IRC logs for #openttd on OFTC at 2020-09-10
⏴ go to previous day
03:47:05 *** y2kboy23 has joined #openttd
04:44:40 *** WormnestAndroid has quit IRC
04:44:55 *** WormnestAndroid has joined #openttd
06:06:58 *** y2kboy23 has joined #openttd
08:03:41 *** sla_ro|master has joined #openttd
08:09:01 *** andythenorth has joined #openttd
09:47:30 *** Progman has joined #openttd
10:33:49 *** ChanServ sets mode: +v tokai
10:58:29 <DorpsGek_III> [OpenTTD/master-server] TrueBrain merged pull request #4: Several bugs found while bringing this to production https://git.io/JUn78
11:04:25 <DorpsGek_III> [OpenTTD/team] TrueBrain merged pull request #2: Add: initial workflow to automate requested access for translators https://git.io/JvGkD
11:05:08 <TrueBrain> let's see how this works .. :)
11:05:34 <TrueBrain> "This is not an issue to request access; ignoring"
11:08:08 <TrueBrain> who of the core developers dares to write a /approve in that ticket ? :D
11:08:40 <andythenorth> oh it's so hard not to troll you :(
11:09:01 * andythenorth must be a better person
11:09:19 <DorpsGek_III> [OpenTTD/master-server-web] TrueBrain merged pull request #5: Several bugs found while bringing this to production https://git.io/JUn7G
11:09:39 <TrueBrain> you are not a core developer andythenorth , so what-ever you reply on that ticket will be ignored :P
11:22:21 <TrueBrain> okay, deployments work .. time to roll this out in pre-production :)
11:23:58 <TrueBrain> oops .. found a "bug" ... staging is using the same database as production
11:24:05 <TrueBrain> that seems like a horrible idea really :P
11:26:18 <TrueBrain> orudge: wow, the new CPanel DNS Zone Manager is TERRIBLE. It paginates the records, but only shows this on the top right :P I know you can do nothing about that, but I just wanted to let you know :D It is a horrible UX :)
11:35:21 <TrueBrain> (it also doesn't allow changing records and pushing them at once, etc etc .. but I guess this is just more "newbie" friendly :P)
11:43:39 <TrueBrain> If anyone wants to test the new master-server, I would really appreciate that: change your "hosts" file such that "master.openttd.org" points to 3.120.179.246 and 2a05:d014:870:d200:93e4:fee6:abba:6b97
11:43:42 <TrueBrain> start your server after that
11:43:50 <TrueBrain> it should register itself without issue
11:44:13 <TrueBrain> same if you want to test the OpenTTD client
11:47:06 <TrueBrain> (and I could really use some testers, to make sure I didn't make any weird boo-boo :D)
11:48:20 <TrueBrain> the server-id is nothing more than the md5("ip:port") of any of the ip/port combinations the server is running on (so either IPv4 or IPv6)
11:48:28 <TrueBrain> is JSON :) Might be easier than HTML :D
11:49:31 <TrueBrain> I see that LordAro cannot read
11:49:43 <TrueBrain> it is not like it is LITERALLY THERE
11:49:51 <LordAro> but "allow" is the opposite of "deny"
11:49:55 <TrueBrain> "Please reply with either /approve or /deny" .. LordAro does /allow ...
11:50:07 <TrueBrain> YOU reviewed this PR :P
11:50:51 <TrueBrain> sometimes I amaze myself :D
11:52:26 <TrueBrain> owh, all these .dorpsgek.yml are wrong .. they don't announce new versions .. oops
11:52:57 <TrueBrain> let me be DorpsGek_III for today :P
12:00:48 <TrueBrain> and that one does announce ... lol
12:04:43 <TrueBrain> right, I am ready to flip the switch and bring this to production .. but I would love to get some feedback first if it works for others :D
12:07:23 <LordAro> have you tested with a variety of ancient versions?
12:09:06 *** Eddi|zuHause has joined #openttd
12:20:38 <TrueBrain> I am honestly more worried from a network perspective that things fail, but it is not a bad idea to run all major versions again :)
12:28:35 <TrueBrain> 0.4.8, 1.0.0, 1.4.4, 1.8.0, 1.9.3, 1.10.3 clients work
12:28:39 <TrueBrain> guess that is a sufficient large set :P
12:31:42 <TrueBrain> master, 1.9.0, 1.4.4 as server work .. 1.0.0 doesn't want to compile .. lets see ..
