Config
Log for #openttd on 26th February 2020:
Times are UTC Toggle Colours
00:45:35  *** Wormnest_ has quit IRC
02:28:20  *** Wormnest_ has joined #openttd
02:45:27  *** Wormnest_ has quit IRC
03:14:30  *** D-HUND has joined #openttd
03:17:51  *** debdog has quit IRC
03:26:57  *** glx has quit IRC
03:50:19  *** tokai|noir has joined #openttd
03:50:19  *** ChanServ sets mode: +v tokai|noir
03:56:58  *** tokai has quit IRC
04:49:58  <DorpsGek_III> [OpenTTD/OpenTTD] spauka opened pull request #8023: Fix #7664: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu4O
04:51:04  <DorpsGek_III> [OpenTTD/OpenTTD] spauka updated pull request #8023: Fix #7664: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu4O
04:53:37  <DorpsGek_III> [OpenTTD/OpenTTD] spauka updated pull request #8023: Fix #7664: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu4O
04:55:30  <DorpsGek_III> [OpenTTD/OpenTTD] spauka updated pull request #8023: Fix #7644: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu4O
04:57:22  <DorpsGek_III> [OpenTTD/OpenTTD] spauka commented on issue #7644: Mysteriously poor performance on macOS https://git.io/fjii3
06:49:10  *** Smedles_ has joined #openttd
06:49:51  *** Smedles has quit IRC
07:00:38  *** tokai has joined #openttd
07:00:38  *** ChanServ sets mode: +v tokai
07:07:17  *** tokai|noir has quit IRC
07:28:07  *** y2kboy23 has quit IRC
07:39:45  *** Smedles_ has quit IRC
07:41:04  *** Smedles has joined #openttd
08:11:26  *** cHawk has joined #openttd
08:27:03  *** tokai|noir has joined #openttd
08:27:03  *** ChanServ sets mode: +v tokai|noir
08:34:00  *** tokai has quit IRC
08:34:19  *** y2kboy23 has joined #openttd
08:45:08  *** y2kboy23 has quit IRC
08:48:05  *** y2kboy23 has joined #openttd
09:09:28  *** Smedles has quit IRC
09:15:29  *** Smedles has joined #openttd
09:37:42  *** Smedles has quit IRC
09:38:43  *** Smedles has joined #openttd
09:43:40  *** Smedles has quit IRC
09:44:55  *** Smedles has joined #openttd
10:03:08  *** tokai has joined #openttd
10:03:08  *** ChanServ sets mode: +v tokai
10:10:00  *** tokai|noir has quit IRC
10:18:27  *** gelignite has joined #openttd
10:44:34  <TrueBrain> so ... I need a savegame with a NewGRF or something in it .. hmmm
10:44:55  *** D-HUND is now known as debdog
10:46:43  <TrueBrain> ha, found one on my disk, w00p
10:49:56  <TrueBrain> okay ... next question .. how the hell do you trigger a CLIENT_CONTENT_INFO_EXTID .. hmmmm
10:51:37  *** gelignite_ has joined #openttd
10:53:01  <TrueBrain> via the console, lol, okay
10:57:56  *** gelignite has quit IRC
10:58:09  *** tokai|noir has joined #openttd
10:58:09  *** ChanServ sets mode: +v tokai|noir
10:59:59  <TrueBrain> OpenTTD keeps finding an OpenGFX, and I have no clue where it is on my disk :D Sweet!
11:01:18  <TrueBrain> ah, "Public". URL in docs is wrong, funny :)
11:04:58  *** tokai has quit IRC
11:10:35  <TrueBrain> so the code has this function: "RequestContentList(ContentVector *cv, bool send_md5sum)"
11:10:42  <TrueBrain> which is never called with md5sum=false
11:10:54  <TrueBrain> we have a whole protocol function that is never used :P funny
11:12:17  <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
11:12:20  <TrueBrain> not what I expected :D
11:19:21  <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
11:26:04  <LordAro> nice
11:26:14  <LordAro> Truelight: bad
11:36:13  *** tokai has joined #openttd
11:36:13  *** ChanServ sets mode: +v tokai
11:37:32  <TrueBrain> lol .. the download mechanism is a bit funky ... client: request A and B
11:37:35  <TrueBrain> server: here you have C and D
11:37:48  <TrueBrain> client: sure, I just told the user I got A and B, but I really stored C and D, no problem
11:42:10  <TrueBrain> something I did not know (and clearly once knew): we store all content as gzip-files .. that saves a lot of bandwidth :D
11:43:08  *** tokai|noir has quit IRC
11:44:01  <TrueBrain> anyway, basics of ottd_content are done. Userflow is implemented. Now all it needs it a database and filestorage :)
11:44:18  *** tokai has quit IRC
11:48:00  <LordAro> :)
12:36:28  *** gelignite has joined #openttd
12:37:32  <TrueBrain> Endianess is a bit tricky with content I noticed ...
