IRC logs for #openttd on OFTC at 2018-07-19
⏴ go to previous day
00:01:21 *** sim-al2 is now known as Guest1420
00:01:22 *** sim-al2 has joined #openttd
00:45:34 *** andythenorth has joined #openttd
00:57:17 *** Suprcheese has joined #openttd
00:58:32 *** CompuDesktop has joined #openttd
00:59:05 *** [1]sim-al2 has joined #openttd
01:01:50 *** [1]sim-al2 is now known as sim-al2
01:01:50 *** Suprcheese is now known as Supercheese
01:02:50 *** rocky1138 has joined #openttd
01:04:52 *** Tharbakim has joined #openttd
01:06:49 *** ccfreak2k has joined #openttd
03:36:06 *** snail_UES_ has joined #openttd
03:44:40 *** Supercheese has joined #openttd
03:53:23 *** Supercheese has joined #openttd
04:32:30 *** Laedek_ has joined #openttd
05:17:27 *** sla_ro|master has joined #openttd
05:43:49 *** ToBeFree has joined #openttd
08:23:29 *** Supercheese has joined #openttd
11:28:08 <TrueBrain> meh; bit stuck ... when I install Debian Stretch on OpenTTD, it simple doesn't want to boot correctly ..
11:32:38 <peter1138> Debian... on OpenTTD...?
11:36:37 <TrueBrain> on the OpenTTD infrastructure
11:38:11 <TrueBrain> tomorrow a call with Google Cloud Platform ... lets hope they can offer something
11:38:43 <TrueBrain> as I simply don't have time anymore to do it myself, all this infrastructure blabla ..
11:41:07 <SpComb> it's fun to write a custom CI/CD system, though? :P
11:41:25 <TrueBrain> meh; I tried really hard to avoid that too
11:41:32 <TrueBrain> but given the constraints ...
11:41:40 <SpComb> what does GCP offer... GCE or even GKE?
11:41:48 <TrueBrain> I asked them what they could offer
11:42:11 <SpComb> you'd expect that having kube would make e.g. CI/CD easier to implement
11:42:14 <TrueBrain> as that would mean a lot more people can help out, without understanding most part of it
11:42:30 <TrueBrain> the CI/CD is done up till the part it needs to deploy something to somewhere
11:42:36 <TrueBrain> and yeah, kubernetes would make that near-trivial
11:42:41 <TrueBrain> (as the configuration is already in there)
11:43:04 <TrueBrain> I was looking if I could do it in docker, which means I have to write my own configuration layer ... which really sucks balls
11:43:20 <TrueBrain> setting up your own kubernetes is .. euh ... yeah, lets not go there
11:49:13 <SpComb> my day job is automating kube setup and maintenance :)
11:50:11 <TrueBrain> so set it up for OpenTTD already :P
11:50:22 <SpComb> no, we offer the tooling that lets you do it yourself :P
11:50:24 <TrueBrain> I am really annoyed CoreOS is no longer really "stable" .. bit uncertain in which state it is ..
11:50:43 <TrueBrain> I looking into kubespray ... I slowly walked away
11:50:57 <SpComb> pharos would be somewhat similar
11:51:00 <SpComb> what was wrong with kubespray?
11:51:12 <SpComb> but yeah, if you can get free GKE resources from google, then it's certainly a good option
11:51:17 <TrueBrain> the complexity .. and the amount of required maintainance
11:51:33 <TrueBrain> mainly, we would need to set it up on-premise
11:51:35 <TrueBrain> we have a single machine
11:51:47 <TrueBrain> all stuff kinda demands you have at least 3 nodes in kubernetes
11:51:56 <TrueBrain> so ... I have to spin up 3 VMs? And another for ansible?
11:52:02 <TrueBrain> I dont have that memory available
11:52:54 <SpComb> you can have a single master host without the NoSchedule taint
11:52:59 <TrueBrain> yeah, I saw you can do that too with kubespray, but than people claim that is not "production ready"
11:53:07 <TrueBrain> and I got just really annoyed of me not having enough domain knowledge :D
11:53:32 <SpComb> dunno why it wouldn't be production-ready assuming you have enough resources to cover the kube overhead, and don't consider the lack of HA to be a problem
11:53:37 <TrueBrain> I guess it is mainly that ... I know shit about this stuff .. and it is a pretty steep learning curve
11:53:49 <TrueBrain> lack of HA is never been an issue :D
11:53:59 <TrueBrain> but okay .. so now I run into the next issue .. that I would need to upgrade the dom0
11:54:14 <SpComb> physical host running some ancient xen?