12:40:42 <TrueBrain> hmm .. 0.4.8 cannot advertise itself
12:41:41 <TrueBrain> it is funny how many issues compilers find these days on old code :P
12:47:05 <TrueBrain> okay, 1.0.0 works, 0.7.0 does not
12:47:14 <TrueBrain> do I care .... well, that is a good question :)
12:56:35 <TrueBrain> it is a freaking one-liner :P
12:57:36 <DorpsGek_III> [OpenTTD/master-server] TrueBrain opened pull request #6: Fix: also make sure servers before 1.0.0 can advertise https://git.io/JUCaG
12:58:02 <TrueBrain> a new level of pedantic, making sure a server more than 13 years old still works, but okay :P
12:58:21 <LordAro> are there any <1.0 servers running at the moment?
12:58:37 <TrueBrain> oldest is 1.3.0 for some reason
12:58:49 <TrueBrain> a whole cluster of them
12:58:59 <TrueBrain> pretty sure 7 years ago someone forgot about them, and they are still running
12:59:17 <TrueBrain> Current date: 5345-12-29
12:59:46 <TrueBrain> now review the PR, so I can deploy it :P
13:08:26 <LordAro> TrueBrain: couldn't possibly, i don't understand the code :p
13:08:36 <LordAro> why is that the only branch to explicitly store the session key?
13:08:48 <TrueBrain> the others do too, in the _next_session_id function
13:09:00 <TrueBrain> _get_next_session_key even
13:09:36 <TrueBrain> which is the bug .. it is the only branch not storing the session-key :)
13:09:50 <TrueBrain> basically, it pre-announces a server, and later on it is checked if the server really still exists
13:10:00 <TrueBrain> allows for more than one instance of master-server to run :)
13:10:22 <LordAro> it's also the only branch to not send a PACKET_UDP_MASTER_SESSION
13:10:30 <TrueBrain> and this is because of the protocol
13:10:33 <TrueBrain> protocol v1 did not support that
13:10:57 <TrueBrain> comments line 123-126
13:11:24 <TrueBrain> there are so many comments in that file .. just reading them should give you a rough understanding of how complex this protocol became :P
13:11:37 <Eddi|zuHause> <TrueBrain> it is a freaking one-liner :P <-- statistically, most things will be one-lineres
13:12:01 <andythenorth> most bugs are just one liners
13:12:11 <andythenorth> if we always committed 2 liners, would we have fewer bugs?
13:12:20 <Eddi|zuHause> it's like if you count the words in a shakespeare play. about half the words will appear exactly once
13:14:09 <Eddi|zuHause> (this is the same underlying mechanics as the 80-20 rule)
13:47:52 *** tokai|noir has joined #openttd
13:47:52 *** ChanServ sets mode: +v tokai|noir
13:58:07 <TrueBrain> LordAro: but I think in general it is safe to say that such PRs don't require an audit on functionality, more if the code quality is sufficient ;) Given I did test it etc :D
14:26:58 *** snail_UES_ has joined #openttd
14:39:05 <DorpsGek_III> [OpenTTD/master-server] LordAro approved pull request #6: Fix: also make sure servers before 1.0.0 can advertise https://git.io/JUCoN
14:39:36 <DorpsGek_III> [OpenTTD/master-server] TrueBrain merged pull request #6: Fix: also make sure servers before 1.0.0 can advertise https://git.io/JUCaG
14:40:22 <TrueBrain> okay, porting the security stuff is easy, but just .. stupid work .. :P
14:40:30 <TrueBrain> I need to copy/paste a shitton of stuff .. bit boring :P
14:41:05 <TrueBrain> and lookup subversion revisions to github hashes ..
14:49:32 <TrueBrain> 1.9.0, 1.4.4, 1.0.0, 0.7.0, 0.6.0 (advertises twice to the master-server, lol), all work. Think that is sufficient amount of testing :D
14:50:34 <TrueBrain> I see someone else also successfully started a server
14:50:50 <TrueBrain> so I guess this is ready to go live :) Tonight, I guess :)
14:51:44 <TrueBrain> hmm ... not sure how often it does DNS resolving .. possibly servers need a restart before it shows up on the new master-server .. worth testing ..