12:38:10  <TrueBrain> the ID OGFX becomes 0x5846474f and not 4f474658 what you would expect when reading OGFX
12:38:32  <TrueBrain> happens when you abuse char for uint32s :)
12:38:56  *** gelignite_ has quit IRC
12:56:41  <peter1138> Just use IDs in the "reserved" range.
13:00:27  * peter1138 copy & pastes code... except when MonoDevelop crashes repeatedly.
13:03:46  *** tokai has joined #openttd
13:03:46  *** ChanServ sets mode: +v tokai
13:04:50  *** Samu has joined #openttd
13:04:52  <Samu> hi
13:22:18  <TrueBrain> hmm .. sometimes it is "Base Graphics", sometimes it is "Base Graphic"
13:22:51  <TrueBrain> same with "Base Sound(s)"
13:22:56  <TrueBrain> all others are singular
13:25:40  * LordAro creates a 477G swap partition
13:26:18  <TrueBrain> LOL! But why?!
13:26:27  <LordAro> Perfectly Legitimate Reasons
13:26:54  <TrueBrain> I hope it is stored on SSDs :P
13:26:57  <LordAro> NVMe :)
13:27:04  <TrueBrain> tomato tomato :)
13:27:12  <LordAro> true
13:27:50  <LordAro> TrueBrain: I'd say they should all be plural - "Music" is both singular and plural
13:28:11  <LordAro> peter1138: has MonoDevelop ever stopped crashing repeatedly?
13:28:13  <TrueBrain> NewGRF? AI? AI Library?
13:28:22  <LordAro> NewGRFs, AIs, AI Libraries
13:28:36  <LordAro> unless... what are we referring to here?
13:28:43  <LordAro> i should probably know that first :)
13:29:11  <TrueBrain> https://github.com/OpenTTD/OpenTTD/blob/master/src/network/core/tcp_content.h#L22
13:29:20  <TrueBrain> https://github.com/OpenTTD/musa/blob/master/example.ini#L3
13:29:32  <TrueBrain> the last one has all entries fully singular
13:29:35  <TrueBrain> the first one has a mix
13:29:59  <LordAro> i see
13:30:05  <TrueBrain> its a mess :P
13:30:15  <LordAro> i guess it comes from "Base *" being collections, rather than a single item
13:30:20  <TrueBrain> I guess so too
13:30:24  <peter1138> LordAro :D
13:30:37  <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
13:30:51  <peter1138> LordAro, and that is one reason why I'm copying _out_ of MonoDevelop, heh
13:31:15  <LordAro> "Base Graphic" and "Base Sound" is sounds weird to me though ("Music" is already both singular and plural)
13:31:31  <TrueBrain> I dont care how things sound; I hate inconsistencies :P
13:31:43  <TrueBrain> existing musa ini files will have "Base Graphic"
13:31:50  <LordAro> peter1138: i don't think i've tried to use MonoDevelop in ...8 years?
13:32:01  <LordAro> TrueBrain: bleh
13:32:24  <peter1138> LordAro, I'm a bit behind, I've got VS Code open on the other half of the screen, heh.
13:47:28  <TrueBrain> argh, uniqueids are really terrible ... endianess conflicts++
13:48:46  <TrueBrain> char 'OGFX' is stored in memory as 0x"X", 0x"F", 0x"G", 0x"O" ... so 0x5846474f
13:48:57  <TrueBrain> euh, not in memory
13:49:05  <TrueBrain> well, yes, but that is not what I meant
13:49:15  <TrueBrain> so the int value is 1481000783
13:49:20  <TrueBrain> @calc  0x5846474f
13:49:20  <DorpsGek> TrueBrain: 1481000783
13:49:21  <TrueBrain> yes
13:49:43  <TrueBrain> so the folder should be named 5846474f
13:50:25  <TrueBrain> so .. struct.unpack("!I", bytes.fromhex("0x5846474f")) .. yes
13:50:37  <TrueBrain> (and yes, ! as it needs to read it in "BE" order ofc)
13:50:46  <TrueBrain> ugh .. this was wayyyy to complicated for something this simple :P
13:51:06  <TrueBrain> so Python tells me b'XFGO'
13:51:12  <TrueBrain> which is correct :) But looks weird :P
14:03:21  *** Flygon has quit IRC
14:43:27  <Eddi|zuHause> just make a function for "grfID('XFGO')" that makes a byteswap?
