybe 2 folders: savegame-only and new-games
20:58:17 <TrueBrain> a bit more friendly for the human in charge
20:59:00 <frosch123> bad idea, that implies moving files in the repository
20:59:09 <LordAro> TrueBrain: yeah, true
20:59:40 <TrueBrain> I am sure we are missing more fields .. but time will tell
20:59:44 <TrueBrain> any other suggestions/comments?
20:59:56 <frosch123> i have models.py opened :)
21:00:06 <LordAro> wait, that's already there
21:00:13 <TrueBrain> except custom license
21:00:16 <TrueBrain> still not happy about that
21:00:25 <TrueBrain> but no way to push them to the client anyway, so meh
21:00:35 <frosch123> TrueBrain: yeah, "licence/license/lisence/lisense" spelling, i have no idea what is what
21:01:22 <frosch123> TrueBrain: dependencies are per version?
21:01:31 <TrueBrain> "In American English, license is both a noun and a verb, and licence isn’t used. For example, one who is licensed to drive has a driver’s license. In all the other main varieties of English, licence is the noun, and license is the verb. So, for instance, one who is licensed to perform dental surgery has a dental surgeon’s licence."
21:02:00 <TrueBrain> litterly never knew there was a difference
21:02:37 <frosch123> oh, i know that there is a difference between licence/license, safe/save, proof/prove, but i never know which is which
21:02:43 <TrueBrain> frosch123: hmm .. dependencies are per version I think yes
21:02:57 <TrueBrain> yeah, it has to, especially for AIs
21:03:14 <TrueBrain> well, in fact, it should not be in global ..
21:03:50 <frosch123> also, i never use yaml... does yaml have save string escaping?
21:04:36 <frosch123> can you put \n \t : | " \ in the description?
21:04:37 <TrueBrain> my client escaped that
21:05:01 <TrueBrain> after |, you can do what-ever
21:05:59 <TrueBrain> I exported the current DB of BaNaNaS .. I was happy to see quoting works in YAML :D
21:06:05 <TrueBrain> people do some stupid shit in their names
21:06:07 <frosch123> TrueBrain: i would expect yaml to eat leading/trailing whitespace, but i wouldn't know what other stuff it eats
21:06:19 <TrueBrain> in yaml you have 2 ways to do things
21:06:30 <andythenorth> hmm this gui2x crap in ogfx
21:06:32 <TrueBrain> the first, it will strip, and do more shit
21:06:34 <andythenorth> I'd better implement it
21:06:36 <TrueBrain> the second, it does everything you expect
21:06:37 <andythenorth> not sure how it works
21:07:01 <TrueBrain> LordAro: yeah, use yaml.safe_load, always, everywhere :D
21:07:39 <LordAro> yaml.safe_load(TrueBrain)
21:07:47 <TrueBrain> frosch123: even UTF-8 works btw :)
21:07:55 <andythenorth> hmm some alternative_sprites thing
21:08:31 <TrueBrain> still on the fence if I should export all history of all GRFs that are currently available-for-new-games
21:08:34 <andythenorth> I'll have to draw them as well
21:08:40 <TrueBrain> pretty sure it should be fine for 99% of the cases
21:08:40 <andythenorth> it's lolz drawing sprites I'll never ever use
21:08:59 <andythenorth> in a style I haven't learnt to do :P
21:09:34 <frosch123> TrueBrain: if you only export stuff available for new games, what does the history include more that is available now?
21:09:52 <TrueBrain> I am pretty sure more than one users meant to do what 0.5.4 and 0.5.5 in my example do
21:10:00 <TrueBrain> but currently you can have only 1 "available-for-new-games"
21:10:10 <TrueBrain> but I guess they can correct that if they want to
21:10:36 <TrueBrain> so basically, it would make it a bit easier for authors
21:10:59 <TrueBrain> hmm, "savegames-only" is basically "archive" :D
21:11:12 <glx> [20:30:24] <andythenorth> did anyone approve glx nml fix yet? o_O <-- not yet
21:11:30 <TrueBrain> now I come to think of it, pretty sure andythenorth doesn't want the world to know that he made 20+ versions of Iron Horse 1 :P
21:11:48 <andythenorth> I want that on my wikipedia page
21:11:57 <andythenorth> can you just hide the brown bag releases, thanks
21:11:57 <TrueBrain> frosch123: the other benefit would be that I know from the repository all MD5s that are available .. now I have to poll the FS to know
21:12:00 <frosch123> TrueBrain: there have been like 70 versions of nuts
21:12:09 <TrueBrain> but yeah ... lets keep it on the safe side :)
21:12:46 <andythenorth> do I have to put the spritenumbers in the ogfx spritesheets?