11:54:20 <TrueBrain> which means renting a second machine, linking them together (with a GRE and some black magic on network level), migrating to one machine, upgrading, migrating back ..
11:54:46 <TrueBrain> it runs XenServer .. upgrading a single-host XenServer is a real pita
11:54:51 <TrueBrain> I regret ever making that choice
11:55:42 <SpComb> I've used ganeti for kvm with drbd and live-migration, but it lacks support for a lot of kvm knobs and it's not really maintained anymore, which is a pity
11:55:47 <TrueBrain> and when you look what it costs to get a kube in cloud and kube on permise ...... cloud is so expensive
11:56:28 <SpComb> does the old version of xen prevent you from running newer domU kernels/distros?
11:56:31 <TrueBrain> so yeah .. what I am trying to create is an infrastructure that is as simple as possible, so more people can support it
11:56:40 <TrueBrain> yeah, latest kernel just flat-out crashes
11:56:53 <TrueBrain> latest Debian Stretch kernel flat-out crashes
11:57:05 <SpComb> is it still some xen-PV mode without any bootloader support?
11:57:22 <TrueBrain> its PV, with pygrub bootloader
11:57:43 <TrueBrain> I can put it in HVM .. but .. I care about my performance
11:58:14 <TrueBrain> the migration is the issue, to any other technoligy
11:58:26 <TrueBrain> so that is why I have the plan to migrate everything inside Dockers
11:58:28 <TrueBrain> which is a lot easier
11:58:32 <SpComb> but yeah, maintaining a single physical machine like that is a challenge once you get to the point where you have to start doing a major upgrade on it... tends to turn into a full rebuild
11:58:34 <TrueBrain> and run Dockers "somewhere"
11:58:55 <TrueBrain> I upgraded the machine twice now .. both times took a month, including planning and execution
11:59:09 <TrueBrain> OVH is a good provider to have in these cases
11:59:40 <TrueBrain> maybe we should rent a second machine, make that a kube-cluster, and move everything in dockers to there or something
11:59:55 <TrueBrain> best thing about Docker, we can just migrate even the shitty-old-stuff we have
12:00:11 <TrueBrain> (some parts still run in Django 1.2)
12:00:17 <SpComb> btw if a single machine is enough, then `kubeadm init` is pretty trivial, all you really need is an appropriate kubeadm yaml
12:00:39 <TrueBrain> people kept telling me I should never do that?
12:00:51 <TrueBrain> (again, I know very little about kube, so I just go with what people tell me)
12:01:06 <TrueBrain> something about hard to maintain, no upgrade path, etc
12:01:21 <SpComb> `kubeadm upgrade` has worked fine so far for 1.9 -> 1.10 -> 1.11
12:01:58 <SpComb> given, the systemd units changed in 1.11, and there were some things like `kubelet --cluster-dns` that you had to tweakin the 1.10 systemd units, and those tweaks no longer work in 1.11... but they're no longer necessary either
12:02:37 <TrueBrain> hmm .. so I guess I could give that a try ..
12:02:49 <SpComb> okay, and you also need to do a two-step dance where you first download the raw kubeadm binary and use that to run the upgrade before upgrading the kubeadm package, or it breaks
12:03:30 <TrueBrain> so ideal I have the kube cluster hosted externally, as that would mean no support from our side ..
12:03:38 <TrueBrain> but for now I can spin one up and see if that works
12:04:00 <TrueBrain> just installing a new debian stretch fails horribly .. hmm ... lets clone an existing :D
12:04:47 <TrueBrain> btw, random question, do you happen to know: can you in a kube cluster have persistant storage?
12:04:55 <TrueBrain> for example, what if I want to run mediawiki in it?
12:05:05 <TrueBrain> (uploads etc ar enot in a database, but on a filesystem)
12:05:38 <SpComb> it took me a couple years, but I finally managed to upgrade our puppet infra from 3.x -> 4.x/5.x this month... means we can finally start migrating from squeeze/wheezy/jessie -> stretch now, because puppet 3.x is no longer available/supported :)
12:06:09 <TrueBrain> ugh, puppet .... I started to really hate puppet :D
12:06:15 <SpComb> there's various approaches to it... trivial approach is to bind the pod to a specific node and use a hostPath volume, of course
12:07:43 <SpComb> after that it starts getting more complicated with persistentVolumes... you can use local storage if you create local persistentVolumes and bind them to specific nodes, but the approches to that which I've seen barely make any sense, it's just more complex than the hostPath approach and I'm not aware of any automagic sync/replication functionality
12:08:27 <SpComb> and then there's all kinds of custom storage providers like OpenEBS, buuuuut I'm not exactly convinced that they would be production-ready yet either
12:08:39 <SpComb> it's a solved problem if you have some external storage like EBS, though
12:08:40 <TrueBrain> do cloud providers solve this?