14:43:37  <Eddi|zuHause> err
14:43:45  <Eddi|zuHause> grfID('OGFX')
14:44:10  <TrueBrain> I don't really care how Python represents it. I was more wondering how to get it right :P
14:54:13  <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
15:01:02  *** nielsm has joined #openttd
15:08:58  <TrueBrain> cool ... except for the minor flaw for savegame compatibility, the new content server works :D Well, that was easy .. :P
15:09:11  <TrueBrain> needs cleaning up and some hardening, ofc .. but .. this is not bad :)
15:17:13  *** cHawk has quit IRC
15:26:40  *** Yee has joined #openttd
15:28:38  *** Yee has quit IRC
15:30:04  <Samu> why's the cost of a corner tile square root of 2 / 2
15:31:42  <Samu> sqrt(2)/2 ~ 0.7071, so for NPF it's 70, for YAPF it's 71? shouldn't it be equal on both?
15:32:09  <TrueBrain> hmmm .... there are 2 NewGRFs with the same uniqueid .. made by 2 authors ... how could this ever been uploaded :o
15:32:35  <Samu> are they mine?
15:33:34  <Samu> what are their names?
15:37:38  <TrueBrain> ah .. one author changed its GRFID .. allowing another different package to upload theirs under the same GRFID
15:37:40  <TrueBrain> that is of course wrong
15:37:52  <TrueBrain> so I found a bug in BaNaNaS. Where can I report it? :D
15:38:48  <DorpsGek_III> [OpenTTD/OpenGFX] kaechele commented on pull request #26: Change: Use xcf2png instead of GIMP https://git.io/Jvu12
15:39:44  *** Wormnest_ has joined #openttd
15:41:16  <heffer> Yeah. Please just don't use xcf2png :D
15:42:37  <DorpsGek_III> [OpenTTD/OpenGFX] TrueBrain commented on pull request #26: Change: Use xcf2png instead of GIMP https://git.io/Jvu19
15:43:48  <TrueBrain> that comment triggered me :P And not because of the content of the PR :D
15:46:07  <milek7> in case gimp stays decision is needed on that 2.10 differences
15:46:13  <milek7> like whether it needs investigation or new blending is fine
15:46:38  *** sla_ro|master has joined #openttd
15:46:58  <TrueBrain> this double GRFID in BaNaNaS is a bit of a mystery to me ... I don't fully understand how it happened ..
15:47:08  <TrueBrain> maybe not worth my time atm :D
15:49:01  <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.
15:49:21  <TrueBrain> not generating the PNGs is still "from source" :)
15:49:40  <TrueBrain> but it makes me really happy you validate the results :D
15:50:53  <TrueBrain> as I noticed this can very easily go really wrong :P
15:51:00  <TrueBrain> we might just need to look at a better solution for this
15:53:34  <TrueBrain> hmm .. more BaNaNaS odities .. heightmaps and scenarios have no uniqueid .. it is the id of the row in the database .. rriiiigggghhhttttt
15:54:43  <nielsm> I'll test the osx colourspace thing in a mone
15:54:46  <nielsm> in a moment
16:07:22  <TrueBrain> openttd-supported-version:
16:07:22  <TrueBrain> minimum: "1.2.0"
16:07:22  <TrueBrain> maximum: "1.7.0"
16:07:35  <TrueBrain> you run OpenTTD 1.7.0 .. is this GRF compatible?
16:10:19  <LordAro> would suggest not
16:10:27  <LordAro> poorly named, perhaps
16:10:47  <TrueBrain> I agree
16:10:50  <TrueBrain> what would be better names?
16:11:05  <TrueBrain> especially as no longer have a "svn version", this at least becomes a bit easier
16:13:57  *** tokai|noir has joined #openttd
16:13:57  *** ChanServ sets mode: +v tokai|noir
16:19:58  *** arikover has joined #openttd
16:20:02  <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
16:20:38  *** tokai has quit IRC
16:21:30  <TrueBrain> openttd-supported-version:
16:21:31  <TrueBrain> supported-from: 1.3.0
16:21:39  <TrueBrain> openttd-supported-version:
16:21:39  <TrueBrain> not-supported-from: 1.6.0
16:21:43  <TrueBrain> guess that is most clear
16:22:16  <LordAro> yeah
16:22:24  <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on pull request #8023: Fix #7644: (Cocoa) Manually set colorspace to sRGB https://git.io/JvuDw
16:22:44  <TrueBrain> question is ... is 1.6.0 including beta and RC?