21:12:51 <TrueBrain> and I like keeping histories .. I never understand why people want to hide stuff .. meh .. ookkkaaaayyyyyyy, I will only do "current" :P
21:13:06 <frosch123> TrueBrain: what is the current status on storing the md5sums? last status i know: store encrypted
21:13:23 <TrueBrain> and I am still open for suggestions :D
21:14:06 <TrueBrain> it is the only piece of information that should be hidden from view
21:14:56 <TrueBrain> come to think of it ... I don't really need to encrypt, do I? I can just hash the hash .. as the user will always give me the md5 ..
21:15:13 <TrueBrain> so if I hash the hash ... with that second hash you cannot request content
21:15:28 <frosch123> you can truncate the md5sum to 8 chars :)
21:15:29 <TrueBrain> well, not true .. on the CDN you can request the download .. meh, nevermind
21:15:51 <TrueBrain> minor chance on collision, but .. minor
21:16:18 <frosch123> i kind of like that, that allows you to locate the right file if you know the md5sum
21:16:23 <TrueBrain> hmm, no, same issue: ottd_content returns HTTP URLs to the CDN
21:16:40 <TrueBrain> hmm, no, I can of course do a * lookup
21:16:54 <TrueBrain> so I can gather the full MD5 by looking on the CDN (as I can list files)
21:16:59 <TrueBrain> from the outside you cannot
21:17:11 <TrueBrain> but what do we do on hash-collision?
21:17:20 <TrueBrain> well ... hash[0:8] collision?
21:17:31 <frosch123> reject the upload :)
21:17:39 <TrueBrain> what can I user do to fix it?
21:17:58 <frosch123> wait... didn't you say ealier that you want to offer both 8bpp and 32bpp files?
21:17:58 <Xaroth> how likely is that to happen?
21:18:11 <TrueBrain> frosch123: I have been told OpenTTD supports that
21:18:13 <dwfreed> you can rainbow table the hashed hash
21:18:17 <TrueBrain> and they return the same md5 ... hmm ..
21:18:19 <frosch123> TrueBrain: yes, but they have the same checksum :)
21:18:21 <TrueBrain> that is a problem :D
21:18:55 <frosch123> well, just means that you have to include the "8bpp/32bpp" in the filename as well
21:19:03 <glx> ah yes md5 is not done on the full newgrf
21:19:13 <frosch123> so you store "newgrf/grfid/md5sum_8bpp.tar"
21:19:22 <TrueBrain> how do I know with what to postfix?
21:19:41 <TrueBrain> but, how files are stored on disk can be different
21:19:48 <TrueBrain> for example, I can store it under the sha256 name
21:20:55 <frosch123> the folder contains some more py files to extract meta data
21:21:03 <frosch123> like dependencies of scenarios
21:21:10 <TrueBrain> what does this do, sorry?
21:21:28 <frosch123> you give it a grf file, and it gives you md5, 32bpp yes/no, trains yes/no, ...
21:22:05 <frosch123> savegame.py gives you md5 of savegame and required newgrf and their md5sum
21:22:26 <TrueBrain> musa had some code for that too
21:22:30 <TrueBrain> just .. really ugly :D
21:23:07 <TrueBrain> possibly not in the first version I am building, but I like that we can just fill in the dependencies for the user .. that is a very nice QOL improvement
21:23:24 <TrueBrain> I like that frosch123 , more meta-data :)
21:23:30 <TrueBrain> what else do you have in BaNaNaS 2? :D
21:23:45 <glx> for me it's sometimes a pain to find my way in nml source
21:24:00 <frosch123> TrueBrain: i listed the best parts already :/
21:24:08 <TrueBrain> well, they are useful, so tnx for that :)
21:24:25 <TrueBrain> okay, so what if we store the files under their sha256 of the full tar
21:24:33 <TrueBrain> and store in the yaml files the first 8 bytes or so
21:24:46 <TrueBrain> we can avoid collision that way easily (as we create the tar)
21:24:59 <frosch123> how does ottd request data?
21:25:24 <frosch123> what do you try to solve with another checksum?
21:25:24 <TrueBrain> 4 methods: list content-type, give based on internal id, give based on uniqueid, give based on uniqueid+md5
21:25:40 <TrueBrain> well, another checksum is always unique, so 32bpp and 8bpp aren't an issue
21:25:44 <TrueBrain> they will have different checksums
21:25:48 <TrueBrain> no disucssion there :)
21:26:25 <TrueBrain> the main issue is the lookup of uniqueid+md5
21:26:25 <frosch123> didn't you want to use the filelisting to get the md5sums?
