IRC logs for #openttd on OFTC at 2021-12-19
⏴ go to previous day
00:37:39 *** gelignite has quit IRC (Quit: Stay safe!)
01:20:22 <dP> svelte has some issues but at least it's somewhat reasonable
02:01:07 *** roadt__ has joined #openttd
02:07:58 *** roadt_ has quit IRC (Ping timeout: 480 seconds)
03:08:45 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:05:03 *** Wormnest has quit IRC (Quit: Leaving)
04:06:55 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
04:40:21 *** snail_UES_ has quit IRC (Quit: snail_UES_)
04:50:23 *** Montana_ has quit IRC (Quit: Leaving)
07:48:26 *** tokai|noir has joined #openttd
07:48:26 *** ChanServ sets mode: +v tokai|noir
07:52:40 *** andythenorth has joined #openttd
07:55:21 *** tokai has quit IRC (Ping timeout: 480 seconds)
08:34:01 *** andythenorth has quit IRC (Quit: andythenorth)
08:49:07 *** andythenorth has joined #openttd
09:27:48 *** roadt__ has quit IRC (Read error: Connection timed out)
09:28:23 *** roadt__ has joined #openttd
10:09:02 *** roadt__ has quit IRC (Ping timeout: 480 seconds)
10:10:36 *** roadt__ has joined #openttd
10:29:09 *** sla_ro|master has joined #openttd
10:37:07 *** sla_ro|master has quit IRC ()
10:52:21 <TrueBrain> I see it is time to hit that Unsubscribe button :P
11:00:37 <michi_cc> Hmm, could the command packet use TCP_MTU_MTU instead of COMPAT_MTU for maximum size, or is there some other limitation?
11:01:33 <TrueBrain> as long as both sides agree on the MTU, there is no issue
11:01:42 <nielsm> does it even matter? set TCP_NODELAY on the socket and just assume the internet handles things right
11:01:54 <TrueBrain> for the admin protocol this was a bit iffy, as that basically means you need to tell all current clients to expect larger packets :)
11:02:12 <TrueBrain> nielsm: the name is picked poorly, it is "max size per packet"
11:02:20 <TrueBrain> it has nothing to do anymore with actual MTU
11:02:52 <nielsm> any reason it can't be MAXINT ?
11:03:11 <TrueBrain> well, no, that is a lie
11:03:16 <TrueBrain> Rb wrote a nice story about the current limit
11:03:18 <nielsm> we support machines with 4 MB RAM?
11:03:19 <TrueBrain> and how to increase is further
11:03:34 <TrueBrain> don't be childish nielsm ;)
11:03:36 <TrueBrain> trying to help here :)
11:04:38 <nielsm> okay I can see a reason to have a kind of limit, since a malicious client could potentially cause DoS stuff by sending big packets, or indicating it intends to send a large packet and not following up
11:05:24 <TrueBrain> basically, we encode the length in each packet so it can be validated, and that currently is limited in size. Rb suggests an auto-grow mechanism, but we have no reason yet to make that change.
11:07:20 <TrueBrain> anyway, admin clients might assume the old ~1500 per packet limit, and as such, they are still on COMPAT_MTU. If anyone is brave enough to change that, technically nothing is holding you back :)
11:08:28 <nielsm> I still haven't seen a document that properly explained what the admin port exactly is capable of or does, I just know that it's possible to send/receive some very specific commands to it to interact with a GS
11:08:42 <nielsm> but does it relay any other commands? does it accept any other commands? arbitrary commands?
11:10:32 <TrueBrain> I know nothing about the admin protocol itself, so I cannot help you there :(
11:10:35 <Eddi|zuHause> i think it was originally designed to replace hacky server control mechanisms that hooked into the dedicated server console, which had ... limits
11:11:40 <michi_cc> Okay, even with COMPAT_MTU, there's still up to 1445 bytes of command data possible.
11:12:23 <nielsm> it talks about various packet names, but never about what the actual format of the packets is
11:12:40 <nielsm> I think that _might_ be documented on the wiki, somewhere hidden away, and otherwise only in the source code
11:14:22 <nielsm> it also doesn't document anything about any maximum message length
11:17:18 <michi_cc> Hmm, this document might be read that we should have incremented NETWORK_GAME_ADMIN_VERSION to indicate the changed command packet format.
