IRC logs for #openttd on OFTC at 2018-11-05
⏴ go to previous day
02:36:30 *** wodencafe has joined #openttd
03:12:56 *** Gustavo6046 has joined #openttd
08:08:18 *** andythenorth has joined #openttd
08:34:38 *** andythenorth has joined #openttd
09:01:04 <andythenorth> tank game salutes :P
09:13:58 *** Progman has joined #openttd
10:04:37 <peter1138> Hmm, well, this is awkward.
10:12:56 <peter1138> I need to put weight on :p
10:24:00 <andythenorth> you've overshot?
10:40:36 <andythenorth> unless it's a medical thing :P
11:05:32 *** andythenorth has joined #openttd
11:23:47 *** andythenorth has left #openttd
11:30:36 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #3848: Orders: different STR_ORDER_GO_TO for each transport type https://git.io/fxj9a
11:59:47 *** matt21347 has joined #openttd
12:13:55 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #5326: NETWORK_REVISION_LENGTH causing "String too long for destination buffer" git branch with length > 8 https://git.io/fxj7D
13:56:39 *** snail_UES_ has joined #openttd
17:42:33 <LordAro> nielsm: busy day at work, i see :p
18:07:12 <DorpsGek_II> [OpenTTD/OpenTTD] Greg-21 opened pull request #6964: Unificate several files with those from release 1.8 branch https://git.io/fpeEN
18:24:19 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #6963: Several improvements in known-bugs file https://git.io/fpezH
18:39:19 *** frosch123 has joined #openttd
18:54:22 <planetmaker> I *think* that would require then separation of version identification between title bar and network version. It's actually not a new issue and not exclusive to git... 2012. Yet the short SVN version numbers made it much less likely to trigger
18:55:03 <nielsm> two people playing diffrent versions of the same branch would probably not be detected before it just desyncs
18:56:02 <planetmaker> However yes, I like your solution, to always use the HASH for network compatibility
19:00:28 <frosch123> and adding the branch name makes it way easier for people to distinguish official, jgrpp, ...
19:00:57 <nielsm> is this for human consumption or version compatibility check?
19:01:04 <nielsm> I understand it as the latter
19:01:16 <frosch123> i think it is also shown in the multiplayer gui
19:06:04 <nielsm> the date in _openttd_revision is commit date, right?
19:06:40 <frosch123> commit date, so monotone incremental
19:07:28 <nielsm> date + dash eats 9 of 15 bytes, leaving 6 for branch or commit hash
19:08:16 <nielsm> and branch names easily consume 6 characters
19:08:24 <nielsm> so you only get date as differentiator
19:08:53 <nielsm> and not insignificant risk of clash between similarly named branches
19:12:56 <planetmaker> currently, yes, there is no distinction between network version and the one used for human consumption
19:14:07 <planetmaker> maybe we should introduce such distinction
19:14:53 <planetmaker> (or maybe just create a version string of whatever length - and use the last 25 bytes as networ version (instead of first 25))
19:15:21 <planetmaker> that makes sure we have the HASH included at all times
19:16:29 <nielsm> ah good, the network game info packet is versioned
19:16:38 <nielsm> so it can in fact be upgraded without total breakage
19:17:03 <nielsm> (old clients not understanding new version packets is fine since they can't connect anyway)
19:18:17 <nielsm> hmm, server_revision is sent as a string, meaning it must be 0-terminated
19:18:27 <nielsm> meaning it only fits 14 characters in the buffer
19:18:44 <nielsm> but also meaning that it's possible to just extend the buffer in the struct
19:19:22 <Wolf01> Yay, I'm now connected at 3 Mbit
19:19:35 <nielsm> now to download some linux isos?
19:19:40 <frosch123> if you change it, it may be useful to add the branch as separate field
19:19:42 <planetmaker> Not sure you can "just extend the buffer in the struct" when it comes to network packets. They have a maximum size for some reasons. IIRC
19:20:19 <nielsm> well yes the SEND_MTU = 1460 bytes for a UDP packet
19:20:33 <planetmaker> what about just using the last N characters of the existing version string?