21:26:38 <TrueBrain> yes, but you pointed out that 32bpp and 8bpp have the same md5 :)
21:26:56 <frosch123> that's why i suggested md5sum_8bpp.tar
21:27:06 <TrueBrain> yeah, but that is more complicated than I would like
21:27:10 <frosch123> and gave you the script to figure that out (which can be added later)
21:27:32 <frosch123> well, your choice, i consider it a lot easier than adding another checksum and conversion tables between checksums
21:27:46 <TrueBrain> conversion tables? who talked about that :)
21:28:02 <frosch123> ottd gives you a md5sum, and you need the matching sha256
21:28:48 <TrueBrain> hmm ... do we need to send the md5sum to the client, I wonder ..
21:29:25 <TrueBrain> so even the partial md5sum is not going to fly
21:30:13 <TrueBrain> "normal" users, I like that wording :D
21:30:49 <TrueBrain> keeping a secret in a public project is always difficult ... "we have this identifier, but we don't want you to know about it"
21:31:27 <TrueBrain> so encryption is simply easiest
21:31:51 <frosch123> i liked the first-8-chars of md5sum
21:32:01 <frosch123> makes it user-readable/debugable
21:32:04 <TrueBrain> the problem is that for 1500 files I have to ask the S3 backend what the full md5sum is
21:32:15 <TrueBrain> that makes it pretty expensive
21:32:26 <frosch123> how often do you need to do that?
21:33:06 <TrueBrain> on reload, was my current idea
21:33:17 <TrueBrain> so that would mean adding caches
21:33:26 <TrueBrain> meh .. it makes everything more complicated ... (encryption too btw)
21:33:30 <TrueBrain> I just want something simple :P
21:34:24 <nielsm> I'd suggest adding caches for the most common result sets at least
21:34:26 <frosch123> how about, let dorpsgek have a private fork of the repository, and store the md5sums there in a separate yaml?
21:34:42 <TrueBrain> frosch123: I was considering something like that too :P
21:35:18 <frosch123> 20170101-1200.yaml is public 20170101-1200-secret.yaml is only in private fork
21:35:19 <TrueBrain> I agree with you that for debugging it would be nice if a partial of the md5sum is in the yaml, as that makes it easier to spot weird things
21:35:37 <TrueBrain> it does come with a risk .. but 32bit is still a huge space to scan
21:35:40 <TrueBrain> so that is fine by me
21:36:02 <TrueBrain> and I need somewhere (tm) that maps that partial to a full
21:36:08 <TrueBrain> and also tells me where on disk those files are
21:36:30 <frosch123> TrueBrain: the md5sum only needs to be unique per uniqueid, not globally, right?
21:36:42 <TrueBrain> so collision chance is near-zero (just not zero :P)
21:36:42 <frosch123> so, 32bit is plenty
21:37:01 <TrueBrain> on the webpage, I would like to have download links too
21:37:07 <frosch123> you can't even mess with other people and deliberately construct files with md5 that match their files
21:37:12 <TrueBrain> but we have to consider how scrapers go about that
21:37:50 <frosch123> website can list the file without md5, and redirect on request
21:38:18 <TrueBrain> not sure yet what is best; just something to keep in mind
21:39:12 <TrueBrain> so you upload a file, the partial md5 is written in the yaml, the file is send to the CDN, and somewhere else is written the partial to the full
21:39:57 <TrueBrain> it is like we need a database!
21:40:27 <LordAro> turns out git repos aren't perfect for everything? :p
21:40:37 <TrueBrain> not for secrets, strangely enough
21:41:18 <TrueBrain> tempted to simply write it in an AWS DynamoDB
21:42:15 <TrueBrain> on the other hand ... this data doesn't change often, and if it changes, there are only things added
21:42:19 <TrueBrain> modifications don't happen
21:42:29 <TrueBrain> so .. why not just a single file on the S3 storage
21:42:39 <TrueBrain> uniqueid,md5-partial,md5
21:43:23 <TrueBrain> and if we ban stuff, this list doesn't need updating .. as it only contains hashes
21:43:32 <TrueBrain> lol .. sometimes it takes a bit of time to dumb down a solution :P
21:44:04 <dwfreed> I'm still trying to figure out what problem you're trying to solve, and I really don't feel like reading 10 screens of scrollback to find it
21:44:24 <TrueBrain> we don't want the full md5sums in the GitHub repository
21:44:25 <nielsm> I think he's trying to avoid using a database for a problem that really wants a database
21:44:49 <dwfreed> TrueBrain: what problem does that solve, though?