12:08:54 <SpComb> yeah, then you just use the integrated kube support for those cloud provider volumes
12:09:16 <TrueBrain> more reason to use that :D
12:10:32 <TrueBrain> so you say, just download kubeadm, and run kubeadm init :)
12:10:54 <SpComb> sure, it's the easiest way to get started, and it *does* have an upgrade path
12:11:09 <SpComb> main limitation with kubeadm is that it doesn't really do multi-master/HA at all, at least not yet
12:11:25 <TrueBrain> when we run kube on premise, that is also never going to happen
12:11:59 <TrueBrain> first, let me see what happens if I upgrade this clone .. see if it does want to boot ..
12:13:33 <TrueBrain> the page on kubernetes.io that tells how to install kubeadm is a 404 :D
12:15:15 <TrueBrain> owh right .. and than you get to "picking a network plugin" .. yeah .... eeuuuuhuhhhhh ... *closes eyes*
12:15:21 <TrueBrain> you happen to have a suggestion?
12:17:35 <TrueBrain> okay, 4.9.0-r6 works, 4.9.0-r7 crashes
12:17:43 <SpComb> but we only support Ubuntu... not Debian
12:17:58 <TrueBrain> is there really a difference, in terms of support?
12:19:52 <TrueBrain> holy crap at these prices .. people pay that? Big business .... :D
12:20:12 *** ChanServ sets mode: +v tokai
12:20:39 <TrueBrain> kubernetes 1.10? What about 1.11!
12:21:22 <SpComb> answer: it's july, everybody except me is on holidays :P
12:23:00 <TrueBrain> okay, now first upgrading while holding the kernel back ..
12:23:57 <TrueBrain> SpComb: why do I read a lot that they want kubernetes to run on Docker 17.03? 18.N has been out for months now?
12:25:02 <SpComb> partly political, kube does not use any of the newer Docker features past 1.11 or so, so the newer Docker releases really only bring more bugs
12:25:37 <SpComb> long-term plan is to replace Docker with some other CRI like containerd or cri-o, but it hasn't happened yet
12:26:37 <SpComb> I mean, it's an option, but not the default
12:28:13 <TrueBrain> can a host be: this current host?
12:30:28 <TrueBrain> seems it always goes over ssh
12:31:47 <TrueBrain> awh, it really refuses the host because it is not ubuntu :D
12:32:02 <TrueBrain> Ubuntu is not a server distribution ffs :P
12:32:26 <SpComb> I'd guess that Ubuntu server is more commol than Debian, tbh
12:32:56 <TrueBrain> it still makes me puke everytime I notice that
12:33:02 <TrueBrain> somewhere something went terribly wrong :D
12:34:48 <SpComb> I'm pretty sure that canonical makes more money from Ubuntu server/cloud than desktop
12:35:54 <TrueBrain> hmm .. kubeadm is not in the google repo, what documentation suggests
12:37:37 <TrueBrain> they too, only added it for Ubuntu
12:38:00 <SpComb> you can probably install the ubuntu dist on debian just fine
12:38:15 <SpComb> the kube packages only contain a sinhle static binary + systemd units
12:39:50 <TrueBrain> so many preflight errors :D
12:39:56 <TrueBrain> it even refuses newer Dockers
12:43:56 <TrueBrain> I love the: swap is not allowed, part
12:46:02 <TrueBrain> well, that seems to run..ish
12:52:56 <TrueBrain> tnx SpComb, kubeadm seems to work .. now I need to figure out how it all works, but I had to do that anyway :D
12:54:16 <SpComb> yeah, and if nothing else a single-node kubeadm setup is a good step towards something like GKE
12:54:37 <TrueBrain> this is a bit weird ... I told it to use 10.100.0.0/16 .. but the sock-shop now has 10.104.N.N
12:54:52 <SpComb> I wouldn't be surprised if GCP were willing to offer the OpenTTD project free GKE
12:55:15 <SpComb> are you sure it osn't 10.100.0.0/12?