11:20:46 <TrueBrain> michi_cc: that is the easiest way to tell others something has changed for sure
11:21:00 <TrueBrain> but, from 12.0 experience .. people tend to just ignore that field and still complain :P
11:45:35 *** Eddi|zuHause has quit IRC (Remote host closed the connection)
11:47:59 *** Eddi|zuHause has joined #openttd
11:49:33 <dP> TrueBrain, wdym ignore? admin_version was never incremented
11:57:31 <andythenorth> did we increase cargo limit to 128 yet? :P
12:04:55 <andythenorth> I only have 45 coded so far in this FIRS version
12:05:03 <andythenorth> but how many kinds of cheese are there?
12:08:52 <andythenorth> do you have any cheddar?
12:09:09 <TrueBrain> you really want to hurt your players, don't you? :)
12:09:25 <andythenorth> why are they mine now? :P
12:10:21 <andythenorth> this christmas, I'll give them to someone special
12:28:52 <Eddi|zuHause> why 128? go big or go home :p
12:29:29 <Eddi|zuHause> all that this is doing is you coming next year asking for 256 :p
12:31:58 *** Eddi|zuHause has quit IRC (Read error: Connection reset by peer)
12:33:16 *** Eddi|zuHause has joined #openttd
12:48:07 <NGC3982> i love cheese more than i love my wife
12:54:13 <TrueBrain> just imagine having to run a network of 128 cargoes ....
12:54:21 <TrueBrain> I don't ..... that ..... yeah :P
12:54:28 <TrueBrain> but I had similar with 64 cargoes, so yeah
13:10:33 *** frosch123 has joined #openttd
13:10:56 <andythenorth> ok so Eddi|zuHause is recommending 256
13:11:36 <Eddi|zuHause> no. i was more aiming for 64k or something :p
13:11:45 <dP> ought to be enough for anybody
13:13:24 <andythenorth> where do we stand on the 'no war' theme today?
13:13:54 * andythenorth added explosives factory, but named 'civil explosives facility'
13:13:58 <andythenorth> IRL it's munitions
13:15:54 <frosch123> compensate it with a SJW academy
13:21:05 <nielsm> demolition technology laboratory
13:56:22 <glx> <+michi_cc> Hmm, this document might be read that we should have incremented NETWORK_GAME_ADMIN_VERSION to indicate the changed command packet format. <-- not really, at the end of the document, for ADMIN_SERVER_CMD_*, it's said "Data provided with these packets is not stable and will not be
13:56:36 <glx> admin protocol itself is unchanged AFAICT
13:57:24 <michi_cc> While the contents of the command packet was not stable, the format and size was until now. As such I think bumping it is the right thing to do.
13:58:54 <glx> "Future improvements might also add additional data to packets. This data should be ignored. Data will never be removed from packets in later versions, except the possibility that complete packets are dropped in favour of a new packet."
13:59:13 <glx> maybe we should also change the packet number
14:08:25 <frosch123> are there other packets affected as well? bumping the version would allow changing other packets as well, like the command-list one
14:09:11 <glx> ADMIN_PACKET_SERVER_CMD_LOGGING is the only affected packet (complete change of format)
14:09:44 <glx> but indeed adding more data to ADMIN_PACKET_SERVER_CMD_NAMES may be a good idea
14:10:13 <frosch123> i think adding a different packet number would just add to confusion, since the old one would never be used again
14:11:54 <glx> at least with a new number, old clients would still "work", even if they don't properly check protocol version
14:13:19 <michi_cc> ADMIN_PACKET_SERVER_CMD_LOGGING has to be explicitly requested, I'd say it is unlikely somebody would request the packets without looking at them at all.
14:15:28 <glx> in the doc we explicitly say we won't remove data, unless we drop the packet
14:27:08 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #9754: Change: [Admin] Bump admin port protocol due to command changes. https://git.io/JDS08
14:31:40 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #9755: Codechange: [Script] Prettify squirrel call helpers by using C++17. https://git.io/JDSEV
14:46:01 <frosch123> yay, 47 added, 571 removed :)
14:46:55 <LordAro> how much added to compilation time? :p
14:58:38 *** andythenorth has quit IRC (Quit: andythenorth)
15:00:16 <TrueBrain> And .. another log4j cve
15:00:22 <TrueBrain> The gift that keeps on giving
15:02:17 <frosch123> it's boring this time
15:03:40 <frosch123> from the discussion on HN 2 days ago, it is barely worth a CVE. the new score of 7 is kind of surprising, so maybe there was more to it
15:04:01 <peter1138> Is that the 2nd one, or is there a 3rd now?