21:44:57 <TrueBrain> nielsm: yes, that is very true :) The whole GitHub repository is used as database :) But that means everyone can modify it, instead of a select few :P
21:45:11 <TrueBrain> dwfreed: you wanted to know the problem we are solving! Not why we are solving it :P
21:45:21 <TrueBrain> no, basically, BaNaNaS database is going to a GitHub repository
21:45:42 <TrueBrain> we don't want to store MD5s there, as that means it is very easy to download older versions of a GRF
21:45:55 <TrueBrain> and tha tis one promise we made to GRF authors, that it is not possible to do that without knowing the MD5 to start with
21:46:19 <TrueBrain> so: if you have a savegame with an old GRF, it is fine that you can download it
21:46:32 <TrueBrain> but you should not be able to go to a GitHub repository, click history, and be able to download it :)
21:46:49 <dwfreed> nothing prevents somebody from posting the old md5s on the internet, outside of bananas
21:46:50 <glx> the main idea is not make it super easy to download any file from bananas
21:47:02 <dwfreed> md5sum is a fact, and thus not copyrightable
21:47:08 <TrueBrain> dwfreed: nope; but I would consider it unlikely that being the case :)
21:47:29 <andythenorth> I hate this tram toolbar sprite
21:47:39 <TrueBrain> so now you know why are are looking for that solution :)
21:48:26 <nielsm> dwfreed: true it's just a fact, but the terms of service offered are that only items the author wants discoverable are discoverable via the service itself
21:48:58 <glx> basically we don't want a trivial way to construct the download url
21:49:02 <TrueBrain> for some reason every time we are looking for a solution for this, people keep argueing if we should be doing this :D I do understand it, but it is not really helping :P :D
21:49:55 <TrueBrain> frosch123: hmm .. one more nasty side-case: a NewGRF that depends on a NewGRF that has a 32bpp/8bpp offering (so 2 downloads)
21:50:07 <TrueBrain> does the UI select them both?
21:50:19 <TrueBrain> no, the server resolves it ..
21:50:22 <TrueBrain> hmm .. so which to select?
21:50:34 <frosch123> TrueBrain: for comfort, it is only a old-people problem. noone who joined the community after bananas was already a thing, ever wanted to ban something
21:50:42 <andythenorth> have ogfx first 2 'build tram' toolbar icons always been broken? Or just me
21:50:46 <TrueBrain> frosch123: truth! :D
21:51:02 <andythenorth> you didn't see simyoulater?
21:51:06 <andythenorth> and all those dramas?
21:51:11 <frosch123> TrueBrain: dependencies do not care about 32bpp/8bpp, both are the same for the dependency
21:51:21 <frosch123> TrueBrain: also, imho 32bpp/8bpp is something for later
21:51:23 <andythenorth> or the inevitable deletion of all canadian content from that other forum? :P
21:51:33 <TrueBrain> frosch123: yeah, but the server tells the client which of the two is the depdency .. so not sure what to do in that case :)
21:51:36 <TrueBrain> but okay, you are right
21:51:41 <TrueBrain> this is more a mental exercise than anything else
21:51:44 <andythenorth> urgh where are these frigging sprites :D
21:52:47 <glx> client could tell its prefered flavour, but defaulting to 8bpp should work too
21:53:04 <TrueBrain> and I made a mistake MD5 is ofc 128bits .. so the search-space is 96bits .. lol .. good luck with brute-forcing that :D
21:53:43 <dwfreed> TrueBrain: let me introduce you to my great friend the GPU cluster
21:53:52 <TrueBrain> dwfreed: it is .... online
21:53:55 <TrueBrain> so ... enjoy your cluster :)
21:54:10 <TrueBrain> you have to guess the URL between 2**96 combinations :)
21:55:03 <glx> check the png on github ?
21:55:15 <andythenorth> the correct pngs are in the spritesheet
21:55:22 <andythenorth> I just don't know what released ogfx does
21:55:26 <andythenorth> I never use it, I hate it
21:55:50 <andythenorth> I guess I'll have to download ogfx :(
21:55:55 <andythenorth> where is it even?
21:56:16 <andythenorth> google found it :P
21:56:20 <andythenorth> we are well SEO-ed
21:56:34 <andythenorth> why was it never 1.0 also?
21:56:36 <dwfreed> you just google openttd things a lot
21:56:51 <TrueBrain> andythenorth: good question
21:57:05 <andythenorth> TrueBrain: I know why, it's not really a valid question :P
21:57:11 <andythenorth> because $reasons
21:57:19 <andythenorth> where does it go
21:58:03 <andythenorth> oh I can get it on bananas :o
21:58:38 <andythenorth> yeah this is a frigging faff
21:58:53 <glx> there's also a bootstrap download if no baseset is present on the computer (but not for your OS I think)
21:58:54 <dwfreed> "Sorry, the most basic graphics set for OpenTTD isn't available on the CDN for OpenTTD graphics sets"
21:59:15 <TrueBrain> where do you read that?