12:56:24 <SpComb> the service cidr is also separate
12:56:45 <TrueBrain> I wrote /16 .. but what-ever ... as long as it stays away from the lower-end of that range, we will be fine
12:57:02 <TrueBrain> OpenTTD infra runs very isolated in terms of networks .. so it cannot do much harm
12:57:19 <TrueBrain> and I hope GCP can do something for OpenTTD .. not having to maintain this ... pfftt .. that would be very nice :)
12:57:29 <TrueBrain> and having everything run in pods ..... even nicer
12:57:57 <SpComb> the default service cidr is 10.96.0.0/12 and that covers 10.104.x.x
12:58:13 <TrueBrain> funny .. I really use another address :P Owh well, that range is fine
12:58:22 <SpComb> odd if kubeadm allows the two to overlap
12:58:49 <TrueBrain> and I can even reach it from our web-proxy :P
12:58:56 <SpComb> having the pod cidr within the service cidr sounds like a bad idea
12:59:23 <TrueBrain> I just loaded the sock-shop demo
12:59:29 <TrueBrain> it gave me a service on a port
13:00:01 <TrueBrain> so now all I have to find out is how to start my test image
13:00:13 <TrueBrain> and a deployment is than as simple as settng a new image for it
13:00:27 <TrueBrain> images are like: truebrain/testing:sha-of-commit
13:01:06 <TrueBrain> tnx a bunch SpComb :)
13:01:36 <SpComb> using nodeport services with an externally configured frontend proxy is probably the common initial setup, but there's also native kube ingress-nginx and letsencrypt stuff
13:01:56 <SpComb> np... where do I have sales send the consulting bill to? :P
14:56:11 *** snail_UES_ has joined #openttd
15:07:51 <Erato> I had a question about BaNaNaS and was told that the people here could help.
15:09:34 <Erato> I would like to change the name of one of my newgrfs, because I incorrectly entered it and it seems like I can't change it without reuploading under a different GRF-ID.
15:22:33 *** Wacko1976 has joined #openttd
15:51:59 <TrueBrain> renaming is a bit of a pita indeed :D
15:55:02 <TrueBrain> Erato: what is the grfid?
15:55:50 <TrueBrain> it already has 4 versions?
15:56:08 <TrueBrain> that makes it even harder :D
15:56:11 <TrueBrain> what would you like changed?
15:56:21 <Erato> "RIMS" should be "Real International Maglev Set"
15:56:59 *** Mahjong has joined #openttd
15:57:24 <TrueBrain> the main issue with renames is that it takes a lot of time and effort
15:57:34 <TrueBrain> the name is part of the filename as it ends up on the mirrors
15:57:53 <TrueBrain> it is also how the files end up at the user on his disk
15:58:23 <TrueBrain> so I am wondering what the easiest approach is .. as renaming 4 versions will take way too much time to make it reasonable
15:58:37 <planetmaker> dumdidei... it would at worst mean users download it again... but grfid is checked before, IIRC?
15:58:39 <TrueBrain> Erato: do you have a new version ready to upload?
15:58:42 <Erato> I'm going to be updating it relatively soon
15:58:54 <Erato> It's not ready yet, unfortunately
15:59:11 <planetmaker> couldn't you make an intermediate update to upload today?
15:59:44 <TrueBrain> planetmaker: I am not sure; last time some custom stuff got a bit confused :D
15:59:50 <TrueBrain> but mostly, it is expensive mirror-wise
16:00:00 <TrueBrain> I have to manually login on them and remove the unwanted files, before they pick up the new name
16:00:11 <TrueBrain> which .. takes a lot of time
16:00:18 <TrueBrain> last we did a rename, we did it in bulk
16:00:23 <TrueBrain> making it somewhat more bareble :)
16:01:32 <TrueBrain> but I cannot remember if we ever did that, and what the after-effects were
16:01:41 <TrueBrain> Toying with a 8 year old system is never fun :D
16:04:02 <TrueBrain> if it breaks, it breaks
16:04:15 <TrueBrain> Erato: if you now upload a new version, it should be named as you asked
16:04:20 <TrueBrain> old files remain RIMS
16:04:34 <TrueBrain> please let me know if you uploaded 1.4 (even if it is a dummy)
16:04:40 <TrueBrain> than I can make sure everything went as it should :D
16:04:56 <TrueBrain> a 1.4 as soon as possible would be appreciated :)
16:07:08 <Erato> I think I can get something acceptable out the door in 5-10 minutes
16:07:21 <TrueBrain> can even be the same content as 1.3 btw, just named 1.4 :)
16:07:29 <TrueBrain> (as long as the md5 isnt identical)
16:13:53 <Erato> I saw in the Bananas manager that the name of 1.3 had already been changed
16:14:26 <Erato> Alright it shows up as "Real International Maglev Set" in the in-game content downloader.
