IRC logs for #openttd on OFTC at 2020-11-12
⏴ go to previous day
00:35:57 *** supermop_Home_ has quit IRC
00:35:58 *** Wormnest_ has joined #openttd
00:40:36 *** WormnestAndroid has quit IRC
00:56:52 *** WormnestAndroid has joined #openttd
01:21:28 <frosch123> TrueBrain: for the morning shift. i pushed the last changes, todo list is empty now :) i'll run the migration with history while i sleep :)
01:33:00 *** snail_UES_ is now known as Guest5602
01:33:01 *** snail_UES_ has joined #openttd
04:07:52 *** Lejving_ has joined #openttd
06:52:40 *** sla_ro|master has joined #openttd
08:14:26 *** longtomjr has joined #openttd
08:46:34 *** ChanServ sets mode: +v tokai
08:47:36 *** andythenorth has joined #openttd
08:47:59 *** HerzogDeXtEr has joined #openttd
08:59:11 <TrueBrain> frosch123: I guess the next sed request would be to change http to https for all tt-forums links :D
09:05:05 <TrueBrain> export is looking really good :D Sweet
09:08:23 <TrueBrain> 8428 http:// links vs 1830 https:// links .. :D Yeah .. that will be some PRs once we are live, to fix that ratio :)
10:08:01 *** Gustavo6046 has joined #openttd
11:24:31 *** frosch123 has joined #openttd
11:35:52 <TrueBrain> so how bad did the export break frosch123 ? :D
11:41:45 *** y2kboy23 has joined #openttd
11:50:24 <frosch123> i realised just enabling something i tested 4 week ago the last time, did not work :)
11:50:32 <frosch123> i fixed it between breakfast
11:51:02 <frosch123> now it is running for 30 minutes, 17k of 85k commits done
11:51:19 <frosch123> 1.4GB memory consumption, i have the entire wiki history/content in memory :p
12:35:34 *** WormnestAndroid has quit IRC
12:35:47 *** WormnestAndroid has joined #openttd
12:55:50 *** WormnestAndroid has quit IRC
12:56:04 *** WormnestAndroid has joined #openttd
13:01:42 *** HerzogDeXtEr1 has joined #openttd
13:20:31 <frosch123> 66k of 85k, looks like it keeps a constant speed
13:20:59 *** andythenorth has left #openttd
13:27:36 <TrueBrain> So another 30 minutes or so
13:35:55 <DorpsGek_III> [OpenTTD/OpenTTD] orudge opened pull request #8340: Draft: Feature: Create Universal (x86_64 + Apple Silicon) build on macOS https://git.io/JkqLt
13:37:03 <DorpsGek_III> [OpenTTD/OpenTTD] orudge updated pull request #8340: Draft: Feature: Create Universal (x86_64 + Apple Silicon) build on macOS https://git.io/JkqLt
13:38:39 <DorpsGek_III> [OpenTTD/OpenTTD] orudge updated pull request #8340: Draft: Feature: Create Universal (x86_64 + Apple Silicon) build on macOS https://git.io/JkqLt
13:41:03 <orudge> Not sure if the homebrew dependencies will be fat binaries though, which might scupper this plan until homebrew is updated (or we can force them to be recompiled?))
13:41:47 <LordAro> will the generated binary still work with OSX?
13:43:19 <orudge> It's still targetted at 10.9
13:43:44 <orudge> I believe you can in theory create a PPC/PPC64/x86/x86_64/aarch64 super-universal binary
13:46:39 <TrueBrain> just when I thought we were done with "universal" builds :P
13:47:23 <TrueBrain> just as a FYI: doing this for CI builds might be a bit pointless and just a waste of CPU time, as the binaries are never available for download; but hopefully it helps for LordAro's branch of moving the nightlies to GitHub Actions :D
13:59:41 <orudge> Handy to ensure that the build isn't broken on aarch64 though
13:59:50 <orudge> and yes, it looks like the homebrew libraries are not universal
14:00:02 <TrueBrain> hence the "a bit" :) there is some gain, for sure
14:00:18 <TrueBrain> I was hoping this time around they would go for a non-universal route ...
14:00:26 <TrueBrain> but they just re-used their old tricks :P
14:01:45 <frosch123> hmm, git log is way faster than i remembered
14:01:56 <TrueBrain> well, that is not a bad thing to find out :D
14:01:56 <frosch123> would lfs slow it down, even when not active?
14:02:39 <TrueBrain> publish your result in gollum-test-data? :)
14:02:51 <TrueBrain> I can do some pre-checks that way .. after that, push to wiki-data :P
14:03:19 <frosch123> i think i fork wiki-data, and push to my account
14:03:43 <TrueBrain> you can PR it after that :P Very curious how BOOM GitHub says :P
14:04:23 <frosch123> won't work. i will drop some of your commits (Main Page etc)
14:04:51 <TrueBrain> (and completely fine)
14:08:21 <frosch123> lol, my fake history contains funny things
14:08:55 <frosch123> the initial "Wiki software successfully installed." main page already contains a "[[Translation:en/Main Page]]"
14:09:14 <TrueBrain> can't start early enough :P
14:09:40 <TrueBrain> how many changes does en/Main Page have? :)
14:10:48 <TrueBrain> I expected more honestly
14:10:50 <frosch123> hmm, that's fewer than on the real wiki
14:11:12 <frosch123> real one has like 70
14:14:16 <TrueBrain> what did those 12 do to not earn a place? :D
14:14:58 <frosch123> hmm, --allow-empty did not work?
14:15:13 <TrueBrain> skipping empty commits is not a bad thing honestly :)
14:15:28 <TrueBrain> so that is totally fine :)
14:15:39 <frosch123> but i did set --allow-empty
14:16:02 <TrueBrain> so by accident you did the right thing? :P
14:16:05 <TrueBrain> the commits can still be there btw
14:16:10 <TrueBrain> but you won't see that in the history
14:16:19 <TrueBrain> as ... the commit was empty :P
14:16:28 <TrueBrain> hence that skipping those is totally fine
14:16:34 <TrueBrain> as nobody is ever going to find them back :P
14:22:29 <frosch123> pff, fetch takes ages. i wonder whether it tries to match the 85k local commits against the 6 remote ones
14:22:48 <frosch123> yeah, now it figured out "no common commits" :)
14:28:54 <frosch123> i never rebased so many commits :)
14:29:11 <TrueBrain> I would assume you would cherry-pick those 2 on top of your work, honestly :P
14:30:08 <frosch123> hmm, ok, i could cherry-pick the one with the actions
14:30:29 <frosch123> well, then forking was pointless
14:31:23 <frosch123> i forgot to add "time" to the rebase, but let's finish this just for the lolz
14:33:40 <TrueBrain> well, guestimate it :P
14:35:27 <frosch123> actually, no, this will take 2.5 hours again
14:35:56 <TrueBrain> yeah ... that is too much lolz
14:38:50 <frosch123> well, i got to the 1.1.0 release article, so maybe it would have been faster
14:39:52 <TrueBrain> not worth the effort and time :)
14:40:51 <frosch123> pushing 489647 objects :)
14:42:03 <TrueBrain> I am still wondering if I have to cache the GitHub checkout
14:42:20 <TrueBrain> I think the total time of clone + metadata scan is going to be over 5 minutes ...
14:42:30 <TrueBrain> depends how GitHub is packing this :)
14:42:42 <frosch123> i ran an aggressive repack
14:42:56 <TrueBrain> GitHub does similar things, I have noticed
14:43:15 <TrueBrain> as you might have noticed, the latest version of TrueWiki does a full metadata scan before the server becomes active
14:43:23 <TrueBrain> I wanted to avoid that, but I couldn't, without losing consistency
14:43:35 <TrueBrain> so the startup deadline is a bit tight
14:43:38 <frosch123> yes, the last check took 50% longer :)
14:43:52 <TrueBrain> yeah, sorry about that .. I postponed it as long as I could :)
14:44:02 <TrueBrain> funny, for me it was 50% too :P From 4 minutes to 6
14:44:15 <TrueBrain> but I/O in a Docker is crazy slow ...