21:59:18 *** Progman has joined #openttd
21:59:27 <dwfreed> TrueBrain: I made that up
21:59:44 <TrueBrain> I don't get it, sorry :(
21:59:52 <andythenorth> I never use it, it's 50% crap
21:59:56 <andythenorth> it's 50% brilliant
22:00:03 <nielsm> I think the condition for being able to download ogfx off bananas in bootstrap is that ottd is built with freetype and it can find a font to use
22:00:08 <TrueBrain> so get to work you lazy bum :P
22:00:33 <TrueBrain> nielsm: bootstrap for all OSes need a bit of love :D
22:00:40 <frosch123> nielsm: there was a deadlock when ottd cannot find a suitable font, i don't think it was fixed
22:00:53 <nielsm> why can't we ship a sprite font? :/
22:01:08 <TrueBrain> like openttd.grf :)
22:01:29 <frosch123> or bundle liberation sans ?
22:01:31 <TrueBrain> or just download a baseset, even if you cannot find a font .. most likely it will work out just fine
22:01:40 <andythenorth> now I need a frigging tram set, but not an NRT set
22:01:48 <andythenorth> I need a whining channel :P
22:02:12 <frosch123> TrueBrain: ottd is from the "phoning home is bad" era. it never openes a network connection without user action
22:02:25 <andythenorth> ok apparently 2 of the build tram sprites have been broken forever
22:02:29 <glx> TrueBrain: but you are supposed to ask the user for valifation first, needing a working gui
22:02:45 <TrueBrain> I guess ... not unreasonable tbh
22:02:48 <TrueBrain> especially in this age and day
22:02:54 <andythenorth> using OpenTTD 1.8.0 and ofgx 0.5.5
22:03:09 <frosch123> andythenorth: your eyes are too trained. i never noticed that those sprites may be wrong
22:03:22 <andythenorth> you have to trigger them into 2x zoom somehow
22:03:25 <andythenorth> to really notice
22:03:43 <andythenorth> I have NFI how to test that but I will click stuff
22:03:52 <nielsm> use an OS-supplied GUI?
22:04:04 <nielsm> a stupid messagebox for windows, equivalent for mac
22:04:12 <nielsm> not sure if SDL offers anything like that
22:04:34 <glx> for windows bootstrapping already works :)
22:04:50 <andythenorth> ok how do I find baseset sprite numbers? The in-game sprite tool is no use for that
22:05:04 <frosch123> andythenorth: it'S actually the 2x zoom auto-tram sprite that is inconsistent
22:05:16 *** supermop_work has joined #openttd
22:05:18 <frosch123> the 1x zoom sprite show the white middle line in all cases
22:05:30 <andythenorth> but I've used the auto-tram sprite in the new convert icon
22:06:09 <frosch123> well, i think many 2x sprites are vastly different to the 1x sprites. they originate from a different set
22:06:36 <andythenorth> the style is wildly different
22:06:39 <andythenorth> let's not worry about that :P
22:06:44 <frosch123> airport icon is compeltely different
22:07:09 <andythenorth> also gui2x contains the needed sprites, they just got missed
22:07:17 <andythenorth> or there is a never-merged commit or something
22:07:20 <glx> I'd say it was silly to merge 2 different styles for 1x and 2x
22:07:39 <frosch123> glx: you don't see them next to each other
22:07:41 <andythenorth> I think it's fine
22:07:43 <frosch123> so it doesn't really matter
22:07:48 <andythenorth> the fidelity on the 2x is much higher
22:07:57 <andythenorth> it was part of the whole high resolution UI thing
22:08:07 <andythenorth> remember when it was all going to be 32bpp EZ ?
22:08:24 <andythenorth> anyway, I need sprite numbers :P
22:08:28 <glx> but yes you're not supposed to see them side by side, unless some 2x are missing and default to zoomed 1x
22:09:46 <frosch123> hmm, 2200 already again :) i guess i cant be on irc if i want to get something done
22:09:51 <glx> I can search the numbers in the source
22:10:02 <andythenorth> oof frosch123 :)
22:10:33 <TrueBrain> frosch123: but we got a lot done! :D
22:12:14 <TrueBrain> okay, exporting of BaNaNaS data in the defined format works fine :)
22:12:58 <LordAro> andythenorth: you can use zbase if you like :p
22:13:58 <frosch123> LordAro: andy is on reddit, so yes he likes zbase a lot
22:14:13 <andythenorth> did anything happen on reddit today?