16:14:41 <Erato> TrueBrain: It seems to work.
16:17:03 <TrueBrain> that was the trick I tried; I renamed 1.3 (the name, not the filename), which is not okay, but fakes the system in thinking on next upload what the name should be :D
16:17:10 <TrueBrain> so now you uploaded 1.4, I renamed 1.3 back to RIMS :P
16:17:17 <TrueBrain> and it seems the system accepted that ;)
16:18:17 <TrueBrain> its on the list of BaNaNaS2 that you as user can just rename it btw :)
16:18:49 <LordAro> inb4 OTTD makes some assumption about names are GRFIDs and now crashes
16:19:16 <TrueBrain> the whole system uses GRFids and MD5s to find GRFs :)
16:19:30 <TrueBrain> names are just what end up on mirrors, and only used when you initiate a http download .. or so it should :P
16:24:21 *** andythenorth has joined #openttd
16:25:23 *** Alberth has joined #openttd
16:25:23 *** ChanServ sets mode: +o Alberth
16:33:26 <andythenorth> glass + plastic -> window factory -> building materials
16:34:29 <Alberth> no metal in windows? :)
16:34:54 <Alberth> or wood, for that matter
16:35:27 <nielsm> modern windows are all about the plastic
16:35:37 <andythenorth> one of the weaknesses in FIRS is that too many cargos could be used ~everywhere
16:36:09 <Eddi|zuHause> gives a challenge for cargodist
16:36:18 <andythenorth> also I worked in a upvc window factory
16:36:20 <Alberth> so you were looking for a use of plastic :)
16:37:27 <nielsm> make an entirely opaque economy where the cargoes are just named A, B, C and industries are "type 1, 2, 3"
16:38:02 <nielsm> and be released from the constraints of having to seems based on a reality
16:38:10 <andythenorth> that's the 'stuff' economy
16:38:29 <andythenorth> reality is quite easy to circumvent
16:38:52 <andythenorth> apply the toy train 'prototype for everything' rule
16:47:11 <Alberth> nielsm: and then you get the problem of assigning cargoes and industry types, not sure how much better that gets
16:51:02 <nielsm> Alberth but you can do that in a completely arbitrary manner
16:52:18 <Alberth> yes, the problem is too much choice indeed
16:53:40 <Alberth> in particular too many boring options
17:03:42 <andythenorth> shall I approve your PRs? :P
17:03:46 *** Wormnest has joined #openttd
17:14:58 <andythenorth> I don't think I'm allowed :P
17:15:05 <andythenorth> we're just in a weird situation now
17:15:12 <andythenorth> where peter1138 used to be allowed to just commit
17:15:33 <andythenorth> we're getting non-core commits through faster (win)
17:15:40 <andythenorth> but core developers are blocked (ooof)
17:31:19 <SpComb> blocked? You mean they have to go through the same PR review as any other plebs?
17:31:46 <andythenorth> previously, often, they kind of didn't
17:36:35 <Alberth> not really, anything non-trivial was reviewed too
17:37:07 <Alberth> ie technically being able to bypass doesn't mean it's a good idea to do so :P
17:38:57 <Alberth> and in my experience, a review does improve a solution, it's beneficial too thus
17:39:22 <andythenorth> I am more laissez faire :P
17:39:47 <andythenorth> experience people usually want to avoid being the one wearing clown shoes
17:40:19 <Alberth> it's also a form of learning
17:40:52 <Alberth> you get new insights in how to solve things beyond what you know or considered
17:57:01 *** HerzogDeXtEr has joined #openttd
17:58:39 <SpComb> it's also possible to do really bad reviews and get stuck on some inconsequential opinionated aspects
17:58:46 <SpComb> reviewing it also a skill that you need to learn
18:02:40 <andythenorth> we are figuring out reviews at work
18:02:56 <andythenorth> reviews seem like they should be about risk
18:03:22 <andythenorth> taste is *really* important when coding
18:03:34 <andythenorth> but if I'm coding it, and I've complied with all/any rules
18:03:44 <andythenorth> then *your* taste shouldn't creep into the review
18:04:01 <andythenorth> just wastes everyone's time
18:04:23 <Alberth> work is simple, as the boss decides the direction
18:04:27 <andythenorth> I should have used Bob and Alice in my example :P
18:04:38 <andythenorth> Alberth: in my case, why would I decide the direction?