19:20:46 <planetmaker> not much change needed for that
19:21:07 <planetmaker> a bit hackish, I grant that :)
19:21:08 <frosch123> planetmaker: don't remove reasons for nielsm to implement multi-package udp announcements
19:21:34 <nielsm> okay 20 bytes consumed per newgrf used
19:21:56 <planetmaker> newGRF is a reason to implement what frosch just said :)
19:21:56 <nielsm> there's stricter version check of newgrfs than the core game engine
19:22:07 <Wolf01> <nielsm> now to download some linux isos? <- I'm downloading destiny 2
19:22:30 <Wolf01> It's 80GB... eta 5 days
19:22:51 <planetmaker> nielsm, yes... because version mismatch between OpenTTD itself hardly was an issue. NewGRF mismatch often was. A BIG PITA. But it got much better now that OpenTTD checks rather rigorously
19:23:22 <nielsm> Wolf01, that would take me about 2 hours :/
19:23:27 <planetmaker> @calc 80000000000/(100000000/8)/3600
19:23:27 <DorpsGek> planetmaker: 1.77777777778
19:27:21 <nielsm> hmm, 198 bytes (max) used by a game info packet with zero newgrfs
19:27:38 <nielsm> @calc (1460 - 198) / 20
19:27:48 <nielsm> so yeah that's 63 newgrfs allowed
19:28:31 <nielsm> drop one to make it 62 newgrfs and the server revision can be extended to 35 bytes
19:28:40 <nielsm> which should definitely be enough
19:29:16 <nielsm> is this number of max newgrfs encoded anywhere in the source?
19:30:18 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #6964: Unificate several files with those from release 1.8 branch https://git.io/fpeoi
19:31:33 <frosch123> static const uint NETWORK_MAX_GRF_COUNT = 62;
19:33:57 <planetmaker> sounds like a good idea. BBL, too
19:52:10 *** andythenorth has joined #openttd
19:55:45 *** yellyFish has joined #openttd
20:06:48 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #5326: NETWORK_REVISION_LENGTH causing "String too long for destination buffer" git branch with length > 8 https://git.io/fpePx
20:09:54 *** Wormnest has joined #openttd
20:14:26 <andythenorth> so after much much thinking
20:14:37 <andythenorth> about vehicle capacities
20:14:40 <andythenorth> the ones I chose
20:14:43 <andythenorth> match the original base set :P
20:15:27 <andythenorth> is that on staging? o_O
20:15:49 <TrueBrain> install it with something like ... 'apt install ruby && gem install bundler && bundle install && jekyll serve'
20:15:57 <TrueBrain> it has no templates nothing
20:16:02 <TrueBrain> it is really as bare as it can be
20:16:17 <andythenorth> I have to read Jekyll CVEs first :(
20:17:01 <frosch123> hmm, you can have dots in repository names...
20:17:35 <andythenorth> policy is not to install anything without checking for obvious vulns first
20:17:46 <andythenorth> except when it's 1 billion upstream deps :P
20:20:57 <TrueBrain> frosch123: surprised me too :D
20:21:11 <TrueBrain> andythenorth: that is what I like about Jekyll .. in production we run nginx + HTML .. fuck any CVE :P
20:21:33 <andythenorth> just don't want my endpoint getting owned :P
20:21:48 <andythenorth> hmm I don't have permission to install to the system ruby
20:21:53 <andythenorth> maybe I can install my own
20:22:35 <TrueBrain> andythenorth: I added you with push rights; feel free to push directly in that branch
20:22:40 <TrueBrain> as often as you like
20:22:42 <andythenorth> I'm currently copy-pasting instructions from the internet
20:22:44 <andythenorth> what could go wrong?
20:22:48 <TrueBrain> before we move it to OpenTTD, I am going to squash it anyway
20:22:52 <andythenorth> at least it doesn't include 'sudo'
20:22:57 <TrueBrain> curl ... | sudo bash, right?