14:44:19 <frosch123> your machine is twice as fast as my >10 old one
14:44:44 <TrueBrain> my single code speed can be faster :P
14:44:58 <TrueBrain> peeks at 5GHz or something
14:45:08 <frosch123> yeah, that's twice as fast :)
14:45:18 <TrueBrain> it is funny ... the specs say it is 3.7GHz
14:45:24 <TrueBrain> in normal operation, it runs at 4.6GHz
14:45:29 <TrueBrain> in stress, it goes higher
14:45:37 <TrueBrain> best investment ever :)
14:46:06 <TrueBrain> oef, okay, I see your point :)
14:46:14 <TrueBrain> what is "fun" about the metadata scan, in a Docker, it is I/O bound
14:46:21 <TrueBrain> outside docker, it is CPU bound
14:46:29 <TrueBrain> somehow we ended just on the edge of what starves the system
14:46:53 <frosch123> oh, i run it inside docker
14:47:04 <TrueBrain> ah .. it is really a lot slower in a Docker
14:47:07 <frosch123> put cpu+ram is still the oldest part of my system
14:47:22 <TrueBrain> mainly as I fully depend on kernel cache, honestly
14:47:38 <TrueBrain> when I add an LRU on the template loading, it is faster in a Docker
14:47:40 <TrueBrain> but .. ugh .. what-ever :P
14:47:56 <TrueBrain> (so those pages that include a lot of the same template, as all opened every single time again and again)
14:49:02 <TrueBrain> your uplink is also limited, I guess :P
14:49:41 <TrueBrain> I guess I shouldn't say I can upload with ~80 MB/s :P
14:49:50 <frosch123> for some reason there is factor 10:1 between down/upload
14:49:59 <TrueBrain> one of the perks of living in a city that close to AIX :)
14:50:25 <TrueBrain> 10:1 is normal .. in The Netherlands you are now seeing this quickly change
14:50:28 <frosch123> i don't need faster download, and faster upload is not worth the price
14:50:46 <TrueBrain> and even cable is changing to 5:1 or 2:1
14:51:47 <TrueBrain> owh, the cable provider here changed to 20:1 .. that is unusual
14:51:54 <TrueBrain> for their 1gbit connection, that is
14:52:09 <TrueBrain> okay, let me run some tests to see how this will run on our cluster
14:52:15 <frosch123> This branch is 84502 commits ahead, 5 commits behind OpenTTD:master :)
14:52:48 <TrueBrain> 30 MiB/s download, not bad :)
14:53:26 <TrueBrain> okay, that works without any incident what-so-ever
14:53:50 <frosch123> i wonder whether i should squash those "cleanup post-import" commits :p
14:54:09 <TrueBrain> owh, you took the commit message directly from mediawiki, not any "modified: NNN" stuff
14:54:11 <frosch123> also, why does it still show author and committer separately
14:54:12 <TrueBrain> I guess that is fine :)
14:54:37 <TrueBrain> wow, that are a lot of cleanup post imports :D
14:55:22 <frosch123> i just attached a revision to every page
14:55:54 <TrueBrain> I am shocked how quick GitHub is in showing this all
14:56:15 <frosch123> TrueBrain: i trashed those categories which are equivalent to a folder
14:56:28 <frosch123> or which were plain pointless
14:56:31 <TrueBrain> awh, the mediawiki render of GitHub -almost- does the right thing
14:57:35 <TrueBrain> LordAro authored and LordAro committed on Mar 17, 2019
14:57:53 <TrueBrain> sometimes it has author and committer
14:57:55 <TrueBrain> sometimes it doesn't?
14:58:04 <frosch123> i set both author and committer to the same value
14:58:13 <frosch123> before it set myself as committer
14:58:22 <frosch123> so, no idea what is the correct setting
14:58:41 <TrueBrain> well, it does this for some commits, not for others
14:59:25 <DorpsGek_III> [OpenTTD/OpenTTD] orudge commented on pull request #8340: Draft: Feature: Create Universal (x86_64 + Apple Silicon) build on macOS https://git.io/JkqcZ
14:59:36 <TrueBrain> I mean ... the first 4 are not
14:59:44 <TrueBrain> and half-way through it is not
14:59:49 <TrueBrain> and I cannot spot why it does this
15:00:55 <TrueBrain> orudge: I hope you can find a workaround for that :D
15:02:24 <TrueBrain> I am looking through the history of other GitHub repos .. I love how it picks the "committer" from your GitHub settings :P
15:03:57 <frosch123> TrueBrain: i claim, github detects names matching org members
15:04:26 <frosch123> hmm, though it's in my account
15:04:26 <TrueBrain> so why doesn't he do it for you, but does for me, as example?
15:04:42 <frosch123> i thought "different spelling/capitalisation"
15:04:57 <orudge> TrueBrain: I imagine there'll be an updated Homebrew available soon; alternatively MacPorts may offer a solution. Once folk start getting their hands on the new hardware and/or Big Sur (which is released today) then I imagine things will start to move.
15:05:11 <orudge> I also have a PR to finish for ARM64 Windows support and builds...
15:05:36 <TrueBrain> now we just need to get ride of Azure Pipelines and do it all in GitHub Actions :D
15:06:06 <TrueBrain> frosch123: okay, everything seems to work from my perspective
15:06:13 <TrueBrain> I guess the only remark I have is the commit messages
15:07:14 <TrueBrain> (not the author vs commiter btw, I couldn't care less how GitHub shows it .. I am just curious why it shows it like this, but that is all)
15:08:29 <frosch123> i would like to keep the wiki message. but i can prepend the change/add/whatever you did
15:08:45 <frosch123> i can also set committer to librarian for everything
15:08:58 <TrueBrain> hmm .. I guess that would make it more consistent with the future commits
15:09:18 <frosch123> TrueBrain: btw. when you upload a file. do you create one commit for .png + .mediawiki, or two?
15:09:20 <TrueBrain> but mostly it would explain the post-import commits
15:09:30 <TrueBrain> I believe one .. lets find out :)
15:09:57 <frosch123> nah, i want to squah the post-import commits into one
15:10:57 <TrueBrain> commit_message = f"new page: {old_page}"
15:10:58 <TrueBrain> commit_message = f"modified: {old_page}"
15:11:03 <TrueBrain> commit_message = f"renamed: {old_page} -> {new_page}"
15:11:35 <TrueBrain> we can also change how TrueWiki commits btw, which is also fine by me
15:11:51 <TrueBrain> I was thinking a bit in the future, if (when?) we add history to TrueWiki itself
15:11:58 <TrueBrain> having consistent commit messages helps in that case
15:12:24 <TrueBrain> what I like about wiki-data now, that per file you see a clear history
15:12:32 <TrueBrain> but the commit history is .. a mess
15:12:39 <TrueBrain> with wiki-data-staging, it is the other way around
15:12:41 *** andythenorth has joined #openttd
15:13:36 <frosch123> i think putting the new page/modified (i do not have renamed) first, and then appending the old message (if any) is best
15:13:50 <frosch123> then it's the same for commits without explicit message
15:14:07 *** arikover has joined #openttd
15:14:08 <TrueBrain> sadly, it does mean doing an export again, but I guess such is life :P
15:14:21 <TrueBrain> if I can help by doing to run or what-ever, that is fine by me; but we are in no rush, so what-ever :)
15:15:24 <TrueBrain> damn, commit are as fast on 15k commits as on 80k commits
15:15:39 <TrueBrain> guess that is not totally surprising, given linux-kernel has most likely even more commits :P
15:15:51 <frosch123> i don't have to watch it doing the commits, i don't care how long it takes :)
15:16:16 <TrueBrain> my hands are itching to fix so many inconsistencies .... :P
15:16:30 <frosch123> do you know what the github daily push quota is?
15:16:38 <frosch123> when do i get banned?
15:16:54 <TrueBrain> from what I understand, they do not really care
15:17:04 <TrueBrain> at a certain point, they contact you, like: wtf dude
15:17:30 <TrueBrain> I am pretty happy how the new wiki turned out
15:17:36 <TrueBrain> also the new layout I kinda fancy
15:17:46 <TrueBrain> (as in, fixed width, instead of what-ever screen size you have)
15:17:52 <TrueBrain> as on a 2k monitor, it becomes ... annoying to read
15:18:19 <frosch123> why do you maximize your browser?
15:18:26 <frosch123> i have browser+irc on one screen
15:18:46 <TrueBrain> because Windows doesn't have the handy CTRL+ALT+6 stuff
15:18:50 <TrueBrain> one thing I really do miss
15:18:52 <TrueBrain> like .. really do miss
15:19:32 <frosch123> haha, i have win at work, so i also miss stuff there :)
15:19:40 <frosch123> for some reason i do not miss anything the other way around
15:19:46 <LordAro> TrueBrain: ctrl+alt+6 stuff?
15:19:55 <TrueBrain> LordAro: quickly move windows to a side of your screen
15:20:22 <TrueBrain> I think it is the CSP causing the link to not render ..