22:14:40 <frosch123> someone posted a video builind a network on a completely flat and empty map
22:14:40 <TrueBrain> okay, so tomorrow I can wrap this up, add the HTTP part to ottd_content, and that should be done. Some code cleanup etc ... and after that .... HTML + Javascript .. ugh ... anyone else wants to do that part?
22:15:00 <andythenorth> it's too much like a day job for me :P
22:15:06 <nielsm> and... this is not the fabled bananas2, right?
22:15:11 <andythenorth> and I have zero tolerance for all the ideological debates
22:15:13 <TrueBrain> your dayjob is yelling at people andythenorth :P
22:15:19 <TrueBrain> nielsm: it surely is not
22:15:33 <andythenorth> TrueBrain: mostly my day job is resisting yelling at people
22:15:39 <TrueBrain> nielsm: this is a .... "different" v1, at best
22:16:07 <nielsm> just a banana-flavoured item shaped like a banana
22:16:09 <TrueBrain> well, pretty sure the user experience will improve :P But more importantly, it can run on a modern infrastructure \o/ :D
22:16:26 <andythenorth> glx: any joy? The 1x sprites are drawn in sprites/png/gui/gui04.png but I can't find them used in ogfx src
22:16:30 <TrueBrain> and GitHub as authenticationt .. hmm ..
22:16:37 <andythenorth> have they ever worked?
22:16:52 <andythenorth> I wonder if they're falling back to openttd.grf
22:17:25 <TrueBrain> what will be cool, it is easy to make a notification now when someone uploaded a new version of your favorite GRF :P
22:17:35 <TrueBrain> THE POTENTIAL IS ENDLESS
22:17:44 <andythenorth> so much POTENTIAL
22:18:09 <andythenorth> everything I touch on ogfx seems to make 2 new things to do :)
22:18:13 <frosch123> TrueBrain: i want to try python3-requests-oauthlib this weekend
22:18:15 <andythenorth> does this continue, or get better? :P
22:18:21 <TrueBrain> andythenorth: I know the feeling :D I had the same with ogfx :P
22:18:32 <TrueBrain> frosch123: that exists?! :o
22:18:48 <frosch123> yes, it is kind of new in debian stable
22:18:54 <andythenorth> zero people are making PRs for opengfx?
22:19:06 <TrueBrain> andythenorth: I did enough PR-work in ogfx ffs :P
22:19:13 <frosch123> two years ago all oauth2 libs were crap
22:19:18 <andythenorth> I mean drive-by contributions
22:19:45 <andythenorth> FIRS probably has more drive-by contributors than ogfx, yet it's the baseset for most players
22:19:57 <andythenorth> just weird is all
22:20:04 <frosch123> andythenorth: i see zbase more often than ogfx
22:20:15 <frosch123> and zbase has tons of missing sprites for years
22:20:28 <andythenorth> also, players who think ogx looks nice don't really have visual acuity
22:20:44 <frosch123> TrueBrain: anyway, copy my code, or let my copy your code
22:21:07 <TrueBrain> I will copy yours :P
22:21:13 <TrueBrain> I have plenty of other code to write :D
22:21:22 <TrueBrain> but I agree, lets not both do this ;)
22:21:27 <frosch123> andythenorth: no, players who think original look nice, are just used to its noisyness
22:22:40 <andythenorth> so tempting to rewrite the whole damn thing
22:22:54 <TrueBrain> improve, don't rewrite!
22:23:24 <frosch123> andythenorth: don't work on stuff you don't use, it's not rewarding
22:23:32 <TrueBrain> frosch123: in general, I am really curious about your new eints code. Let me know when it works etc, would love to copy that :D
22:23:54 <andythenorth> frosch123: too true :)
22:24:38 <frosch123> well, i already now that new eints is only a step to newnew eints. i am sure you will request a version without need for persistent local file storage
22:24:53 <glx> andythenorth: 1339 and 1340
22:25:09 <TrueBrain> frosch123: persistent? Most likely :D What does it store persistently?
22:26:09 <andythenorth> fuck knows how the sprite numbering is supposed to avoid reading errors
22:26:11 <frosch123> TrueBrain: albert had a different vision for eints, and designed it without vcs in mind. so eints stores translation history in its own files, and vcs is more like a mirror of that
22:26:13 <andythenorth> or what those sprite numbers mean
22:26:27 <frosch123> so if you delete eints local files, all history is lost
22:26:53 <glx> numbers in ogfx png really means nothing I think
22:27:22 <frosch123> yes, they mean nothing
22:27:38 <andythenorth> so I am pointless diligently extending them as I add sprites
22:27:38 <TrueBrain> frosch123: ah; but the history doesn't really mean anything, does it?
22:27:48 <andythenorth> but they're just bullshit from grfcodec decompiles?
22:27:48 <TrueBrain> or is that used by users?