20:22:57 <andythenorth> also noted, thanks
20:23:08 <andythenorth> curl myurlisdefinitelyfine
20:23:26 <andythenorth> oh there's some kind of rubyenv thing
20:23:33 <TrueBrain> I am not first going to try to get this on a staging system of some kinds
20:23:44 <TrueBrain> gem install --user-install .... works fine :)
20:24:57 <andythenorth> fails for me, I'm following jekyll instructions instead
20:28:02 <andythenorth> now we watch ruby build :P
20:28:10 <andythenorth> ok, so I learnt something
20:28:18 <andythenorth> it's not just python that has these problems :P
20:29:19 <andythenorth> obvs. JS has them, but that's not a serious tool
20:30:08 <TrueBrain> is there a pyup for Ruby ..
20:33:33 <andythenorth> jekyll is getting busy with clang now
20:33:44 <andythenorth> "This could take a while..."
20:34:20 <TrueBrain> yeah .... it insist on coming with 'jekyll serve', which requires em_websocket (??) which needs httpparser, which needs a compiler
20:39:26 <TrueBrain> still annoyed we need Jekyll for the small dynamic part .. but still dont know a way around :P
20:39:36 <TrueBrain> we can leave placeholders and do nginx trickery I guess
20:40:25 <LordAro> why ruby/jekyll instead of python/some-python-static-site-generator?
20:40:55 <TrueBrain> what do you have against ruby or jekyll?
20:41:04 <TrueBrain> is any flavor of the month better than the next?
20:41:07 <LordAro> nothing, but existing stuff is python
20:41:15 <TrueBrain> and Jekyll is non-existing?
20:42:49 <TrueBrain> anyway, Jekyll is just the tool; so it shouldnt matter really
20:43:22 <TrueBrain> and github-pages is really populair, so I guess there is something there
20:43:36 <TrueBrain> the framework itself is pretty nice tbh; clean, simpel, to the point
20:43:46 <TrueBrain> that a Ruby app runs it .. meh ..
20:44:29 <TrueBrain> stupid github keeps Ruby populair I guess :D
20:52:10 <frosch123> stupid mediawiki keeps php popular?
20:52:31 <TrueBrain> and script kiddies, dont forget about those! :P
20:54:04 <TrueBrain> still wonder if we should not just sed over the files every 5 minutes to update the banner, instead of running Jekyll every 5 minutes or something :P
20:54:15 <TrueBrain> not sure .. having only a very small part dynamic is annoying :(
20:56:18 <LordAro> wouldn't be the worst thing in the world...
20:57:47 <andythenorth> yeah jekyll in prod, or just have the output committed in the repo and publish those from a path?
20:58:10 <TrueBrain> no, the Jekyll source files will always be in the repo
20:58:15 <frosch123> the template engines have very similar syntax. so it is easy to replace jekyll with other engines later
20:58:24 <TrueBrain> the choices are: run Jekyll build on startup, or once during build and sed the dynamic parts in
20:58:36 <andythenorth> frosch123: that was my thinking, they're all jinja-ish or whatever
20:58:41 <andythenorth> everything is for % {{}}
20:58:41 <TrueBrain> but given GitHub Pages suggest Jekyll .. we will be fine :P
20:59:00 <andythenorth> maybe we can adopt that syntax for nml 2
20:59:13 <andythenorth> oh I had some horrible ideas for nml
20:59:15 <andythenorth> but maybe not now
21:00:31 <andythenorth> I wondered about scoping all the blocks by hacking an ID into the declaration
21:00:37 <Wolf01> Say a number between 1 and 894 (and hope it's installed :P)
21:00:45 <andythenorth> and then using that to see if makes any difference to parser / lexer performance
21:00:57 <andythenorth> by only recompiling what's changed
21:01:05 <andythenorth> the idea makes more sense if I explain it better
21:01:41 <andythenorth> oh jekyll has been busy failing to install while I wait