15:20:29 <LordAro> well it can do some of it with win+arrow keys
15:20:43 <TrueBrain> "some" being the prober term
15:20:50 <frosch123> TrueBrain: if you show different icons for http and https, lease reverse the logic compared to mediawiki
15:21:04 <frosch123> mediawiki makes https look weird
15:21:10 <frosch123> it should be the other way around :)
15:21:19 <TrueBrain> but there should be an icon already
15:21:36 <TrueBrain> LordAro: fancy .. will check it out :)
15:22:33 <TrueBrain> background: url(data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAoAAAAKCAYAAACNMs+9AAAAVklEQVR4Xn3PgQkAMQhDUXfqTu7kTtkpd5RA8AInfArtQ2iRXFWT2QedAfttj2FsPIOE1eCOlEuoWWjgzYaB/IkeGOrxXhqB+uA9Bfcm0lAZuh+YIeAD+cAqSz4kCMUAAAAASUVORK5CYII=) center right no-repeat;
15:22:47 <TrueBrain> I would expect a CSP warning if it is CSP
15:23:04 <TrueBrain> Chrome does indicate it is a CSP error
15:23:20 <TrueBrain> well, guess it is time to replace it with an UTF-8 char
15:23:43 <TrueBrain> what UTF-8 char to use ...
15:24:34 <TrueBrain> looks VERY weird on iOS :)
15:24:59 <frosch123> wtf. your os choice gets worse and worse
15:25:13 <TrueBrain> shows you how it renders
15:27:21 <TrueBrain> owh, it is one of those UTF-8 chars that cannot be coloured
15:27:26 <TrueBrain> ugh ... do we have a better one? :P
15:27:48 <frosch123> there are many arrows. but no real link icon
15:28:09 <TrueBrain> something to indicate it is external, that we do need
15:28:12 <TrueBrain> just anything, honestly
15:29:08 <TrueBrain> happy iOS refuses to render it :P
15:30:20 <frosch123> 1f500 to 1f51d are super weird
15:30:39 <FLHerne> TrueBrain: And of course it's different from 'three lines converging left'
15:31:04 <frosch123> i think one of the arrows is our best bet
15:33:01 <frosch123> put the icon in-front of the link imo
15:36:36 <frosch123> is that the same font size? i expected it to be full height
15:36:43 <TrueBrain> full font looked odd
15:37:21 <TrueBrain> owh, it was still selected, owh well
15:37:51 <TrueBrain> making it a bit smaller feels better
15:37:55 <TrueBrain> but I think the icon is sufficient
15:37:59 <TrueBrain> putting it before the link also looks odd
15:38:23 <TrueBrain> right, one more icon I need .. one for the chapter anchors
15:38:26 <TrueBrain> when you hover over a chapter
15:38:31 <TrueBrain> it should show an anchor on the left
15:38:37 <TrueBrain> also an inline data: element :)
15:39:00 <TrueBrain> good enough for now
15:39:05 <TrueBrain> I will wait for real CSS people to make it proper
15:39:15 <TrueBrain> cannot be coloured, those chars
15:39:17 <TrueBrain> which look very odd
15:39:51 <frosch123> rest for any then :p
15:41:30 <frosch123> no pony, and then css
15:45:20 <frosch123> oh, you can set font sizes in percentage now?
15:45:36 <TrueBrain> you can for a long long long time
15:45:38 <frosch123> last time i did font sizes in html, i had to use numbers -7 to 7 or so
15:45:44 <TrueBrain> it used to be something you didn't want to do
15:45:49 <TrueBrain> but these days, using em/%
15:46:05 <TrueBrain> it makes mobile experience better
15:46:23 <TrueBrain> have you tried visiting the wiki via mobile?
15:46:47 <TrueBrain> you can kinda tell what is pixel based and what is em based :D
15:47:29 <TrueBrain> it does need some more work btw, but .. it is not terrible at least
15:57:56 <TrueBrain> right, so now we wait for the export .. some quick pre-check .. and after that we can bring it to wiki-new :)
15:59:06 <frosch123> do we keep it read-only for a few days after switching wiki.openttd.org? just to see how it takes the load?
15:59:36 <TrueBrain> hmm .. pros and cons to that
15:59:47 <TrueBrain> something new often also comes with people willing to change things
15:59:56 <TrueBrain> but indeed, if it doesn't hold, having it read-only allows us to roll back
16:01:43 <frosch123> hmm, though since we keep the old one readonly anyways, it does not matter
16:02:00 <TrueBrain> well, worst case, we have to duplicate changes made in the new
16:02:02 <frosch123> if we switch back we essentially show an older version
16:02:36 <TrueBrain> yeah .. lets just switch to a wiki in read/write mode
16:02:45 <TrueBrain> if it cannot handle the load ... we have to work fast to fix it :P
16:02:56 <TrueBrain> the load of 404s will be heavy :D
16:03:23 <frosch123> oh, we should record those
16:03:32 <frosch123> then we actually know what is linked externally
16:04:23 <frosch123> website also has some links, we need a pr for it
16:05:47 <frosch123> sure, shouldn't be much
16:06:02 <frosch123> just something we should not forget :)
16:06:06 <TrueBrain> 12 hits; but you know better than me where they are now :D
16:08:08 <frosch123> oh damn, you tricked me into writing the new post as well
16:08:29 <TrueBrain> haha, I did not think of that yet,but no, you just did that yourself >:D
16:08:56 <TrueBrain> I was thinking we could also poke the Translations team, asking help in translating the wiki
16:09:03 <TrueBrain> but that is maybe something for a bit down the road
16:09:36 <frosch123> yep, i want to write a new main page on staging, one that has less links to Archive :p
16:09:49 <TrueBrain> that is a perfect use of Staging :)
16:10:03 <TrueBrain> we should also move Sandbox there, for example
16:10:50 <frosch123> well, i trashed it, but linking there is a good idea
16:11:08 <frosch123> but that means we also have to put all the templates on staging at some point
16:29:51 <frosch123> turns out (a1, b1, c1 == a2, b2, c2) needs more () to mean what i want :p
16:37:28 <TrueBrain> so, restart of the export? :P
16:39:08 <TrueBrain> lowercase the email, I would say
16:39:22 <TrueBrain> and yes, looks excellent
16:39:40 <TrueBrain> not sure about IPs in the history like this
16:39:51 <TrueBrain> maybe we should change them into "anonymous" or something? Not sure
16:41:09 <frosch123> is testing the first char for digit enough?
16:41:55 <TrueBrain> there are ~20 usernames starting with a digit
16:42:31 <frosch123> there are also ipv6 ones
16:42:42 <TrueBrain> start with digit and either "." or ":" in them
16:43:27 <TrueBrain> or I can give you the usernames starting with digit :P
16:43:42 <frosch123> i have them myself :)
16:44:09 <TrueBrain> of which a few never authenticated, lol
16:44:54 <TrueBrain> 5 usernames starting with a digit made edits
16:44:57 <TrueBrain> to word it properly :)
16:45:16 <TrueBrain> there is also a "arie" (I did not add the quotes .. his username has quotes)
16:45:31 <TrueBrain> and 2 starting with a ^
16:46:07 <frosch123> i only have 4 on the wiki
16:46:53 <TrueBrain> SELECT * FROM `user` WHERE user_editcount > 0 ORDER BY `user`.`user_name` ASC
16:47:13 <TrueBrain> 02walshe , 07es26 , 11Runner , 2talltyler , 7036
16:47:19 <frosch123> arie is funny. python or git ate some of the quotes
16:47:34 <frosch123> oh, i missed 7036 :)
16:47:47 <TrueBrain> 1 edit, database says
16:49:58 <frosch123> should i slugify them?
16:51:32 <TrueBrain> that solves your problems for sure :)
16:51:37 <TrueBrain> well, for email-addresses, that is
16:51:46 <frosch123> i'll slugify the email
16:51:53 <frosch123> i don't like the lowercase for names
16:52:51 <TrueBrain> I meant only for emails, to be clear
16:53:01 <TrueBrain> owh, also what I said, pfew :P
16:53:13 <TrueBrain> but the slugify will do that for you :) (or at least, should :P)
16:55:33 <frosch123> haha, just noticed that the first page created after "main page" is already in archive :p
16:56:48 <frosch123> let's blame Dominik
17:01:00 <DorpsGek_III> [OpenTTD/OpenTTD] James103 commented on issue #8339: [Game-play] Stopped vehicles started after reloading network game https://git.io/JkTYo
17:02:52 <TrueBrain> think we did enough bikeshedding :)
17:03:21 <TrueBrain> if you already start with the 2 commits from origin/master, you don't have to fix things afterwards :P
17:03:34 <TrueBrain> but that is really tomato tomato :D
17:04:02 *** Wormnest has joined #openttd
17:04:04 <TrueBrain> no, what you have now is better
17:04:11 <TrueBrain> the current commit 0 is at the wrong date
17:04:13 <TrueBrain> so that would be bad
17:10:10 <frosch123> i picked the date of the initial commit very carefully
17:15:07 <TrueBrain> Do I want to know why it is months before the first page? :p
17:26:34 <frosch123> aw, i thought you would recognise it immediately
17:28:13 <frosch123> 13k/79k, afk now :)
17:55:31 <TrueBrain> Ah, yeah, the date I could understand, but the time took a bit longer :D
18:26:31 <frosch123> haha, i don't think we have fan boys who would know the time :)
18:28:27 <TrueBrain> Not a bad thing honestly :p
18:32:45 *** Progman has joined #openttd
18:51:54 <DorpsGek_III> - Update: Translations from eints (by translators)
19:06:58 <longtomjr> awh man, did I miss a bikeshedding session.