16:23:06  <LordAro> i guess it should do really...
16:23:41  <TrueBrain> so that means versions should always be major.minor.patch
16:23:46  <TrueBrain> as we sometimes introduce new stuff in patches, ofc
16:23:47  <LordAro> i mean, we basically do semver, could just say <1.6.0
16:24:00  <TrueBrain> semver is horrible with betas :)
16:24:04  <TrueBrain> 1.6.0-beta1 < 1.6.0
16:24:04  <LordAro> we *shouldn't* introduce new stuff in patches
16:24:24  <TrueBrain> so if you say >=1.6.0, it is excluding betas
16:24:47  <LordAro> well you'd do >1.5 :p
16:24:54  <TrueBrain> which includes 1.5.1
16:25:00  <LordAro> no it doesn't
16:25:03  <TrueBrain> (at least, in pip it does)
16:25:17  <LordAro> maybe i'm forgetting the correct syntax
16:25:43  <TrueBrain> seems semver doesn't define this at least
16:25:52  <LordAro> https://www.python.org/dev/peps/pep-0440/#version-specifiers
16:26:31  *** Progman has joined #openttd
16:27:10  <TrueBrain> " >1.7 will allow 1.7.1"
16:27:23  <TrueBrain> so yeah ... welcome to the confusion-land, where nobody is sane, etc
16:27:39  <TrueBrain> all I remember when I had to understand this shit, that it is horrible
16:27:44  <LordAro> ok, so you'd do >=1.6 then :p
16:27:50  <LordAro> (instead of >1.5)
16:27:51  <TrueBrain> which is excluding 1.6-beta1
16:28:04  <LordAro> yes, but 1.6-beta1 isn't a thing
16:28:07  <LordAro> 1.6.0-beta1 is
16:28:17  <TrueBrain> >=1.6 is excluding 1.6.0-beta1, fine :P
16:28:36  <TrueBrain> but okay, we don't have to do what-ever it was that Python tried to do
16:28:42  <TrueBrain> we need to do something people understand :D
16:28:58  <TrueBrain> so we can make that >= 1.6.0 means including betas
16:29:09  <TrueBrain> as that is most likely what people mean
16:29:33  <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
16:29:36  <TrueBrain> so I do: >= 1.6.1
16:29:43  <TrueBrain> I mean, including the betas and RCs ofc
16:29:47  <DorpsGek_III> [OpenTTD/OpenGFX] planetmaker commented on pull request #26: Change: Use xcf2png instead of GIMP https://git.io/JvuDH
16:30:08  <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
16:30:47  <nielsm> I wonder if we can delete/simplify the blocks that support ancient macos versions?
16:31:02  <nielsm> e.g. checks for 10.5 and 10.7
16:31:03  <TrueBrain> we only support N-2, N-1 and N, right? :)
16:31:19  <TrueBrain> I tried to remove them a few months ago .. someone told me "it works, so why remove it"
16:31:24  <TrueBrain> just as a heads-up :)
16:31:44  <FLHerne> nielsm: 10.5 was (IIRC) the last version that worked on PowerPC hardware, so someone might be upset :P
16:31:54  <TrueBrain> LordAro:
16:31:55  <TrueBrain> openttd-compatibility:
16:31:55  <TrueBrain> - ">= 1.2.0"
16:31:55  <TrueBrain> - "< 1.7.0"
16:31:57  <TrueBrain> ? :)
16:32:05  * FLHerne [finally] doesn't have any working anymore
16:32:12  <nielsm> we also removed support for dos builds and various other weird OS nobody seriously use
16:32:26  <nielsm> and I'm not sure if windows 9x builds are really possible any longer either
16:32:32  <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
16:33:13  <FLHerne> Hm, both those things are older, to the point where it's really impossible to use them for normal stuff
16:33:14  <planetmaker> hm... PRs welcome :P
16:33:34  <FLHerne> But no, I'm not certain it's worth caring
16:33:54  <nielsm> nobody cares whether we support linux 2.2 either
16:34:21  <TrueBrain> lol, some people did this for their NewGRF:
16:34:22  <TrueBrain> openttd-compatibility:
16:34:22  <TrueBrain> - "< 0.6.0"
16:34:30  <TrueBrain> the content-service only exist from 0.6 and upwards :D
16:34:30  <nielsm> having a ton of ifdef blocks makes code harder to maintain and bugfixes harder to verify
16:34:36  <TrueBrain> basically, making it not available :)
16:34:48  <planetmaker> TrueBrain, that's means to hide it
16:34:58  <planetmaker> and make it available for savegames which need it
16:35:02  <planetmaker> at the very same time
16:35:02  <TrueBrain> others used "< 0.0.0"
16:35:06  <TrueBrain> which is easier to understand :)
16:35:10  <planetmaker> well :) yeah
16:36:18  <TrueBrain> openttd-compatibility:
16:36:18  <TrueBrain> - ">= 3.11.9"
16:36:18  <TrueBrain> - "< 0.0.0"
16:36:19  <milek7> does "not-supported-from" is sensible at all? it shouldn't be all backwards compatible with old newgrfs?