21:01:48 <TrueBrain> meh, will have to think about how to get this to staging; either way, andythenorth (and others if they like ofc :D) can already start porting the website :P
21:02:55 <Wolf01> I still have like 15 games not redeemed
21:03:28 <andythenorth> ugh yair, jekyll is broken
21:03:31 <andythenorth> I love open source
21:03:36 <Wolf01> Others, just bored of inserting codes in steam
21:03:50 <TrueBrain> you can run the docker, but how it is currently is not meant for development :D
21:03:51 <andythenorth> .rbenv/versions/2.5.1/lib/ruby/gems/2.5.0/gems/bundler-1.17.1/lib/bundler/spec_set.rb:91:in `block in materialize': Could not find concurrent-ruby-1.0.5 in any of the sources (Bundler::GemNotFound)
21:04:06 <andythenorth> I could play the random install crap game
21:04:19 <andythenorth> this is all reassuringly pythonic :P
21:04:25 <TrueBrain> I have to install json and bignumbers before it worked :P
21:04:30 <TrueBrain> but concurrent-ruby I havent seen
21:04:37 <andythenorth> neither has the gem installer
21:04:40 <andythenorth> it can't find any version
21:04:53 <TrueBrain> what you can do, is forget about 'bundle install'
21:05:00 <TrueBrain> and just run: gem install jekyll
21:05:08 <andythenorth> yeah I did that already :)
21:05:14 <TrueBrain> possibly the Gemfile.lock is Linux only or something
21:05:25 <TrueBrain> if you have jekyll installed, just run: jekyll serve
21:05:28 <TrueBrain> and forget about the rest :P
21:05:42 <andythenorth> yeah that's how I get the concurrent error :)
21:06:16 <andythenorth> if it's like python, I probably need a different bundler version
21:06:29 <TrueBrain> lolz; that mostly is the case for Python yes :D
21:06:30 <andythenorth> I'm assuming bundler is setuptools, with all the related 'it doesn't really work' shit
21:06:44 <andythenorth> gem is presumably pip
21:06:47 <TrueBrain> I run Ruby 2.5.0 and Bundler 1.17.1
21:07:11 <andythenorth> I have ruby 2.5.1 and bundler 1.17.1
21:07:16 <TrueBrain> so that should be fine, I would guess
21:08:49 <TrueBrain> gem install concurrent-ruby doesnt work?
21:08:56 <TrueBrain> internet suggests that should "just work"
21:09:20 <andythenorth> works fine, gives me 1.1.1
21:09:26 <andythenorth> but this recipe wants 1.0.5
21:09:55 <TrueBrain> 1.0.5 is VERY old :P
21:10:40 <TrueBrain> "gem install concurrent-ruby -v 1.0.5" fails, I assume?
21:13:04 <andythenorth> gem install jekyll -v 3.8.4
21:13:30 <andythenorth> gem install jekyll-feed -v 0.11.0
21:13:40 <TrueBrain> dependencies should take care of this :P
21:13:43 <TrueBrain> try: bundle install, btw
21:13:43 <andythenorth> I started looking for a manifest to edit
21:13:54 <TrueBrain> Gemfile.lock contains all the dependencies
21:14:55 <andythenorth> ok bundle install worked
21:15:05 <andythenorth> I think I should bin this rbenv and start from scratch tho
21:16:23 <andythenorth> that's the nice thing about these packaged envs
21:16:28 <andythenorth> instead of blitzing the system install
21:17:01 <andythenorth> so many years of pain due to messing with system python etc
21:18:24 <andythenorth> ok back in 15 mins or so, after ruby finishes running my fans
21:25:26 <andythenorth> TrueBrain: it works again, but still wasn't a clean install :P
21:26:44 <TrueBrain> as long as you can get to work
21:26:47 <TrueBrain> I am fine with it :)
21:26:58 <andythenorth> I like a repro :)
21:27:07 <andythenorth> too many years of 'it works, never touch it again'
21:27:34 <andythenorth> gem install jekyll gives 3.8.5 or so
21:27:41 <andythenorth> bundle install gives 3.8.4 or so
21:27:44 <TrueBrain> most likely if we build the website once, we never run Jekyll ourself again :P
21:29:31 <andythenorth> can anyone explain physics to me?