19:11:34 <frosch123> pretty sure andy has some yaks left over for shaving
19:13:05 <andythenorth> am I volunteered for something?
19:14:27 <longtomjr> I read that as 'needs a map' for a second
19:27:27 <andythenorth> did anyone release a FIRS 4 beta yet?
19:28:20 <TrueBrain> you really do not want me to :P
19:31:34 <andythenorth> well I guess I'll do it :P
19:31:46 <TrueBrain> sounds like the better plan, yes
19:31:54 <andythenorth> beta support on bananas?
19:32:01 <andythenorth> 'this content is probably broken' :P
19:32:26 <TrueBrain> aren't all GRFs beta?
19:32:55 <TrueBrain> please, never do that again :P
19:33:00 <TrueBrain> it just cracks me up every time
19:33:24 <andythenorth> do I miss something :P
19:34:35 <TrueBrain> it isn't Friday yet!
19:34:42 <TrueBrain> I lose count sometimes ..
19:34:49 <longtomjr> call Japan, maybe it is
19:35:47 <longtomjr> 5 o clock somewhere, and friday somewhere
19:35:49 <TrueBrain> does Japan has a phone? :P
19:35:57 <TrueBrain> just one HUGE phone in the middle of the country
19:36:05 <TrueBrain> my imagination is weird :P
19:36:18 <longtomjr> How else do countries communicate with each other
19:36:41 <TrueBrain> as the earth is flat, they never have to travel far
19:37:03 <longtomjr> ah right, forgot about that
19:37:35 <TrueBrain> pretty sure that was for the best, to forget about that :P
19:37:54 <longtomjr> never forgot about the flat earth, just the pigeons
19:38:10 <longtomjr> but I am sure next time I walked over to my car the pigeons would have reminded me on my windscreen
19:38:36 <TrueBrain> They are sending you a message!
19:38:59 <longtomjr> It is what they do best I heard
19:39:09 <longtomjr> that and planting seeds all over the place
19:42:34 <longtomjr> what am I looking at
19:43:03 <longtomjr> fewer users = more edits. More users = fewer edits?
19:43:36 <frosch123> 10 users have > 1000 edits, 100 users have > 100 edits
19:44:02 <longtomjr> is that per user or just total edits on the wiki?
19:44:07 <frosch123> > 50% of users have 1 edit
19:44:28 <frosch123> longtomjr: every position in the X axis is one user
19:44:39 <frosch123> they were so nice to sort themself by number of edits
19:45:22 <longtomjr> thanks for explaining
19:46:05 <TrueBrain> frosch123: you want all math on us again, haven't you? :D
19:46:32 <frosch123> 68k of 79k, what else shall i do?
19:46:44 <TrueBrain> y = 10000 / x, nice :D
19:46:52 <TrueBrain> good to know we have an upper limit :D
19:48:05 <frosch123> TrueBrain: you only have the 17th most edits on the wiki
19:48:08 <TrueBrain> do I dare to ask who is #1 and #2?
19:48:17 <Eddi|zuHause> frosch123: my initial thought is towards a poisson distribution
19:48:21 <TrueBrain> wuth? 17th? Holy crap .... can't remember I added anything useful to it :P
19:48:39 <longtomjr> is the pre, or post cleanup?
19:48:42 <Eddi|zuHause> frosch123: but i'm not really familiar with that
19:48:54 <frosch123> i think #1 is a french translator (not the one you know), #2 is our favorite pl translator
19:49:05 <TrueBrain> translations, well, that is nice
19:49:12 <TrueBrain> can you make the graph with only english pages?
19:49:41 <frosch123> hmm, i guess, git log Page/en will do
19:49:43 <TrueBrain> I am happy btw that 50+% only makes a single edit; their contribution was something that annoyed them, and it is good to see they could fix it :)
19:50:16 <frosch123> well they are also unregistered :) so mandatory login stopped that
19:50:23 <TrueBrain> I am surprised btw that the GitHub runners are still not running into the Docker limiter .. I expected that to happen already ..
19:51:05 <longtomjr> Maybe GH has some caching on their side
19:51:28 <TrueBrain> could be, or they have enough public IPs they cycle through ..
19:51:32 <TrueBrain> or they have an agreement with Docker Hub, ofc
19:51:46 <TrueBrain> I asked the Docker Library dudes if they can also publish on GHCR, but .... no answer
19:51:46 <frosch123> TrueBrain: lol, it actually does not change much. i guess you also have to edit every english page to link the translation:p
19:52:04 <TrueBrain> frosch123: but that is stripped out of Page/en, not?
19:52:23 <TrueBrain> (As they would be empty commits)
19:52:26 <frosch123> yeah, i hope, not sure
19:52:40 <TrueBrain> but I would expect only the top few to drop out
19:52:45 <TrueBrain> and that nothing really changes on the grap
19:53:07 <frosch123> TrueBrain: anyway, only counting english you are taken over by many known people from this channel
19:53:10 <TrueBrain> as my assumption is that only a few care about translations, where the many care about the english
19:53:17 <frosch123> looks like you edited a lot of non-english pages
19:53:23 <TrueBrain> well, this month, sure
19:53:42 <TrueBrain> I never understood mediawiki, so I always avoided doing anything with it honestly :P
19:54:05 <longtomjr> just have a md -> mediawiki -> md script handy
19:55:17 <longtomjr> can you set the bg color with md
19:55:25 <longtomjr> otherwise nvmd, bad idea
19:55:27 <TrueBrain> curious how long GitHub is going to do about making its insight graphs :D
19:55:48 <TrueBrain> you can clearly see our activity spike frosch123 :D
19:56:08 <TrueBrain> 2006-2010 were quiet .. 2010-2012 were busy ... up to 2014 there was some activity .. died out till 2020 :P
19:56:51 <TrueBrain> awh, Contributors only works for accounts that are linked to GitHub
19:57:35 <longtomjr> Since you have the infrastructure stats, is Ottd growing, going steady or declining in popularity? Or is it too difficult to say?
19:57:43 <TrueBrain> it seems I did 500 edits this month :P
19:57:56 <TrueBrain> longtomjr: very difficult to say in real useful numbers
19:58:07 <TrueBrain> but our bandwidth has been kinda the same since .. 2011 or something
19:58:09 <longtomjr> Ah ok, maybe that is for the best
19:58:14 <glx> usually it's very flat with some peaks
19:58:26 <TrueBrain> frosch123: yeah, what I would expect :)
19:58:45 <TrueBrain> longtomjr: few months back we got posted on Hacker News
19:58:46 <longtomjr> it is kinda trippy what it does towards the end
19:58:51 <TrueBrain> that increased traffic with 50% for a while
19:59:14 <longtomjr> I mean I were inactive for quite a while, and I am now playing more than ever.
19:59:17 <TrueBrain> we had more of those spikes .. where our average hit/s went from ~10 per second to ~150 per second
19:59:38 <TrueBrain> but we don't really collect stats
20:00:01 <TrueBrain> some people crawl the multiplayer stats, not sure what they are telling
20:00:13 <TrueBrain> but from my perspective, OpenTTD is as popular as it was 10 years ago
20:00:47 <TrueBrain> (despite there being almost 0 development activity in the official repo the last few .. months now?)
20:01:10 <longtomjr> it just means it is stable :)
20:01:20 <TrueBrain> it doesn't, but I like how you think :)
20:01:40 <TrueBrain> JGR mostly keeps the more advanced players interested
20:01:49 <TrueBrain> GRFs do the job for another group
20:01:53 <TrueBrain> new players always come and go
20:02:38 <longtomjr> Looks like an example of stability :)
20:03:00 <TrueBrain> but that kind of stability for games is rarely a good thing
20:03:06 <TrueBrain> and you see that in the OSX support, for example
20:03:19 <TrueBrain> and I think we still run an ICU version that is not really supported anymore?