22:28:13 <glx> and they should have split the png by action 5, or whatever
22:28:16 <TrueBrain> anyway, a bit of persistent storage is not the worst; just less ideal. But I can attach an EFS to a container and have persistent storage :)
22:28:18 <andythenorth> I wrote a thing for Horse that finds sprites and re-lays them out
22:28:27 <TrueBrain> it just means there wont be any replicas running :)
22:28:29 <andythenorth> without all this weird crap
22:28:44 * andythenorth steps away from change
22:29:12 <andythenorth> super, the sprite number is a constant TRAMWAY
22:29:19 <frosch123> andythenorth: some spritesheets also contain wip sprites, or components that were used to compose sprites
22:29:20 <andythenorth> but that's not defined :)
22:30:11 <andythenorth> I need to convert that to a decimal for alternative_sprites I think
22:31:14 <andythenorth> so the 2x sprite format is
22:31:15 <andythenorth> alternative_sprites (ottd_gui179, ZOOM_LEVEL_IN_2X, BIT_DEPTH_8BPP) { tmpl_gui2x_file_toolbar(626, 776, "sprites/png/gui/gui2x.png") }
22:31:31 <andythenorth> but I don't know how to reference 'replacenew(TRAMWAY, "sprites/png/infrastructure/infra08.png")'
22:31:48 <andythenorth> and I've only got two sprites to replace
22:32:06 <andythenorth> but TRAMWAY has 113 or so
22:32:14 <glx> I think you can split TRAMWAY
22:32:50 <frosch123> andythenorth: you insert an identifier before the (
22:34:10 <frosch123> note how that file splits the replacenew into two ranges
22:34:18 <frosch123> those with 2x sprites, and those without
22:34:48 <andythenorth> the 38 is an offset?
22:34:56 <andythenorth> I did replacenew ottd_tramway2(TRAMWAY + 2, "sprites/png/infrastructure/infra08.png") {
22:35:00 <andythenorth> which is wrong then :)
22:35:06 <glx> you only need to replace the first 2 sprites
22:35:14 <frosch123> yes, ottd_gui1 contains sprites 0-37, then the next block starts at 38
22:35:27 <frosch123> actually, there are more than 2 blocks, there are dozens in that file
22:35:29 <glx> replacenew [<block-name>](<type>, [<image-file>[, <offset>]]) {
22:35:49 <andythenorth> controversial :P
22:36:23 <glx> or just look how it's done in extra-openttd-gui
22:36:45 *** supermop_work has joined #openttd
22:37:42 <andythenorth> hurrah, fixed thanks :)
22:38:23 <glx> you name the existing block, keeping the to first sprites in it, then add a new block (no name needed) with an offset of 2
22:38:32 <andythenorth> I wonder what I broke :P
22:38:45 <andythenorth> do I have to test all other tram sprites? :P
22:42:35 <glx> if cursors are concerned with 2x you should also replace sprites 2 and 3 :)
22:44:01 <andythenorth> they seem to be 1x afaict
22:44:10 <andythenorth> oof ogfx tram depot is not same size as default :P
22:45:47 * andythenorth hopes no-one ever tries to maintain FIRS
22:45:50 <andythenorth> there would be .... comments
22:46:45 <glx> ahah I think I looked at FIRS files and ran away
22:47:59 <andythenorth> it's a one-person project :P
22:48:16 <frosch123> andythenorth: modern depot sprites allow building multiple depots in a row/column, to make the depot look bigger
22:48:33 <frosch123> all newgrf have done that for 10+ years. original graphics are the only one who dont
22:49:17 <andythenorth> ogfx uses smaller bounding boxes though
22:52:35 <andythenorth> now how to test :D
22:54:48 <glx> hmm the building doesn't seem to be on its concrete spot
22:55:17 <andythenorth> "feature not bug"
22:57:41 <DorpsGek_III> [OpenTTD/OpenGFX] andythenorth commented on pull request #34: Change: add missing nrt gui sprites https://git.io/JvuNL
23:02:22 <andythenorth> oof, no parking though :P
23:08:09 *** supermop_work has joined #openttd
23:10:51 *** ChanServ sets mode: +v tokai
23:12:57 <TrueBrain> yes, go, make pictures, share!
23:14:23 <andythenorth> probably visible from my office TBH :P
23:19:33 <andythenorth> now I have to go :P
23:19:37 <andythenorth> to see if she is real
23:19:49 <TrueBrain> at least that remark of S explains so much
23:20:05 <TrueBrain> make sure to report back to us andythenorth :P
23:20:07 <andythenorth> I suspect, on balance, the internet is a negative force
23:20:19 <TrueBrain> remember that if you say she is real, you are part of the consperency
23:20:22 <andythenorth> my taxi driver told me this morning that Covid was manufactured
23:20:28 <andythenorth> to keep China's population down
23:20:38 <andythenorth> it always fails the simplicity test :P
23:20:44 <TrueBrain> I read today that it was manufactured to get Trump out of the office
23:21:41 <TrueBrain> (yes, someone really was thinking that ....)