21:30:22 <andythenorth> oh he can't get in the channel anymore
21:30:27 <frosch123> gravity? did you drop your notebook?
21:31:05 <andythenorth> you got a bridge?
21:31:26 <andythenorth> frosch123: weight of object vs internal volume
21:31:40 <andythenorth> claim is that narrow gauge wagons have less dead weight per unit cargo capacity
21:31:44 <andythenorth> I don't follow that
21:31:46 <Wolf01> If it's full it weights more
21:32:02 <TrueBrain> Wolf01: unless they are anti-gravity particles, of course
21:32:27 <TrueBrain> or anything lighter than air
21:32:38 <TrueBrain> YOUR LOGIC IS FLAWED
21:32:41 <andythenorth> volume increases as a cube?
21:32:57 <andythenorth> weight of sides, floor, roof, less so?
21:33:15 <TrueBrain> volume can also increase like a cylinder
21:33:23 <TrueBrain> depends on the form, really
21:33:46 <andythenorth> what about a gem?
21:33:59 <TrueBrain> right, it seems I am not going to be useful anymore this evening
21:34:04 <andythenorth> gem install bundler first?
21:34:12 <TrueBrain> gem install bundler && bundle install
21:34:16 <TrueBrain> should be sufficient
21:34:23 <TrueBrain> (that is what the Dockerfile does anyway :P)
21:37:25 <andythenorth> -bash: .rbenv/shims/jekyll: No such file or directory
21:37:35 <andythenorth> after bundle install
21:37:37 <LordAro> andythenorth: stop breaking everything
21:37:42 <andythenorth> but if I gem install jekyll
21:37:47 <andythenorth> then uninstall it
21:37:55 <TrueBrain> so if that works for you, use it
21:38:00 <TrueBrain> dont overthink this part too much
21:38:01 <andythenorth> so presumably it's just missing a shim or path or something
21:38:03 <TrueBrain> you are special on your OSX :P
21:38:26 <TrueBrain> dont waste too much of your brain cycles on getting this perfect :)
21:38:34 <andythenorth> I am just using plain ruby here though, this is not Apple ruby
21:39:01 <nielsm> orudge yes :) I have a working (I think) patch for fluidsynth music driver on my linux laptop that changes it to output via the mixer too
21:39:46 <andythenorth> followed by gem install -v 3.8.4
21:39:58 <nielsm> and I think being able to use fluidsynth on any supported platform that way would be awesome
21:40:06 <orudge> Yes, that would be handy
21:40:23 <nielsm> also want to add in the opl2/3 synth from dosbox and use it for the adlib.cat music
21:41:06 <nielsm> (just need to reverse the adlib.com driver for that... ugh)
21:41:58 <frosch123> andythenorth: smaller things are more firm in general i guess
21:42:09 <TrueBrain> that's what she said?
21:42:09 <frosch123> like ants can carry stuff
21:44:06 <TrueBrain> right, I am off. andythenorth, I expect results tomorrow :P
21:44:07 <frosch123> if the wagon weights less, you need less wagon to support itself
21:47:30 <andythenorth> "the advantages claimed for narrow gauge proved untrue"
21:47:34 <andythenorth> I read my sources a bit more :P
22:55:50 *** ToffeeYogurtPots has joined #openttd
22:59:27 *** andythenorth has left #openttd
23:31:08 <DorpsGek_II> [OpenTTD/OpenTTD] Greg-21 updated pull request #6964: Unificate several files with those from release 1.8 branch https://git.io/fpeEN
23:49:19 <DorpsGek_II> [OpenTTD/OpenTTD] Greg-21 commented on pull request #6964: Unificate several files with those from release 1.8 branch https://git.io/fpvUG
23:52:09 *** TheMaster has joined #openttd
23:55:33 <DorpsGek_II> [OpenTTD/OpenTTD] Greg-21 commented on pull request #6963: Several improvements in known-bugs file https://git.io/fpvUP
23:56:26 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #6964: Unificate several files with those from release 1.8 branch https://git.io/fpvU9
23:57:19 *** TheMaster is now known as Unit193
continue to next day ⏵