20:03:29 <TrueBrain> we create HTTP (not HTTPS) connections
20:03:39 <TrueBrain> there is this "hard stop" on this kind of stability
20:04:04 <longtomjr> Yep, unfortunately the world keeps changing, creating the need to change stable software
20:04:06 <TrueBrain> I just wish we could help JGR a bit more with infra
20:04:34 <longtomjr> What is missing on the infra side that would help JGR?
20:04:45 <TrueBrain> well, we created this new infra that we can also host any patchpack
20:05:01 <TrueBrain> we can build binaries, publish them on the CDN, announce them, etc
20:05:29 <TrueBrain> a while back (over a year) we tried to get him to update some configuration scripts so we could, but he clearly is not that knowledge about these things :) (which is not a complaint to him)
20:05:39 <TrueBrain> we just have nobody bridging between us and him to close that gap
20:06:14 <longtomjr> Where is that code sitting currently?
20:06:28 <TrueBrain> Azure Pipelines currently builds binaries
20:06:39 <TrueBrain> which hopefully is going to be GitHub Actions soon
20:06:49 <TrueBrain> I hope that helps JGR too to make the changes needed in his repo to publish
20:06:53 <TrueBrain> (as GitHub Actions are just easier)
20:08:21 <longtomjr> Is the plan to migrate the pipelines to actions?
20:08:33 <TrueBrain> well, I hope that is the plan :D
20:08:54 <longtomjr> yep just look, but the build steps not yet
20:08:54 <TrueBrain> releases is a bit more fiddling around
20:09:08 <TrueBrain> but GitHub Actions are just easier to work with, especially in forks etc
20:09:10 <longtomjr> Might be nice to have docker releases as well
20:09:18 <TrueBrain> it would just need some secrets from me, and it will publish :)
20:09:25 <TrueBrain> "docker releases", that is kinda undefined :P
20:09:43 <TrueBrain> if you want an OpenTTD server via Docker, there are 2 versions out there that I know of
20:09:45 <TrueBrain> most likely more :)
20:10:38 <longtomjr> I meant image release, but there is no real need if there is existing ones.
20:10:50 <TrueBrain> the server-market is also very niche
20:10:54 <TrueBrain> they can take care of themself :)
20:11:25 <TrueBrain> ha, cool: if you search for jgrpp openttd, the 3rd hit on duckduckgo is the server listing for JGRPP server specific
20:11:33 <TrueBrain> exactly why we made that feature, happy it also works :D
20:12:08 <TrueBrain> but basically, from an infra perspective, the "official" binaries are just a setting
20:12:16 <TrueBrain> it can handle any amount of patchpacks, and can switch what-ever is promoted
20:13:16 <TrueBrain> I mean, JGR is using PRs to merge translations .. which is fine, but we build a translator system for that
20:13:29 <TrueBrain> I would love to link up those two, so his patchpack is also translated by the ones doing "official"
20:13:58 <TrueBrain> but okay: migrate everything to AWS first (3 things left), condense the community after :)
20:14:43 <longtomjr> are those things on the trello?
20:15:17 <longtomjr> yep, not the other plans.
20:15:24 <TrueBrain> no ... 1 thing at the time :)
20:15:46 <TrueBrain> if people THINK you are going to do something, they won't :P
20:16:06 <longtomjr> hehe just saw you move things around
20:16:19 <TrueBrain> yeah, noticed it wasn't up-to-date
20:16:35 <TrueBrain> I like looking at what is done (aver the last 2 years)
20:16:39 <TrueBrain> vs what needs doing
20:16:43 <TrueBrain> puts things in perspective
20:17:26 <TrueBrain> the insane amount of time that went into all of that ... I don't even dare to count the hours frosch123 put into this migration of wiki data :)
20:20:59 <TrueBrain> talking about GitHub Actions, LordAro , how are the releases going? Getting somewhere? I guess in a few weeks you would like to ship a 1.11 beta?
20:23:21 <JGR> Reading up a bit, what specifically did you have in mind as regards JGRPP and infrastructure?
20:24:44 <TrueBrain> still the same as we talked about over a year ago, I guess :)
20:25:06 <TrueBrain> mostly of course a please where binaries can be produced for all the OSes
20:25:13 <TrueBrain> (and stored, and distributed)
20:25:44 <TrueBrain> but sky is the limit .. currently there are GRFs that possibly only works with jgrpp .. we have backend support for that in BaNaNaS, but it needs a bit of a tweak in the protocol
20:26:15 <longtomjr> I know of a few that only works in JGRPP, bridge grfs
20:26:17 <TrueBrain> we have a news section that could use: new JGRPP released
20:26:28 <TrueBrain> and I am pretty sure if you give me long enough, that list will be a lot longer :P
20:27:03 <TrueBrain> a please where? a place where .. wow, typing is hard ..
20:27:12 <TrueBrain> I don't like this is not Discord ... :P
20:27:31 <longtomjr> no message editing?
20:27:44 <TrueBrain> my Matrix client allows it
20:27:48 <TrueBrain> but weird shit happens if I try :P
20:27:49 <longtomjr> Yep, I know, but is that what you miss?
20:27:57 <JGR> Ideally, all GRFs which use extra features in JGRPP would also work to at least some level in trunk, I'm not sure that quite worked though
20:28:35 <longtomjr> I can test the bridge GRF, dunno how it behaves in trunk
20:28:41 <TrueBrain> in general, something that was said back in the beginning of 2018, openttd.org could be a bit more verbose to other packpacks out there
20:28:59 <TrueBrain> at least, that is how we rewrote most stuff
20:29:31 <TrueBrain> and, well, there is only 1 active patchpack :D
20:29:38 <JGR> The latest MWBS bridge GRF should work in trunk, all the others, probably not
20:29:52 <JGR> On binaries, having MacOS builds would be nice
20:30:08 <TrueBrain> yeah .. but I suggest we wait till we migrated fully to GitHub Actions
20:30:24 <TrueBrain> after the next sync, that would mean your fork only needs some secrets
20:30:31 <TrueBrain> and the rest will be complete magic :P
20:30:41 <TrueBrain> I love GitHub Actions :) I really do
20:30:48 <TrueBrain> they make my life so much less complicated
20:32:06 <TrueBrain> well, "after the next sync" -> "after OpenTTD/OpenTTD is switched fully to GHA, after the next sync, .."
20:32:55 <JGR> I've no objections in principle to getting involved in that sort of thing
20:33:11 <TrueBrain> good :) Well, as I mentioned earlier, we first have an AWS migration to finish
20:33:15 <TrueBrain> which will take a few more weeks
20:33:26 <TrueBrain> after that, we can hopefully connect the communities a bit, where useful, ofc :)
20:34:15 <TrueBrain> in most of the infra there are now all kinds of way to add these kinds of filters
20:37:01 <JGR> TrueBrain. Yes, I've seen that, it is rather nice of you
20:37:16 <JGR> I loosely follow this channel via the logs, even if I don't usually log in.
20:37:18 <TrueBrain> I would expect such GRFs to simply only show up on BaNaNaS for download in the JGRPP build, tbh :P But I might be a bit radical in that :D
20:37:49 <longtomjr> I am off to bed. Night all!
20:37:54 <TrueBrain> sleep well longtomjr
20:38:18 <TrueBrain> holy crap, that looks cool
20:38:42 <frosch123> TrueBrain: insane catenary :)
20:38:56 <JGR> On the BOBB bridges, you probably need to use a railset GRF as well to get the rail on top of bridges, longtomjr
20:39:09 <TrueBrain> is there an echo echo echo echo
20:39:17 <TrueBrain> shit, there is! is .. isss ... iiissss
20:39:19 <Koala> I think there might be be be
20:39:42 <Koala> Maybe you could help how me around?
20:39:52 <TrueBrain> well, here on the left is the bar; drinks are free
20:39:52 <longtomjr> JGR, aah ok, thanks. :)
20:39:57 <TrueBrain> on the right is a window ... not much of a view
20:40:01 <TrueBrain> and longtomjr is leaving for the night
20:40:19 <Koala> So you guys come here often then?
20:40:32 <TrueBrain> (it is funny because it is true)
20:40:43 <Koala> Lol, you might see me here often:D
20:41:04 <Koala> Is there some sort of registering I need to do?
20:41:10 <LordAro> TrueBrain: i've not looked at it in a while
20:41:32 <TrueBrain> LordAro: was mostly curious :)
20:41:36 <Koala> What's this, a youtube video!
20:42:01 <TrueBrain> I still claim that the artist should be "Matchbox 20", but that argument is old
20:42:51 <Koala> TrueBrain: Should I have registered or anything?:D Just don't wanna break any rules:P
20:43:37 <TrueBrain> just let us know if we can help with anythoug
20:43:40 <Koala> So is there a social hierarchy here? xD
20:43:56 <TrueBrain> well, yes, we have 1003 kicks in this channel, and I hate I caused the last 3
20:44:00 <TrueBrain> 1000 was such a nice number
20:44:19 <TrueBrain> otherwise we have andythenorth at the bottom of the hierarchy
20:44:25 <Koala> You mean you kicked people or?