15:13:40 <TrueBrain> okay, DNS is resolved every time
15:14:50 <TrueBrain> so what is going to happen: I am going to change the TTL to < 15 minutes. At change, servers will be registered to the old master-server for the next 15 minutes, or when-ever their next broadcast is. After that, they appear on the new master-server. They remain on the old for another ~20 minutes, till the timeout there kicks in
15:17:12 <TrueBrain> found a bug in the old master-server .. it does not update your server if there is UTF-8 chars in your servername :P
15:39:01 <TrueBrain> interesting .. a CVE, with a patch, but without a commit .. so .. we never patched this upstream, or what is the message here :D
16:08:53 <Eddi|zuHause> when has an unpatched CVE ever hurt anyone... :p
16:13:07 * LordAro reads Eddi|zuHause's private keys
16:32:14 <DorpsGek_III> [OpenTTD/website] TrueBrain opened pull request #164: Remove: we no longer use a mirror network that depends on donations https://git.io/JUC1u
16:32:40 <DorpsGek_III> [OpenTTD/website] TrueBrain opened pull request #165: Add: port the security patches we have for CVEs https://git.io/JUC1g
16:32:48 <TrueBrain> there ... that should replace "security.openttd.org" :)
16:34:51 <DorpsGek_III> [OpenTTD/website] LordAro approved pull request #165: Add: port the security patches we have for CVEs https://git.io/JUC1V
16:35:06 <TrueBrain> that was ... scary quick LordAro
16:35:12 <TrueBrain> you are okay with the CSS too? :P
16:35:24 <DorpsGek_III> [OpenTTD/website] LordAro approved pull request #164: Remove: we no longer use a mirror network that depends on donations https://git.io/JUC1r
16:36:04 <LordAro> TrueBrain: that's what staging is for!
16:36:34 <TrueBrain> well, it looks fine; that is not my issue :P I expected something along the lines: ugh, that CSS looks wrong :P
16:37:38 <DorpsGek_III> [OpenTTD/website] TrueBrain merged pull request #164: Remove: we no longer use a mirror network that depends on donations https://git.io/JUC1u
16:37:51 <DorpsGek_III> [OpenTTD/website] TrueBrain merged pull request #165: Add: port the security patches we have for CVEs https://git.io/JUC1g
16:38:02 <TrueBrain> hmm, I might have merged those two too quickly
16:38:52 <TrueBrain> let me cancel one before I make things worse :D
16:55:25 <TrueBrain> LordAro: find the bugs :D
16:55:45 <LordAro> page loads, doesn't it?
16:55:47 <LordAro> what more do you want
16:56:30 <LordAro> i have to say, the fact that we've not "had" a CVE in 2013 is a bit odd
16:56:44 <LordAro> i guess that's about the time that RB stepped down?
17:00:05 *** Wormnest has joined #openttd
17:00:32 <DorpsGek> LordAro: Rubidium was last seen in #openttd 1 year, 17 weeks, 2 days, 20 hours, 58 minutes, and 36 seconds ago: <Rubidium> are those from the same NewGRF?
17:04:30 <TrueBrain> LordAro: I think he made most if not all of these .. and if you see the amount of work involved, I can imagine nobody wants to :P
17:04:47 <TrueBrain> and this morning you were overly picky, and now you found the other side of the balance .. not sure if I like that :P :P :D
17:06:30 <LordAro> well in between then i've had to try to convince someone that taking an offline copy (via various recursive wgets) of a Bamboo instance is really not the correct solution to the problem they're trying to solve
17:06:36 <LordAro> so i just don't care anymore
17:07:32 <TrueBrain> gave up on life, poor LordAro :(
17:07:38 <TrueBrain> it is going to be okay
17:10:18 <LordAro> other today fun was trying to work out why our zfs snapshots are taking so long to copy to another server
17:11:00 <LordAro> it's either "compression too high" or "compression not enough"
17:17:47 <TrueBrain> Or just too much data :p
17:18:34 <LordAro> that is an option also
17:24:23 *** gelignite has joined #openttd
18:05:38 <TrueBrain> right .. guess I am going to move the master-server to production .. I can't find a reason why not, except that I don't want to :P
18:08:24 <TrueBrain> and the first servers are already switched .. that went quicker than I expected :P
18:10:03 <TrueBrain> hmm, also servers that are not responding to game-info requests
18:13:48 <TrueBrain> hmm .. possible IPv6 is not getting through ..
18:13:55 <TrueBrain> and my home IPv6 stopped working today :(
18:18:33 <TrueBrain> ah, lol .. NAT gateways didn't have IPv6 assigned .. oops
18:19:29 *** Gustavo6046 has joined #openttd
18:20:07 <TrueBrain> at least it seems all (most?) IPv4 servers have reproted in :)
18:48:05 <TrueBrain> okay, and it appears we have a bug .. which I do not understand really :P
18:48:10 <TrueBrain> 1 server doesn't have a map_name ..
18:58:55 <TrueBrain> okay, switching back to the old infra .. I have to look into this in more detail ..