18:04:42 <andythenorth> I pay other people to do that
18:04:50 <andythenorth> my job is mostly spreadsheets and signing stuff
18:05:03 <Alberth> open source has the additional problem of holder of the direction of development
18:05:18 <andythenorth> Guido is having an interesting moment with that
18:05:23 <Alberth> in that case, the people that you pay decide the direction :)
18:05:42 <andythenorth> so we look at risks
18:05:47 <nielsm> IMO reviews should also look at the context of the changes and consider if all relevant documentation/comments are being updated to match the new state
18:06:05 <andythenorth> - infosec risk (confidentiality, availability, integrity)
18:06:16 <andythenorth> - and stasis risk: does this tie our hands in future
18:06:17 <Alberth> suppose I submit a patch to remove some in my view obsolete part
18:06:37 <nielsm> I had a bit of an argument on a review a while ago on a pull request here, about whether/how much comments should be changed
18:06:47 <Alberth> even if I follow all rules, there has to be agreement that the part is indeed obsolete
18:07:54 <andythenorth> I'm not disputing that :)
18:08:00 <andythenorth> the map array PR is a good case for that
18:08:10 <andythenorth> it's a fundamental change, even if it's not dramatic
18:09:23 <Alberth> map array is not a real change, it's just an expansion of possibilities, we can disagree on how much cpu time it may cost, but that's about it
18:09:41 <Alberth> shunting would be bigger in my view
18:10:24 <andythenorth> consent and consensus are always interesting
18:10:39 <Alberth> but people just submit patches, without making sure it's acceptable, so that comes with the review
18:10:45 <andythenorth> I tend to apply consensus as 'people have stopped disagreeing with me'
18:10:51 <Alberth> while work has already set that direction
18:11:30 <andythenorth> "Consent means that one has an opportunity to present a “reasoned no.” It does not require “a yes,” an affirmation, nor does it imply unanimity or full agreement. It is also not a veto; it must be explained."
18:12:25 <Alberth> oh, there are many level of agreement I agree :p
18:12:58 <andythenorth> the upside of my projects is instant 100% consent :P
18:13:09 <andythenorth> the downside is, I have to do nearly everything :P
18:13:36 <Alberth> but at some point it's about direction, and if some guy doesn't want, it's not going to happen with that guy
18:14:48 *** Progman has joined #openttd
18:17:52 <Alberth> but we may be deviating from not getting code reviewed :p
18:18:19 <andythenorth> it's all related :P
18:18:39 <Alberth> I am sure FIRS is the center of it :p
18:20:27 <SpComb> how common is it to review your own code with `git diff` before each `git commit`?
18:20:51 <Alberth> 99.5% of the time for me
18:21:33 <Alberth> but that's just technical correctness, is it all there, nothing weird?
18:22:02 <SpComb> I wouldn't necessarily assume that to always be the case
18:22:17 <Alberth> not "he, if you take this and this approach, it's much simpler, and you can do that and that too"
18:23:11 <Alberth> nope, I usually find 1 or 2 things every 3 commits or so
18:23:28 <Alberth> just an additional empty line usually
18:24:35 <Alberth> on the other hand, when I revisit code, I do sometimes think "oh man this is all wrong" :p
18:24:56 <Alberth> broken in some obscure way :)
18:27:37 <Eddi|zuHause> the "which idiot coded THAT?!?" problem :p
18:27:52 <peter1138> With reviews, you now have "which idiot approved THAT?!?"
18:28:15 <Eddi|zuHause> i meant, the "idiot" is you from 6 months/5 years/whatever ago
18:28:47 <andythenorth> 'former us' were stupid
18:31:51 <Alberth> everything is broken, for most parts we're still working out how :)
18:32:06 <peter1138> So maybe I should make maintain a branch with everything merged in my forked.
18:32:33 <peter1138> But maybe I won't bother, because that way lies madness.
18:32:58 <peter1138> As it is, I've effectively given up coding anything because of the review system.
18:33:17 <peter1138> Great principle, it's just not working in practice.
18:33:45 <peter1138> Map array changes. No comment. Nothing. Not even a "holy shit you shouldn't do this"
18:33:46 <SpComb> the worst part of review systems is when you have open PRs that never get reviews, either negative or positive
18:34:21 <Alberth> number of people that can give meaning full comments on that is close to 0
18:34:24 <peter1138> GUI timing changes. Reviewed briefly, changed made, not looked at again. Stagnant.