20:44:25 <frosch123> Koala: only rule: if you want to talk the talk, you have to walk the walk
20:44:26 <TrueBrain> and LordAro on the top
20:44:45 <TrueBrain> Koala: yeah ... andythenorth was not being nice to me :(
20:44:45 <Koala> Well, where am I walking:D
20:44:52 <TrueBrain> can't remember why/what, but that is not relevant
20:44:55 <TrueBrain> the number is 1003 :(
20:45:06 <LordAro> TrueBrain: 1003 since how long?
20:45:14 <TrueBrain> LordAro: since the beginning of time
20:45:27 <TrueBrain> well, not really .. believe it was in logs till 2007
20:45:31 *** gelignite has joined #openttd
20:45:47 <LordAro> what an excitable person
20:45:52 <TrueBrain> frosch123: sweet! Let me quickly run some tests :)
20:45:56 <LordAro> remember when we were like that?
20:46:56 <TrueBrain> frosch123: pfft .. "Cleanup post-import" breaks the commit checker! HOW DARE YOU :P
20:47:08 <andythenorth> was that software?
20:47:20 <frosch123> most prs will look like that
20:47:29 <TrueBrain> Cleanup: post-import
20:47:35 <TrueBrain> commit-checker refuses that commit message :P
20:47:43 <frosch123> what about the first commit then?
20:47:53 <LordAro> andythenorth: well they connected via the OFTC web interface, so probably unlikely
20:48:01 <TrueBrain> frosch123: makes me wonder if the last commit should have that "Add:" :P
20:48:15 <TrueBrain> LordAro: was part of the test :)
20:48:42 <TrueBrain> okay, everything starts and works as I would expect ..
20:48:51 <LordAro> *apparently connected
20:48:51 <frosch123> ok, i reword the cleanup
20:48:53 <TrueBrain> for good measure, lets run --validate-all
20:49:44 <TrueBrain> history looks good, author/commiter looks good
20:50:05 <TrueBrain> frosch123: I think a bigger question is: will we install the commit-checker on that repo
20:50:27 <frosch123> no idea what it should check :)
20:50:44 <andythenorth> ok I played enough tanks for now
20:50:49 <andythenorth> what shall I do now?
20:50:53 <frosch123> certainly not --validate-all, since you cannot really detect what errors were present before, or added new
20:51:09 <TrueBrain> frosch123: not true; I think what I will add, is a run on master, and a run on the PR
20:51:12 <TrueBrain> and tell the difference
20:51:17 <TrueBrain> what CodeQL also does
20:51:24 <TrueBrain> (there is a reason it outputs YAML :P)
20:51:26 <andythenorth> is it FIRS 4 Alpha 4, or Beta 1?
20:52:32 <TrueBrain> frosch123: and we can cache the master result in the Actions cache, so that would be relative quick too :P (~5 minutes, I guess)
20:53:02 <TrueBrain> but ... not for now :)
20:54:01 <TrueBrain> okay, all my prechecks say: WE SHOULD DO THIS
20:54:06 <TrueBrain> so ..... push to upstream?
20:54:31 <TrueBrain> I enabled "allow force push" on master
20:55:23 <TrueBrain> the push should be relative quick btw
20:55:28 <TrueBrain> if I understand the GitHub infra a bit
20:55:28 <frosch123> let's see how smart gh is
20:55:38 <TrueBrain> your fork should just be a branch, what people keep telling
20:56:21 <frosch123> TrueBrain: and yes, i had the exact same idea of "let's check this myth" :)
20:56:46 <TrueBrain> they were implementing some protections because of the DMCA stuff
20:56:49 <TrueBrain> so it might have been true :P
20:57:48 <TrueBrain> what still makes me giggle a bit, even I have no clue how I can commit "Verified" commits :P
20:58:11 <TrueBrain> they talk about that nobody looks at it, but it is also so no clear how you would go about adding it yourself .. so why would I check commits for it :P
20:58:33 <frosch123> i think you have to add your pgp key or something
20:58:46 <TrueBrain> but "git show" shows me NOTHING of any signing of any kind
20:58:55 <TrueBrain> so it is GitHub only?
20:58:58 <TrueBrain> or .. I mean .. wuth?
20:59:11 <frosch123> i would think that pgp signing affects the commit hash
20:59:14 <LordAro> it's definitely somewhere
21:00:13 <TrueBrain> well, the fact that you would need a different command shows how bad git is in this
21:00:18 <TrueBrain> you cannot blame GitHub for that :D
21:00:23 <LordAro> you don't need a separate command
21:00:33 <TrueBrain> GitHub publishes their docs via Heroku, it seems
21:00:36 <TrueBrain> I just hit a Heroku error :)
21:00:44 <TrueBrain> LordAro: I want to VERIFY if a commit is signed
21:00:52 <TrueBrain> git show tells me NOTHING of that sorts
21:01:53 <TrueBrain> funny, that they use Heroku :)
21:01:55 <TrueBrain> nothing wrong there
21:02:01 <TrueBrain> but okay, I have a commit that is "Verified" on GitHub
21:02:04 <TrueBrain> locally I do "git show"
21:02:12 <TrueBrain> the comitter is GitHub, as expected
21:02:27 <TrueBrain> git log --show-signature -1
21:02:36 <TrueBrain> gpg: Can't check signature: No public key
21:03:32 <TrueBrain> so yeah, sorry git, if you are not making this easier, don't expect people to do this
21:04:29 <TrueBrain> owh, I need to download the public key from some PGP keyserver first
21:04:48 <TrueBrain> (there is a reason I hate PGP ... I am sure it is very robust, but they made it impossible for the average person to figure out how it works)
21:06:06 <TrueBrain> ah, you can upload your PGP key to GitHub
21:06:23 <TrueBrain> owh, GPG key, sorry
21:06:26 <TrueBrain> (wtf is the difference :P)
21:08:06 <TrueBrain> server is reloading ... :D
21:08:49 <TrueBrain> took ~30 seconds to download
21:08:52 <TrueBrain> now reloading metadata :)
21:09:18 <TrueBrain> (just languages and categories are not there yet)
21:13:06 <TrueBrain> yeah, this takes longer than 5 minutes, so I do need to fix cold-startup .. but not now :D
21:13:14 <TrueBrain> as this was a hot-startup, it is fine :)
21:13:30 <TrueBrain> 2020-11-12 20:13:23 INFO Loading metadata done; took 295.90 seconds
21:13:38 <TrueBrain> there we go, languagebar there
21:14:13 <TrueBrain> see if you find anything wrong/odd/broken
21:15:01 <frosch123> TrueBrain: <small> is apparently a thing
21:15:12 <TrueBrain> I use it in some places even
21:15:28 <LordAro> might be difficult to find quite a lot of pages
21:15:31 <TrueBrain> no, search will be done via search-engines when they indexed the wiki
21:15:44 <TrueBrain> frosch123: ah, tag is not whitelisted :D Nice
21:15:47 <LordAro> what if there are pages that aren't linked from anywhere?
21:16:02 <frosch123> LordAro: at the bottom, click "folder"
21:16:12 <TrueBrain> hmm, it is whitelisted ...
21:16:57 <TrueBrain> I will have to investigate what is going on there :)
21:17:31 <TrueBrain> I think it is because <small> is inside the link
21:18:10 <TrueBrain> everything in a title is escaped
21:18:23 <TrueBrain> which, honestly, is I think the proper thing to do
21:19:35 <andythenorth> TrueBrain sure :)
21:20:17 <LordAro> that external link icon looks terrible
21:20:23 <andythenorth> shall I throw 320 crawler threads at it? :P
21:21:15 <TrueBrain> LordAro: we are open for suggestions; but it is hard to find a good utf-8 icon :)
21:21:20 <TrueBrain> andythenorth: no need to crawl; many dead links
21:21:40 <LordAro> why not use an image?