18:58:59 <TrueBrain> I do have a lot of test-data now :D
19:03:47 <TrueBrain> oops .. the old master-server is not accepting the session-keys ..
19:03:51 <TrueBrain> and it doesn't retransmit a new one ..
19:04:45 <TrueBrain> so ... guess we are IPv4 only for a bit :)
19:05:20 <Wolf01> Fine, like my API auth system, I discovered like one month after doing it that I couldn't login because I didn't implement a renew token feature
19:12:27 <andythenorth> is it pixels or tanks?
19:12:37 * andythenorth had a break from both due to profound lack of interest
19:13:18 <andythenorth> in the meanwhile I played a lot of Warcraft 1
19:13:22 <andythenorth> trying custom campaigns
19:13:42 <andythenorth> using save-load I managed to win every silly combination of army I could find
19:13:50 <andythenorth> although I didn't try 1 peon and 1 catapult
19:13:53 <andythenorth> I don't think that's viable
19:14:16 <andythenorth> it's about £5 on GOG.com and is one of the 1994 big 3
19:14:27 <andythenorth> along with TTD and Doom 1 shareware edition
19:14:42 <andythenorth> Railroad Tycoon in 4th Place
19:40:57 <TrueBrain> error on socket: [Errno -9] Address family for hostname not supported <- grr @ ECS .. why doesn't this container have IPv6 ..
20:01:30 *** frosch123 has joined #openttd
20:06:49 *** HerzogDeXtEr has joined #openttd
20:09:14 <TrueBrain> hmm .. it might be that I cannot get IPv6 enabled on these containers ... mehhhhh
20:17:21 <TrueBrain> okay, to fix this properly, I do have to switch back to the old infra .. how am I going to do that in a smart way ..
20:20:25 <TrueBrain> there we go .. a small patch later
20:20:41 <TrueBrain> okay .. found 2 major blockers I need to investigate ..
20:20:45 <TrueBrain> 1 AWS related, 1 code related ..
20:20:55 <TrueBrain> owh well .. you can't always have them all :D
21:06:39 *** andythenorth has joined #openttd
21:38:23 <andythenorth> TrueBrain thanks for doing infra stuff :D
22:09:45 *** supermop_Home has joined #openttd
22:21:29 *** Progman has joined #openttd
22:57:28 <supermop_Home> I am quite bored
22:57:38 <supermop_Home> i even have a fair amount of work to do
22:57:46 <supermop_Home> but i do not want to do it
23:08:09 <andythenorth> I won with 2 catapults and 1 peon
23:08:24 <andythenorth> it's not so hard, you can build a barracks immediately
23:08:44 <andythenorth> the hardest is a lot of useless conjurors at start
23:10:58 <supermop_Home> castles 2: siege and conquest?
23:14:18 <supermop_Home> very little of the game seemed to be the battles
23:14:48 <supermop_Home> also i would always try to build too elaborate of a castle that would take years to finish
23:15:17 <supermop_Home> so by the time it got attacked it just had like 4' high unfinished walls
23:16:11 <supermop_Home> most of it was this and having to pay off the pope
23:23:24 <andythenorth> oh I think there was a BBC Basic version
23:23:27 <TrueBrain> meh ... getting IPv6 on ECS is .. more difficult than I would like :(
23:23:36 <andythenorth> I remember building castles too big, so the walls were too low
23:25:10 <TrueBrain> easiest is to upgrade our instance type .. which is vastly more expensive
23:25:14 <TrueBrain> (about 5 times more expensive)
23:25:30 <TrueBrain> still nothing compared to the price of bandwidth, ofc :D
23:31:48 <TrueBrain> well, this is a puzzle
23:33:07 <TrueBrain> but regarding IPv6, basically AWS is working on getting IPv6 to work with ECS, but it is not yet there
23:33:23 <TrueBrain> I can use what is called an awsvpc, by the looks of it anyway, but .. because of the instance type, I have a very limited amount I can use
23:33:38 <TrueBrain> and it is such a small number, I can not really run the service in an easy to maintain form
23:33:49 <TrueBrain> I wonder if I can think of other hacks/tricks ...
23:34:11 <TrueBrain> fake ipv6 interface in the docker with NAT64 out :P
23:34:55 <TrueBrain> it is funny .. OpenTTD supports IPv6 since 2007 ... AWS does not support IPv6 for all common services yet .. and very little documentation / questions about it (showing not many people care) .. it is weird
23:35:25 <andythenorth> my wife seems to get up at 5.30am
continue to next day ⏵