18:34:27 <SpComb> because reviewing also takes effort, maybe more effort than writing the code itself sometimes
18:35:44 <peter1138> NRT, I may just merge this later, becuase it's had zero objections.
18:35:55 <Eddi|zuHause> understanding other people's code is a pain
18:36:17 <peter1138> However, I wouldn't be able to fix any issues myself, because it'd just get stuck.
18:36:25 <andythenorth> NRT, we already had TB review
18:36:36 <andythenorth> all the issues TB found were fixed afaik
18:36:36 <SpComb> well, the motivation for reviewing has to also come from somewhere
18:36:40 <andythenorth> then you reviewed it peter1138
18:37:19 <andythenorth> I am +1 to reviews, but I like to know what risk they're preventing them
18:37:32 <andythenorth> shitloads of things are done in software because 'someone said it's correct'
18:37:37 <andythenorth> that's not engineering
18:37:58 <SpComb> I personally appreciate having someone else review my code and give comments
18:38:37 <SpComb> doesn't answer your question, though
18:40:50 <nielsm> map array changes is the nuclear plant the famous bike shed belongs to
18:40:58 <Eddi|zuHause> maybe for "trusted devs" the review should be more optional? like you make a pull request, and if no other dev has said "hold on, i want to review that first" within a certain time, then you get to just commit it?
18:41:27 <andythenorth> yes. the argument isn't "are reviews good or bad"
18:41:35 <andythenorth> it's about gates to progress
18:42:09 <SpComb> it certainly requires someone to also commit to reviewing stuff
18:44:09 * nielsm is downloading an actual linux iso via bittorrent
18:45:59 <nielsm> huh apparently I've had a linux vm running in HyperV for possibly years here and totally forgotten about it
18:46:08 <nielsm> and I mean actually running
18:47:09 *** frosch123 has joined #openttd
18:48:19 <peter1138> Bit out of date then?
19:09:21 <Alberth> safe to say it works :p
19:25:03 <peter1138> Current Game Speed Factor: 1.02x
19:26:25 <LordAro> peter1138: what's the future like?
19:29:48 *** Wacko1976_ has joined #openttd
19:44:21 <TrueBrain> peter1138: you always "had a patch for that", even when there was no review system; are you sure the review system is to blame? :)
19:44:22 <andythenorth> 30 days, no objections, clear to merge?
19:44:29 <andythenorth> ^ prone to being gamed
19:44:32 <andythenorth> oh it's TrueBrain :)
19:45:06 <TrueBrain> I always wonder why, especially full-fletched-devs, can be so pessistic yet action-less
19:45:11 <TrueBrain> if you don't like something, change it
19:45:25 <TrueBrain> don't sit on the side of the road like a grumpy old man holding your stick up high
19:45:29 <TrueBrain> do something about it!
19:45:34 <TrueBrain> don't expect someone else to do it or something
19:45:54 <TrueBrain> this community has the huge tentency to go back to his hibernation when someone is not yelling loud enough to move forward
19:46:06 *** gelignite has joined #openttd
19:46:32 <TrueBrain> is your PR not getting a review? Tell people and "force" them to review, how-ever nice you can do that (not by passive-aggresively making statements in this channel btw; that is not a good option :D)
19:46:46 <TrueBrain> is your PR not being merged? Ask yourself why not! Talk about it! Come up with ideas to solve it!
19:46:55 <TrueBrain> fix your own problem! This is Open Source!
19:47:04 <TrueBrain> be active! Even better: be PROACTIVE!
19:47:27 <TrueBrain> but most of all: grow some balls and take action
19:47:32 <TrueBrain> if you consider NRT good enough: fucking merge it
19:47:36 <TrueBrain> if it blows up, it blows up
19:47:46 <TrueBrain> OpenTTD has a history of shit being blown up
19:47:48 <andythenorth> there's always revert
19:48:09 <TrueBrain> no movement is always worse than any movement, in which ever direction
19:48:17 <TrueBrain> but worse than no movement, complaining about the lack of moment
19:48:50 <TrueBrain> the only (!!) thing that is currently enforced, is that the CI approves a PR, and that a dev merges the PR
19:48:57 <TrueBrain> there is no restriction on WHO reviewed it
19:49:25 <TrueBrain> and if you think that system sucks: come with a suggestion!