21:22:01 <TrueBrain> because CSP was blocking my inline image
21:22:06 <TrueBrain> and I don't know how to make it into a proper image
21:22:21 <TrueBrain> "A process in the process pool was terminated abruptly while the future was running or pending"
21:22:28 <TrueBrain> hmm .. that is not good
21:24:54 <TrueBrain> "FileExistsError: [Errno 17] File exists: '/data/.git/index.lock'"
21:25:20 <TrueBrain> so nothing is now being send to GitHub :p
21:25:24 <TrueBrain> so any changes will be void after reload
21:25:48 <TrueBrain> so that might be an issue if it creates a new process
21:26:35 <TrueBrain> someone indexed A LOT of 404s in a very short time :P
21:26:47 <TrueBrain> right, going to provision more memory I guess :)
21:27:07 <TrueBrain> indexing a wiki for 404s is not very useful sadly
21:27:11 <TrueBrain> I can already tell you: 14k+ dead links
21:27:21 <TrueBrain> ah, memory usages is 100 MiB
21:27:23 <TrueBrain> yeah, that will go OOM
21:28:34 <TrueBrain> okay, going to reprovision with a bit more memory
21:28:50 <TrueBrain> I really need to add another node :)
21:29:00 <TrueBrain> I did not see that with the new data-set, memory increased to 100 MiB
21:29:07 <TrueBrain> it needs to run the process at least twice
21:29:11 <TrueBrain> and it had 192 MiB assigned
21:30:34 <TrueBrain> not really sure why a commit also consumes that much memory, as a fork() should be cheap
21:30:56 <TrueBrain> well, possibly a "git commit" just consumes a bit of memory, given the size of the repo
21:31:36 <Koala> I have classes, so I had to go
21:32:01 <TrueBrain> you sounded like the joker ... I'm BBAAAACCCCKKKK
21:32:38 <Koala> SO this is the irc channel for coop right?
21:32:47 <Koala> Or that's a different channel
21:32:51 <TrueBrain> sorry to disappointed :(
21:33:08 <TrueBrain> this feels like: "you are Felicity right?" "No, I am Anna" "Owh, sorry, BYYEEEE"
21:33:32 <Koala> I'm still gonna be on here, just never used an irc channel before
21:33:33 <TrueBrain> this is #openttd, talking about the development of #openttd .. sometimes
21:33:53 <Koala> Development as in game coding?
21:33:57 <TrueBrain> does coop still have an IRC channel? I dunno, honestly ..
21:34:09 <TrueBrain> most of us haven't played in MONTHS if not years :P
21:34:15 <TrueBrain> doesn't mean anything, honestly
21:34:22 <Koala> Well I'm very active lol
21:34:38 <Koala> Still in school, so not playing near as much as I'd like
21:34:51 <TrueBrain> after school you will neither, sad truth of life :P
21:35:14 <Koala> That is a very sad truth indeed
21:35:27 <TrueBrain> server redeployed, memory seems a bit better now
21:35:31 <TrueBrain> (as you can see, I can commit again :P)
21:36:18 <frosch123> TrueBrain: what a useful editing test :)
21:36:33 <TrueBrain> LordAro: I am a bit disappointed that all you could muster so far is to point out one thing you consider "terrible" :P
21:36:59 <TrueBrain> 40 seconds to redeploy with a hot-cache .. that is not bad :)
21:37:29 <andythenorth> coop has a channel, but it's mostly vacant
21:37:53 <TrueBrain> frosch123: what just broke is a bit of a terrible way of things to break .. as it will be nearly invisible to the user .. only I get a sentry report that tells me something is wrong ..
21:38:02 <TrueBrain> but it does prevent users waiting for 4+ seconds when pressing Save
21:38:07 <TrueBrain> not sure what the right thing to do is here
21:38:21 <frosch123> TrueBrain: can you detect it somehow, and put the wiki into readonly mode?
21:38:39 <Koala> Oh, I'll have to check out the coop channel sometime, must go for now!
21:38:41 <TrueBrain> if "pushing" fails, it simply reloads GitHub
21:38:46 <TrueBrain> and your change just disappears
21:38:55 <TrueBrain> but I think your solution is better :)
21:40:09 <frosch123> well, 20 years ago i used to run to the computer room between lessons. but now you probably have a mobile device
21:40:12 <TrueBrain> something to fix tomorrow :)
21:41:20 <TrueBrain> peak memory is now 70%
21:42:49 <TrueBrain> andythenorth: okay, run your crawler again for a bit please? Lets see how it handles stress :P
21:43:21 <TrueBrain> I have seen this wiki too much now, I don't even know where to click to see if things are working/broken :P
21:45:09 <andythenorth> I've pretty much maxed out my connection :P
21:45:18 <andythenorth> 164 threads running currently
21:45:26 <TrueBrain> that explains the latency :P
21:46:03 <andythenorth> yeah it's hurting it a bit
21:46:13 <TrueBrain> okay, tnx andythenorth :)
21:46:14 <TrueBrain> you can stop now :P
21:46:38 <TrueBrain> the difference is insane :D
21:46:42 <andythenorth> surprised my kids weren't complaining
21:46:47 <andythenorth> roblox probably stopped
21:47:12 <TrueBrain> but, by clicking around, found anything wrong LordAro / andythenorth ?
21:47:16 <TrueBrain> like a real show-stopper or anything?
21:47:38 <TrueBrain> memory went from 52% to 57% during your "attack" andythenorth :P
21:47:42 <TrueBrain> so it seems that has very little impact
21:48:42 <andythenorth> I didn't find anything obviously broken
21:48:50 <andythenorth> $somebody ought to do some more CSS
21:48:55 <andythenorth> but the community, right?
21:49:23 <TrueBrain> if you are the community, sure
21:49:45 <TrueBrain> that really does need fixing :D
21:50:49 <TrueBrain> it was a nice idea, to do stuff out-of-process
21:53:22 *** jottyfan has joined #openttd
21:53:24 <TrueBrain> okay, 384MB RAM assigned now
21:53:29 <TrueBrain> that for sure should be enough :P
21:55:31 <TrueBrain> lol ... this has corrupted the cache-file
21:55:36 <TrueBrain> and ... there is no code to recover from that :D
21:59:13 <andythenorth> frosch123 you used to have a newsletter about how bad smooth economy is?
21:59:16 <andythenorth> I have lost my copy
21:59:40 * andythenorth considering a random production change in FIRS, alongside supplies
22:01:06 <FLHerne> Random production changes are incredibly annoying for making pretty networks :-(
22:01:37 <FLHerne> Because it forces you to build huge stations everywhere to absorb the train backlog
22:01:53 <FLHerne> Or do that thing where long trains vanish into 1-tile depots repeatedly
22:04:15 <andythenorth> I use those a lot :)
22:06:40 <glx> <TrueBrain> that for sure should be enough :P <-- yeah like the old 640kB limit ;)
22:09:13 <TrueBrain> "Cache was corrupted; reloading metadata ..."
22:09:22 <TrueBrain> okay, not ideal, but at least it can now recover on its own :)
22:10:59 <TrueBrain> owh, and I guess now it is not going to make it's deadline ... lol ..
22:11:20 <TrueBrain> well, this turned out to be fun :D
22:11:41 <TrueBrain> I am annoyed that this did work on staging
22:11:51 <TrueBrain> but that we fiddled just enough for it to break on pre-production
22:12:35 <TrueBrain> yeah, it failed to make the deadline, so now it is endlessly trying to start :D
22:16:06 *** WormnestAndroid has quit IRC
22:24:32 *** WormnestAndroid has joined #openttd
22:25:20 <TrueBrain> frosch123: well, I downloaded the repo .. 30 times now? :P
22:25:26 <TrueBrain> so we will see if we get an email from GitHub :P
22:27:25 <frosch123> well, it is actually not that big
22:27:57 <TrueBrain> okay, wiki is now running on 384MB RAM
22:28:02 <frosch123> otoh, i am kind of surprised how big the .git folder of openttd is
22:28:06 <TrueBrain> tomorrow I will see if I can reduce the memory requirement
22:28:45 <TrueBrain> but otherwise, this should be able to go live honestly
22:28:54 <TrueBrain> don't want to before we have that read-only protection in place :P
22:29:54 <frosch123> i'll push some templates to staging, so it can be used as sandbox in a meaningful way
22:30:24 <TrueBrain> I did limit staging a bit more on memory btw, as it was using a lot less with this emptiness :D
22:30:38 <TrueBrain> I think tomorrow I first provision bigger ECS nodes
22:30:46 <frosch123> i won't push everything :p
22:31:01 <TrueBrain> we are now over 80% memory reservation
22:31:03 <TrueBrain> I don't like that :)
22:32:14 <TrueBrain> frosch123: what was your plan, to point every language to the english README.md?