16:36:19  <TrueBrain> haha :D
16:41:37  *** BOB has joined #openttd
16:46:24  <BOB> are there any plans for new Vehicles to openttd. i have the idea for a airlander like vehicle
16:47:45  <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?
16:49:46  *** Etua has joined #openttd
16:49:59  <BOB> so i need to draw a new texture and put that in the NewGRF library?
16:51:53  <nielsm> if you mean something that's neither road vehicle, train, ship, or aircraft, then no that's probably never happening
16:52:13  <nielsm> but as long as you can place the vehicle you imagine under one of those categories, go hog wild
16:52:32  <TrueBrain> hmm .. either the default font cannot render Korean chars, or something is wrong with UTF-8 :)
16:55:35  <TrueBrain> holy crap, pressing "Check Online Content" is hella slow, for the next 10+ seconds OpenTTD is not behaving nicely :)
16:55:47  <TrueBrain> maybe it shouldn't update the UI for every new entry :D
16:55:47  <LordAro> TrueBrain: list has gotten quite big
16:55:55  <LordAro> though yes, that doesn't help :p
16:56:05  <TrueBrain> or maybe it should only fetch per category
16:56:09  <TrueBrain> or only "what can be update"
16:56:11  <TrueBrain> or something :P
16:56:29  <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.
16:56:55  <BOB> i have found the NewGRF make documantion
16:57:09  <BOB> i will take a look a it.
16:57:20  <TrueBrain> LordAro: something for you to fix in 1.11? :D
16:57:30  <LordAro> TrueBrain: maybe!
16:57:43  <TrueBrain> the server needs 0.03 seconds to tansmit the whole list
16:57:45  <TrueBrain> 30msec
16:57:46  <TrueBrain> that is fair
16:58:00  <LordAro> though not updating the UI for every new entry could reasonably go in 1.10
16:58:06  <nielsm> BOB: also the term is sprite here, OTTD does not use 3D models so there are no textures
16:58:20  <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
16:59:06  <LordAro> TrueBrain: :p
16:59:12  <LordAro> TrueBrain: you should file a bug
16:59:16  <BOB> @nielsm, with texture i mean the image of the Vehicle i see on screen when playing the game
16:59:23  <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 ..
16:59:58  <nielsm> BOB yes you're going to confuse everyone else as long as you use the wrong word :)
17:00:32  <BOB> than i will use sprite for it
17:02:20  <DorpsGek_III> [OpenTTD/OpenTTD] TrueBrain opened issue #8024: "Check Online Content" lags the UI https://git.io/Jvuyj
17:02:29  <TrueBrain> LordAro: as requested ^^. Surprised nobody else had :)
17:09:02  <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #8024: "Check Online Content" lags the UI https://git.io/Jvuyj
17:09:33  <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
17:09:47  <BOB> http://dev.openttdcoop.org/projects/ttdviewer
17:17:20  *** glx has joined #openttd
17:17:20  *** ChanServ sets mode: +v glx
17:20:26  *** BOB has quit IRC
17:22:21  *** cHawk has joined #openttd
17:22:47  <Samu> glx, i tried diferent costs to curve90, it made little to no difference https://pastebin.com/VhRmUuiY
17:24:52  <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 :(
17:25:06  <glx> weird the cost is higher than 100/600
17:25:39  <Samu> that's not the cost, it's the number of search nodes
17:25:40  <glx> I guess with lower cost it tries more paths
17:25:44  <Samu> yes
17:26:11  <glx> when the cost is high, paths are rejected earlier
17:27:19  <Samu> but for savegame 1 it meant no difference
17:27:41  <Samu> considering my fix is in
17:27:44  <Samu> of coures
17:27:49  *** tokai has joined #openttd
17:27:49  *** ChanServ sets mode: +v tokai
17:30:00  *** Progman has quit IRC
17:30:17  <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
17:33:44  *** cHawk has quit IRC
17:34:38  *** tokai|noir has quit IRC
17:41:28  <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
17:42:14  <Samu> maybe i should look better
17:45:35  *** HerzogDeXtEr has joined #openttd
17:49:47  <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on pull request #8023: Fix #7644: (Cocoa) Manually set colorspace to sRGB https://git.io/Jvu9Z
17:54:35  <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh opened pull request #8025: Remove: Support for macOS before 10.9 https://git.io/Jvu9R
17:54:50  <nielsm> kill it!