23:21:41 <TrueBrain> yeah ... there is a chance we will be wipedout because stupidity won
23:23:24 <andythenorth> so is everything ready for 1.10 LordAro? o_O
23:23:33 <andythenorth> does anyone except pm know how to release nml?
23:23:49 <LordAro> there's a couple of pathfinder things i'd like to see fixed
23:24:19 <andythenorth> when's code freeze? :)
23:24:38 <LordAro> feature freeze was RC1
23:24:44 <TrueBrain> just before release andythenorth :P
23:25:04 <LordAro> could be RC2, i guess
23:25:48 <LordAro> would be nice if someone made a title game
23:26:01 <LordAro> since apparently no one cares enough to do a competition this year
23:26:21 <glx> already 14 PRs to backport
23:27:16 <DorpsGek_III> [OpenTTD/OpenTTD-git-hooks] LordAro merged pull request #5: Improve diagnostics from commit checker https://git.io/Jv0b0
23:27:49 *** arikover has joined #openttd
23:29:23 <LordAro> someone should review #8006 :p
23:30:17 <TrueBrain> LordAro: -1, bad commit message
23:30:22 <TrueBrain> "only store 1024 news message"
23:30:26 <TrueBrain> like ... yeah ... that is not helpful at all!
23:31:05 <TrueBrain> what does it solve? Why 1024? Why limit it? Etc etc :D
23:31:35 <glx> ignore the title, the most important thing is the other commit in it
23:31:51 <TrueBrain> I ignored the title, and was bitching about the commit message :P
23:32:01 <LordAro> They could be split up fairly easily, tbh
23:32:02 <TrueBrain> "Limit the amount of news messages to avoid overwhelming the user"? :D
23:32:12 <LordAro> TrueBrain: that's exactly what i was just writing
23:32:23 <TrueBrain> I HACKED YOUR COMPUTER :P
23:32:57 <DorpsGek_III> [OpenTTD/OpenGFX] andythenorth commented on pull request #34: Change: add missing nrt gui sprites https://git.io/JvuAg
23:34:36 <TrueBrain> YOLO review best review :D Couldn't have done it better tbh :)
23:36:08 <andythenorth> did you build it LordAro :P
23:36:22 <LordAro> and i trust that you tested it
23:36:24 <TrueBrain> you got to love CI :)
23:36:32 <andythenorth> I tested what I changed
23:36:38 <TrueBrain> we do need to add a 'git status' to see if the PNGs are unchanged after GIMP btw
23:36:41 <TrueBrain> but not sure how yet
23:36:42 <andythenorth> I didn't test I didn't break ALL THE TRAM SRPITES
23:37:18 <glx> ah yes a way to ensure new png are commited
23:39:12 <TrueBrain> <3 commitmessage :)
23:39:35 <LordAro> if preferred, the change commit should be backportable on its own, if you don't want the scrollbar changes
23:39:44 <LordAro> (which weren't hugely tested, i'll be honest)
23:40:05 <TrueBrain> if you already aren't really sure ..
23:40:40 <LordAro> because i have no issues with breaking master in some edgecases :p
23:40:45 <glx> I think the scrollbar stuff should be in a separate PRs
23:41:45 <TrueBrain> breaking master is fine, but making backports more difficult less so :P And we know .. lazy ... I would be lazy too :D
23:41:46 <glx> it probably works fine, but who knows
23:42:21 <LordAro> i wouldn't say it'd be more difficult...
23:42:52 <TrueBrain> it assumes brains on the side of the backporter! THE HORROR!
23:42:52 <LordAro> perhaps ever so slightly easier to miss a big comment saying "Only backport this commit, not the other"
23:42:55 <glx> if someone else backports the PR both commit may be included
23:43:13 <LordAro> the commits would have to be individually cherry-picked regardless
23:43:33 <LordAro> and besides, it would be caught in the review :p
23:43:50 <glx> yeah we never fail a review
23:47:06 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro opened pull request #8026: Change: Keep News Window usable by only storing the 1024 latest news messages https://git.io/JvuxJ
23:49:54 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 approved pull request #8026: Change: Keep News Window usable by only storing the 1024 latest news messages https://git.io/Jvuxk
23:50:11 <glx> this one is easy to review, and it's trivial :)
23:50:48 <glx> just a moved function and an extra check
continue to next day ⏵