22:32:55 <frosch123> that was my plan for english. i did not consider translations
22:33:15 <TrueBrain> well, translators can figure that out :P
22:33:25 <frosch123> but i don't want copies of stuff on the wiki, which are more up-to-date somewhere else
22:35:00 <TrueBrain> "Manual is an open source project, published under the GPL. "
22:35:11 <TrueBrain> guess it is GFDL, not GPL? :)
22:35:24 <TrueBrain> owh, no, that should read OpenTTD
22:36:23 <TrueBrain> FIXED SOMETHING MORE :D
22:36:23 <frosch123> yeah, i noticed... many pages linked [[OpenTTD]], but that actually linked to the manual
22:36:55 <frosch123> so the link was misleading before as well, just noone clicked it
22:37:27 <TrueBrain> okay, everything is a lot more snappy now it has sufficient memory :D
22:37:30 <TrueBrain> shocker, I know right
22:37:34 <frosch123> TrueBrain: i'll make a PR to use the {{forum}} template everywhere
22:37:40 <frosch123> that way everything will be https
22:37:59 <frosch123> editing the wiki with sed is so cool :)
22:38:02 <TrueBrain> I would like some more testers, but it seems nobody really wants to :P
22:38:34 <TrueBrain> after a commit, memory increases by 6% .. that is odd :P
22:38:45 <TrueBrain> (especially as the commit is done out-of-process)
22:38:49 <andythenorth> we'd get better value testing if we edited some content
22:38:59 <TrueBrain> andythenorth: go for it :)
22:38:59 <andythenorth> where is that page that lies about how cdist works?
22:39:12 <andythenorth> oh is the lolz FIRS page still there?
22:39:22 <TrueBrain> right bottom corner is "Folder"
22:39:31 <TrueBrain> I am guessing "Archive" is a good place to start looking
22:40:08 <andythenorth> why does the front page have a list item 'NewGRF and NewGRF'
22:40:12 <TrueBrain> awh, Community .. not Archive :)
22:40:25 <TrueBrain> andythenorth: hover over them .. :)
22:40:27 <frosch123> andythenorth: frontpage needs cleaning up
22:40:28 <TrueBrain> we have stuff to fix ;)
22:40:41 <frosch123> andythenorth: i will do that on staging first though
22:41:57 <frosch123> i liked that other one, that put so much focus on how your hq should look like
22:42:37 <frosch123> oh, i guess you are not in the sc2 news loop. but there was a funny rant about player feedback
22:44:06 <andythenorth> oh that video features me
22:44:29 * andythenorth pretends to look like that
22:49:05 <frosch123> last time you showed them they were using flash
22:52:20 <TrueBrain> I don't really get why this is using so much memory, honestly .. it is using fork()
22:52:29 <TrueBrain> that doesn't duplicate memory, does it?
22:52:51 <TrueBrain> hmm, the question of course is: does Docker count it as a CoW, if it is
22:54:03 <TrueBrain> yeah, CoW is shown as still being used by the child
22:54:08 <TrueBrain> so that is where I went wrong, gotcha
22:54:19 <frosch123> iirc security people usually disable KSM. no idea whether that affects this case
22:55:17 <TrueBrain> pretty sure I can use "spawn" in this case, instead of "fork", for git operations
22:55:25 <TrueBrain> as it doesn't need anything from the parent process
22:55:38 <TrueBrain> for the metadata scan, however, it does need some information
22:57:27 <TrueBrain> memory usage is at 42% now .. so that does fit nicely twice in memory
22:57:31 <TrueBrain> but it is still a bit meh
22:59:18 <TrueBrain> maybe I should approach this a bit different, and let the out-of-process analyze a single file
22:59:22 <TrueBrain> and return what it has found
22:59:26 <TrueBrain> and process that in the main process
22:59:56 <TrueBrain> as updating the dicts is quick
23:00:52 <TrueBrain> would also allow me to use sys.internal()
23:00:57 <TrueBrain> which, as we know, saves a lot of memory :P
23:01:11 <frosch123> TrueBrain: a process for each file? sounds like fork death
23:01:34 * andythenorth has a Horse idea
23:01:44 <frosch123> can you use async to yield after processing each file?
23:01:48 <TrueBrain> I set the worker to a max of 5, or what-ever
23:01:51 <TrueBrain> and it will process the queue
23:02:20 <TrueBrain> well, async would yield the main process every time it starts to analyze a file
23:02:21 <frosch123> if you return to the main thread after each file, i see no point in a separate process
23:02:25 <TrueBrain> which gives plenty of time to handle a request
23:02:43 <TrueBrain> a single file can take 4+ seconds
23:02:45 <TrueBrain> that is the issue :)
23:02:55 <frosch123> maybe i misunderstood what you meant with "single file"
23:02:57 <TrueBrain> 4 seconds is a long time to stall EVERYTHING :)
23:03:03 *** tokai|noir has joined #openttd
23:03:03 *** ChanServ sets mode: +v tokai|noir
23:03:09 <frosch123> 4 seconds? but they render in 30ms?
23:03:17 <TrueBrain> MOST render in that time
23:03:23 <TrueBrain> those who include 100+ templates
23:03:44 <frosch123> so it would also stall when someone views them?
23:03:51 <TrueBrain> it takes the server 288 seconds to do the metadata
23:03:54 <DorpsGek> TrueBrain: 0.0261984899481
23:04:16 <TrueBrain> rendering in the background too you say? :D
23:04:27 <frosch123> then i see no reason to not also stall for indexing
23:04:38 <frosch123> if you want non-stalling viewing, go the render-to-disk route
23:04:52 <TrueBrain> but that does require I can release async for a sec
23:04:55 <TrueBrain> but that should be no issue
23:05:02 <TrueBrain> that means I only have to do git operations out-of-process
23:05:05 <TrueBrain> that simplifies it a bit
23:08:11 <TrueBrain> guess a sleep of 0.1 or something should be enough
23:09:38 <TrueBrain> and for git commits I can just use "spawn" instead of "fork"
23:09:46 <TrueBrain> that consumes only 40MB of RAM
23:09:51 <TrueBrain> instead of what-ever the parent was using :P
23:13:56 <frosch123> someone should improve the css, h2 and h3 almost look the same :)
23:14:10 <TrueBrain> so many more things that really could use ... someone with the eye for CSS :P
23:14:23 <TrueBrain> but in this case, I think I stole that from gollum or mediawiki
23:27:49 <TrueBrain> okay, without doing it out-of-process, it is a bit slugish
23:28:14 <TrueBrain> especially images take a while
23:28:35 <TrueBrain> between 500ms and 800ms response times
23:28:43 <TrueBrain> this happens when someone for example edits Template:- :P
23:29:38 <andythenorth> Christmas Stylesheet Santa
23:30:16 <TrueBrain> so that solution means it is noticeable someone pressed "Save" frosch123
23:30:19 <TrueBrain> otherwise it does work
23:30:49 <TrueBrain> not sure I really like the slowness it adds, but I agree: if that is a real issue, pre-render :)
23:31:25 <TrueBrain> well, and you know sys.internal() is going to help, when a fresh scan consumes 80MB of RAM
23:31:28 <TrueBrain> and a JSON load does 100MB
23:31:29 <frosch123> that change halfes the memory requirement?
23:31:32 <TrueBrain> we have seen that before!! :P
23:32:55 <TrueBrain> and you now slowly see the languages loading :P
23:35:05 <frosch123> feels pretty fast to me
23:35:54 <TrueBrain> I did not commit it yet :P
23:36:18 <frosch123> ah, so i can't see it loading :)
23:37:16 <TrueBrain> didn't we add "sys.internal()" in eints?
23:37:33 <frosch123> eints uses the STR_xyz as keys in dicts
23:41:31 <TrueBrain> 101928 after restart
23:41:37 <TrueBrain> without sys.intern is a restart
23:41:43 <TrueBrain> what-ever "ps aux" reports
23:41:50 <TrueBrain> I think it is KB, might be KiB, dunno
23:47:56 <TrueBrain> what a difference :D
23:55:21 *** WormnestAndroid has quit IRC
23:55:46 *** WormnestAndroid has joined #openttd
23:55:49 <TrueBrain> funny ... the more sys.intern(), the higher memory usage :P
23:55:54 <TrueBrain> but it is now 75044
23:56:23 <frosch123> there must be some downside to it
23:56:33 <frosch123> it only makes sense if strings are actually used multiple times
23:56:56 <TrueBrain> and that is why it is using ~5000KB more memory
23:57:01 <TrueBrain> (over a full rebuild)
23:57:35 <TrueBrain> but we go from 100MB to 75MB
23:57:39 <TrueBrain> that 5MB ... is fine, honestly :)
23:58:55 <TrueBrain> okay, time to make a PR out of this
23:59:02 <TrueBrain> this is a huge memory-improvement
23:59:11 <TrueBrain> at the cost of some increased latency
23:59:38 <TrueBrain> just if people edit Template/en/- a lot, it is noticeable :)
23:59:44 <TrueBrain> otherwise ... pretty sure nobody will notice
continue to next day ⏵