17:59:14  <LordAro> cast it into the fire!
18:01:03  *** frosch123 has joined #openttd
18:01:15  <Samu> oh, right, yapf_roads works with segments, not tiles
18:01:51  <Samu> entire different beast than ships
18:04:02  <LordAro> nielsm: and yet you seem to have left support for PPC OSX in?
18:04:41  <LordAro> was PPC OSX even a thing?
18:05:43  <nielsm> I guess I should remove that part too :P
18:05:51  <nielsm> and yes it was
18:10:57  <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
18:12:39  <glx> also add the time of newgrf scanning before even get access to the main GUI ;)
18:12:56  * glx has way too many newgrfs
18:13:18  <frosch123> grf scan also differs a lot between optimised and non-optimised builds
18:13:49  <glx> last_newgrf_count = 1213
18:14:12  <frosch123> 1401 :p
18:14:37  <nielsm> can some of the local content md5 calculations be offloaded to a thread if they aren't already?
18:14:38  <glx> and some of them are even unusable in openttd
18:15:13  <frosch123> nielsm: at some point i wanted to store the md5sum together with file size/date in a ini file
18:15:28  <frosch123> but i dropped that idea when i noticed that optimised build is way faster
18:15:55  <TrueBrain> frosch123: this slowness is without any GRFs ;) So yeah .. lot of things to improve :)
18:16:26  <glx> BTW I think some repos need more people with write access
18:16:43  <LordAro> ^
18:16:51  <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
18:16:53  <frosch123> glx: you are owner, you can add people
18:16:59  <frosch123> LordAro: also, get 2fa
18:17:12  <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 :)
18:17:23  <LordAro> frosch123: :<
18:18:20  <frosch123> glx: just don't add individual/external contributors, use the teams please
18:19:35  <frosch123> TrueBrain: does the name differ between versions?
18:19:43  <frosch123> i thought you can't rename stuff?
18:20:11  <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 :)
18:20:40  <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
18:20:50  <TrueBrain> but there too .. the savegame download will be missing metadata
18:20:55  <TrueBrain> not sure that is a bad thing tbh ..
18:21:07  <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
18:21:29  <TrueBrain> but what if we have that ID also not on record?
18:21:49  <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
18:22:02  <frosch123> TrueBrain: how?
18:22:12  <TrueBrain> Eddi|zuHause: frosch123  mentioned that usecase; something for future extensions
18:22:27  <frosch123> the repo contains meta data per 4-byte id. that is only removed if the content is banned
18:22:43  <TrueBrain> frosch123: I just mentioned the usecase ;) We have people who want to remove their content :)
18:23:00  <TrueBrain> which is fine by me honestly, if they just want to remove their content .. it should be delisted etc
18:23:04  <frosch123> yes, but in that case all is removed, so you do not need to report the name
18:23:06  <TrueBrain> just .. the savegame should still be able to get it :)
18:23:15  <TrueBrain> so when you open such savegame, what name should I report?
18:23:24  <frosch123> none, because you can't download it
18:23:27  <TrueBrain> you can
18:23:41  <TrueBrain> you can ALWAYS download GRFs based on ID+MD5 if it was once uploaded (and not banned)
18:23:47  <frosch123> i think you mixed up some things
18:23:48  <Eddi|zuHause> TrueBrain: so what's the problem if fetching older versions doesn't have a name?
18:23:49  <TrueBrain> that is the whole point of the savegame content download thingy :)
18:24:08  <TrueBrain> frosch123: I think it is more that we are not talking about the same thing :D :)
18:24:36  <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
18:25:06  <Eddi|zuHause> something isn't right, it's reporting 98% traffic flow...
18:25:08  <TrueBrain> that is not the part I am talking about :)
18:25:28  <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
18:25:30  <TrueBrain> the normal flow is working just fine :) Let me make a few screenshots ..