15:05:02 <frosch123> 1st and 2nd were RCE, 3rd is infinite loop when (intentionally?) misconfiguring
15:05:12 <frosch123> imo 3rd does not really count as remote exploit
15:07:57 <frosch123> but hey, when this is done, there is a really well reviewed java logging framework :)
15:10:04 <frosch123> i wonder what would have happend if there was no active maintainer, log4j 1.x saw about no commits from 2004 until appearance of log4j 2.x in 2014
15:10:43 <frosch123> so there may as well have been noone fixing the CVEs
15:20:09 <glx> score is lower for each new CVE
15:27:00 *** roadt__ has quit IRC (Ping timeout: 480 seconds)
15:29:26 *** roadt__ has joined #openttd
15:31:59 *** jottyfan has joined #openttd
15:48:13 *** Kitrana1 has joined #openttd
15:51:49 <DorpsGek> [OpenTTD/OpenTTD] michicc merged pull request #9753: Cleanup #9725: Clean up Doxygen docs and fix typo. https://git.io/JD1lp
15:52:59 <DorpsGek> [OpenTTD/OpenTTD] michicc approved pull request #9738: Fix #9735: Fix OrderBackup::Reset in non-GUI case https://git.io/JDS7S
15:53:07 *** Kitrana has quit IRC (Ping timeout: 480 seconds)
15:53:14 <DorpsGek> [OpenTTD/OpenTTD] michicc merged pull request #9738: Fix #9735: Fix OrderBackup::Reset in non-GUI case https://git.io/JMxjE
15:53:17 <DorpsGek> [OpenTTD/OpenTTD] michicc closed issue #9735: [Bug]: Multiplayer desync due to incorrect OrderBackup handling https://git.io/JMAM2
16:02:25 *** esselfe has joined #openttd
16:18:44 <nielsm> hmm I'm looking at the route restrictions (Trace Restrict) feature in JGR's patch pack hoping to upstream it, it looks like there's been enough changes since the original patches that it may be easiest to re-implement following the original sequence of patches, rather than trying to apply them
16:30:37 *** sla_ro|master has joined #openttd
16:32:25 *** andythenorth has joined #openttd
16:49:30 *** Wormnest has joined #openttd
17:05:33 *** Flygon has quit IRC (Read error: Connection reset by peer)
17:07:25 *** gelignite has joined #openttd
17:29:51 *** andythenorth has quit IRC (Quit: andythenorth)
17:30:52 *** andythenorth has joined #openttd
17:42:38 *** esselfe has quit IRC (Read error: No route to host)
17:45:07 <glx> seems ANIM_TRIGGER_OBJ were added for a doc update, while OBJ_ANIM were already present but never documented
17:45:48 <frosch123> probably someone forgot to add a comment to OBJ_ANIM stuff
17:46:02 <frosch123> various things were renamed and the old one kept as "deprecated"
17:47:09 *** esselfe has joined #openttd
17:47:20 <frosch123> hmm, looks like that "deprecated" stuff was only added with sweets/candy
17:50:00 <glx> I think OBJ_ANIM was overlooked when the other were added
17:50:40 <glx> not surprising with the size of global_constants array
17:51:32 <frosch123> ah, i forgot there were html docs before wiki
18:01:28 *** jottyfan has quit IRC (Quit: jottyfan)
19:00:10 <DorpsGek> [OpenTTD/OpenTTD] danskidb commented on issue #9743: [Crash]: OpenTTD crashes on startup (MacBook Pro 2020, M1) https://git.io/JDCuI
19:02:43 *** jottyfan has joined #openttd
19:11:43 *** andythenorth has quit IRC (Quit: andythenorth)
19:27:56 *** sla_ro|master has quit IRC ()
19:39:07 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on issue #9743: [Crash]: OpenTTD crashes on startup (MacBook Pro 2020, M1) https://git.io/JDCuI
19:43:57 <frosch123> is there some way to remember the meaning of ZOOM_LVL_NORMAL? i never know what it means, what is normal?