18:26:02  *** arikover has quit IRC
18:26:25  *** arikover has joined #openttd
18:26:35  <TrueBrain> https://pasteboard.co/IWvke18.png https://pasteboard.co/IWvkro0.png
18:26:39  <TrueBrain> that is what I am talking about
18:26:58  <TrueBrain> if you load a savegame which was saved with an old GRF, it allows you to retrieve it
18:27:14  <TrueBrain> this can be an uniqueid (your 4-byte id), which is not publicly listed anymore
18:27:30  <DorpsGek_III> [OpenTTD/OpenTTD] frosch123 commented on issue #8024: "Check Online Content" lags the UI https://git.io/Jvuyj
18:27:32  <TrueBrain> https://github.com/TrueBrain/OpenTTD-BaNaNaS/tree/master/NewGRF/4341121f
18:27:37  <TrueBrain> as example of how the repo will look
18:27:56  <TrueBrain> so I have an ID+MD5, I have the tarball I can serve to the client, that is all working fine
18:28:09  <TrueBrain> but what name etc am I going to give it .. where do I get that metadata from
18:28:45  <DorpsGek_III> [OpenTTD/OpenTTD] TrueBrain commented on issue #8024: "Check Online Content" lags the UI https://git.io/Jvuyj
18:28:46  <frosch123> TrueBrain: you still  insist that listing the name publicly implies that something can be downloaded
18:28:52  <TrueBrain> huh?
18:28:59  *** arikover has quit IRC
18:29:12  *** arikover has joined #openttd
18:29:17  <nielsm> one of the promises of banana is that an author can delist an item so it's completely invisible
18:29:26  <TrueBrain> ^^
18:29:30  <frosch123> TrueBrain: the case you make "uniqueid (your 4-byte id), which is not publicly listed anymore" does not exist
18:29:31  <nielsm> unless a would-be downloader knows its id and md
18:29:44  <TrueBrain> frosch123: it does? Where do you get that information from?
18:30:04  <TrueBrain> it is already the case, that we have uniqueids that are not publicly listed, yet do exist
18:30:08  <frosch123> one of use has switched the universe
18:30:10  <TrueBrain> (and are not blacklisted)
18:30:44  <frosch123> TrueBrain: no, there are only cases where the latest version is set to "not compatible with your openttd"
18:31:22  <TrueBrain> this is not completely true; we also delisted items
18:31:42  <TrueBrain> but okay, that is not really important tbh
18:31:46  <TrueBrain> the question is: what do we want to do
18:31:57  <TrueBrain> on that GitHub I linked, I am pretty sure people will make PRs to remove their GRFID completely
18:32:02  <TrueBrain> because they got mad, or what-ever
18:32:09  <TrueBrain> that is a use-case I fully understand and I completely see happening
18:32:22  <TrueBrain> so you say: we are not going to support that usecase, if I understand you correctly?
18:32:48  <frosch123> no, banned stuff is removed completely
18:32:57  <TrueBrain> we are not talking about banned stuff
18:33:02  <frosch123> or are you now trying to do stuff with a public repo?
18:33:26  <frosch123> TrueBrain: "banned" and "deleted" are the same thing
18:33:40  <TrueBrain> they are? Okay .. there is a difference in my opinion, but sure
18:33:40  <Eddi|zuHause> i'm fairly convinced you two are still not talking about the same thing
18:33:43  <frosch123> both are 4-byte ids that can neither be uploaded nor downloaded
18:33:45  <TrueBrain> Eddi|zuHause: I am too :)
18:34:22  <TrueBrain> frosch123: the above usecase I just gave, shows that won't be the case
18:34:28  <TrueBrain> or are you against that usecase?