19:54:26 <Rubidium> I guess the old normal
19:55:30 <frosch123> no, it's not that, that's why it's so confusing
19:57:17 <TrueBrain> so first explain it to us what it means
19:57:23 <TrueBrain> and we can tell you if that was normal :P
19:58:22 <frosch123> the old default zoom is ZOOM_LVL_VIEWPORT = 4X
19:58:43 <TrueBrain> isn't it still the default? :P
19:58:49 <TrueBrain> so NORMAL is what we would call 4x zoom-in?
19:59:09 <frosch123> newgrf/nml define sprite zoom levels relative to that, so you have zoom-in 4x, zoom-in 2x, default, zoom-out 2x, ...
19:59:21 <frosch123> but ottd only counts zoom-out, which shifts all levels by 2
20:01:09 <TrueBrain> sadly, as it turns out, I was that weirdo
20:01:17 <TrueBrain> fucking hate when that happens
20:01:53 <TrueBrain> well, no, in my world NORMAL was the current default zoom
20:02:29 <frosch123> TrueBrain: are you TruePeter?
20:02:30 <TrueBrain> so yeah, that commit made it weird
20:02:44 <TrueBrain> frosch123: no, I tracked down who introdced ZOOM_LVL_NORMAL
20:02:49 <TrueBrain> as "normal" is just a weird word
20:02:53 <TrueBrain> but that is on me :P
20:03:41 <TrueBrain> and 815982 made it have another meaning .. so that is not on me \o/
20:04:09 <TrueBrain> pointing fingers, the most important thing to do!
20:05:13 <TrueBrain> ZOOM_LVL_MOST_ZOOMED_IN
20:05:15 <TrueBrain> so that is the way to remember it!
20:05:16 <frosch123> the other day i saw an article on the author of mediawiki. turns out it was a student on their first project trying to learn php
20:05:31 <TrueBrain> that explains ... A LOT
20:07:54 <frosch123> that just came second to my mind when you mentioned "pointing fingers"
20:08:05 <frosch123> i thought the first thing was too easy :p
20:08:24 <TrueBrain> I think this is the moment I say: FU
20:09:11 <frosch123> no, i meant that popular java framework
20:14:17 <TrueBrain> ported Cargoes to the new framework for TrueGRF .. now only Industries is remaining ..
20:14:19 <TrueBrain> something for tomorrow :)
20:14:51 *** andythenorth has joined #openttd
20:14:56 <andythenorth> regions then? :P
20:15:18 <frosch123> did you try firs2? it has regions, called "clusters"
20:17:28 <frosch123> we could do a FIRS week, play one evening each with FIRS1..FIRS5
20:20:50 <frosch123> i lost counting 10 years ago
20:21:21 <andythenorth> I did try assigning towns as 'types'
20:21:28 <andythenorth> can't remember why it doesn't work so well
20:21:38 <andythenorth> oh there's no appropriate callback
20:21:45 <andythenorth> first constructed industry per town has to do it :P
20:24:42 <glx> oh this zoom thing reminds me the intro game crash
20:36:02 *** andythenorth has quit IRC (Quit: andythenorth)
20:36:23 *** andythenorth has joined #openttd
20:42:35 <michi_cc> zoom_min is the gift that keeps on giving.
20:43:38 <frosch123> in this case it does not crash though, right? it disables the touchbar completely
20:44:46 <michi_cc> "OpenTTD crashes on startup"
20:45:25 <frosch123> that's the race condition with spriteloading, it's not the zoom_min thing
20:46:44 <michi_cc> Hmm, last comment of danskidb mentioned crash with zoom_min 2, but maybe that in fact the race condition.
20:48:08 <frosch123> hmm, no, you are right, it does not check for the zoom level existing
20:48:24 <frosch123> i had the spriteexists() call in mind, but that one does not check zoomlevels
20:48:42 *** andythenorth has quit IRC (Quit: andythenorth)
21:01:59 *** jottyfan has joined #openttd
21:30:31 *** jottyfan has quit IRC (Quit: jottyfan)
21:30:49 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
22:04:08 *** andythenorth has joined #openttd
22:11:29 *** andythenorth has quit IRC (Quit: andythenorth)
22:19:16 *** sla_ro|master has joined #openttd
22:21:30 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:36:17 *** gelignite has quit IRC (Quit: Stay safe!)
23:09:10 *** Etua has quit IRC (Quit: Etua)
23:12:56 *** sla_ro|master has quit IRC ()
continue to next day ⏵