18:35:06  <TrueBrain> (which is fine, but clearly we are talking cross eachother, so I am just trying to sync up :D)
18:35:12  <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
18:35:18  <frosch123> either you leave, your you dont
18:35:26  <frosch123> there is no middle "somewhat mad"
18:35:35  <TrueBrain> okay, now I get where you are coming from :)
18:35:50  <TrueBrain> how ever mad an author will be, the ToS allows us to always distribute their content for savegame compatibility
18:35:55  <TrueBrain> so there is where our paths split :)
18:36:12  <TrueBrain> there is banned content, content that was never allowed to be uploaded because of license issues
18:36:16  <Eddi|zuHause> frosch123: the idea is: however mad they got, we can still distribute it to people who have a savegame
18:36:21  <TrueBrain> and there is deleted content, content that was delisted because someone got mad
18:36:25  <TrueBrain> the first case, data got removed
18:36:30  <TrueBrain> the last case, data exists, and is retrievable
18:36:35  <frosch123> TrueBrain: https://github.com/frosch123/bananas2/blob/master/docs/data.md <- see first section
18:37:01  <frosch123> Eddi|zuHause: that is wrong, that is betraying them and distributing the content beind their back
18:37:04  <frosch123> i hope we dont do that
18:37:20  <TrueBrain> our ToS is explicitly written to allow that
18:37:24  <TrueBrain> unless your content is banned
18:37:42  <TrueBrain> so that "promise", is not what BaNaNaS v1 does
18:37:44  <Eddi|zuHause> frosch123: you can't revoke licenses just becaue you got mad
18:37:47  <frosch123> TrueBrain: i very much hope that delisted content is not downloadable
18:37:52  <frosch123> otherwise i think we have a big issue
18:38:06  <TrueBrain> it is, that is what I am saying :) And no, we have no issue, it is explicitly written to be like that
18:38:30  <TrueBrain> basically, the ToS says: players come first, unless a license has been broken
18:38:34  <frosch123> i was convinced that it is explicitly written to not do that, and i still hope you are wrong
18:38:52  <TrueBrain> http://bananas.openttd.org/en/manager/tos/ <- not sure that works?
18:38:53  <Eddi|zuHause> i'm gonna side with TrueBrain on that
18:39:16  <TrueBrain> basically, in 2007 or 2008, the issue was people wanted their content gone, as they were mad
18:39:24  <TrueBrain> and that would invalidate 20+% of the people using NewGRF savegames
18:39:31  <TrueBrain> a LONG discussion followed
18:39:34  <TrueBrain> and the ToS got changed
18:39:43  <TrueBrain> to allow content, once uploaded, to always distribute, under conditions
18:39:51  <TrueBrain> basically, OpenTTD get a distribution license
18:40:07  <TrueBrain> the conditions are simple: only based on uniqueid+md5, the content can be made available
18:40:19  <TrueBrain> in other words: it only works for existing savegames
18:40:24  *** Wormnest_ has quit IRC
18:40:27  <TrueBrain> so if you get a savegame for 10 years ago, it will load in OpenTTD
18:40:28  <frosch123> the tos was never changed, it was extended for gs at some point, but there is only one version of the tos
18:40:42  <frosch123> you may remember drafts before bananas was started
18:40:45  *** tokai|noir has joined #openttd
18:40:45  *** ChanServ sets mode: +v tokai|noir
18:40:45  <glx> except for banned content that should not have been uploaded
18:41:44  <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
18:42:34  <frosch123> i think we live in different universes
18:42:50  <TrueBrain> well, what I described is what is currently implemented :)
18:43:06  <TrueBrain> I am fine changing it etc etc
18:43:13  <TrueBrain> I am just moddeling around the current system
18:44:18  <TrueBrain> https://gist.github.com/TrueBrain/3564623507cc30735c22781863e30f73
18:44:21  <TrueBrain> styling got lost :(
18:44:22  <TrueBrain> dammit
18:45:28  <TrueBrain> anyway, I get now why we didn't understand each other, we work from a different set of rules :D
18:45:54  <TrueBrain> - You grant the OpenTTD team to retain older versions of your content for the purpose of loading save games with said older version.
18:45:54  <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.
18:46:01  <TrueBrain> these two rules is were the difference come from, I guess
18:46:46  <frosch123> i checked svn, in my universe the tos have been unmodified since the start of bananas in 2009
18:46:46  *** arikover has quit IRC
18:46:56  <TrueBrain> frosch123: again, timeline is really not that relevant
18:46:57  <frosch123> if your universe has a different svn, i cant help
18:47:05  <TrueBrain> if this is the only ToS you have ever seen, even better :)
18:47:07  *** arikover has joined #openttd
18:47:14  <TrueBrain> I am not sure why that became a focus point :)
18:47:39  *** tokai has quit IRC
18:47:46  <frosch123> because we disagreed on what was decided how in the past, and i see my understanding backed up by vcs
18:47:52  <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?
18:48:05  <frosch123> so, unless you hacked the vcs, i have no idea where you are coming from
18:48:23  <frosch123> TrueBrain: yes, keep all, or remove all. no middle ground
18:48:44  <TrueBrain> okay. As that is not what is currently implemented, so the question is there: do we want to change the current implementation?
18:48:48  <frosch123> the author can always set stuff "not available for ottd < 0.3", but that does not delete stuff
18:48:53  <frosch123> it just makes the gui hide it