19:49:43 <TrueBrain> (am I rattling this cage hard enough? Or should I continue?)
19:50:06 <TrueBrain> why isn't this accepted yet?
19:50:08 <peter1138> I'm not complaining about the system!
19:51:53 <TrueBrain> [18:32] <peter1138> As it is, I've effectively given up coding anything because of the review system.
19:51:59 <TrueBrain> sorry, you could have fooled me :D How should I have read it? :)
19:53:40 <TrueBrain> (and just to be clear, the above was aimed at any person in this channel; get active! Stop being so passive!)
19:56:29 <TrueBrain> peter1138: PR#6780 -> there are some comments that are not addressed yet; if they are fixed, that seems mergable?
19:59:30 <nielsm> lol, just discovered that the configure script doesn't actually check the user has make installed
20:09:23 <TrueBrain> no comments about the content of PR#6821 .. that is a bit sad
20:09:33 <TrueBrain> who to slap for that .... hmmmmmmmmmm
20:10:00 <TrueBrain> frosch123: does it make sense to, on load of a savegame in Single Player, convert all other companies to AIs?
20:11:23 <frosch123> i somewhat doubt that
20:11:44 <frosch123> likely ais won't make sense of the already built stuff
20:11:48 <TrueBrain> it feels a bit weird, but I dont know why
20:12:01 <frosch123> and some people like to play multiplayer with themself
20:12:27 <frosch123> by switching companies
20:12:42 <frosch123> or playing a bit on their own and then continue in multiplayer or whatnot
20:12:46 <nielsm> is there an (easy) way to convert an existing player company to an AI if you ever want to?
20:13:02 <frosch123> so. in essence, i can imagine more people complaining about conversion than prasiing :p
20:13:21 <frosch123> nielsm: i don't think so
20:13:27 <frosch123> but maybe adding a console command makes sense
20:13:44 <frosch123> force ai onto existing company, and remove an ai without klling the company
20:13:45 <TrueBrain> so we are going to reject the PR than :)
20:13:56 <Alberth> simplest option for the ai is to sell everything and start anew
20:14:04 <frosch123> currently we have "startai" and "stopai". maybe people want "attachai" and "detachai"
20:19:25 <TrueBrain> frosch123: tnx; updated tickets, closed PRs
20:20:14 <nielsm> interesting it runs at slightly faster than 33.33 here
20:23:38 <TrueBrain> nielsm: yeah, tri-state is the bare minimum; it could be a lot better. But I am already happy if someone adds tri-state tbh :D
20:26:26 <TrueBrain> frosch123: tldr, #6846, did he fix all your mentioned points?
20:26:53 <frosch123> i'll check on the weekend
20:26:58 <frosch123> did not manage last weekend
20:27:35 <TrueBrain> was mostly wondering, and didnt feel like reading :D
20:28:09 <TrueBrain> and ugh, I did so little of what I planned .. the stupid weather is really draining me :(
20:28:14 <TrueBrain> I like the forest PR :)
20:30:21 <TrueBrain> tempted to merge it, but I should validate it first ..
20:30:29 <TrueBrain> wish I had a bot I could tell to make me a binary
20:30:33 <TrueBrain> owh wait, that was what I was building
20:32:38 <TrueBrain> I wonder about 6784 .. feels like someone needs to say "left" or "right" on it ..
20:33:35 <nielsm> unless anyone has objections for fps-meter I think it's also ready to merge now?
20:33:44 <nielsm> or someone thinks it needs re-testing on mac
20:33:46 <Wolf01> Ships need to be fixed in many ways
20:35:11 <TrueBrain> Wolf01: there is a bit of politics involved; if we can make J0anJosep happy by adding a simple statement, it makes his work in his patchpack easier
20:36:33 <TrueBrain> something about an upwards spiral of hapiness
20:36:40 <TrueBrain> if you help someone else, he is more likely to help you :)
20:37:13 <TrueBrain> so what I wonder, how much code is it
20:41:13 <andythenorth> TrueBrain: I am -1 to the forest patch
20:41:25 <andythenorth> I would rather tree algorithm was newgrf content
20:42:06 <LordAro> perhaps, but if "tree placement as newgrf" is going to take years and years, why not add this now?
20:42:08 <TrueBrain> which is unlikely to happen ... that is a thoughy!
20:42:14 <LordAro> it can always be taken out later
20:42:44 <TrueBrain> it depends .. do we want "master" to be as close to original as reasonable, and extend by addons (read: newgrfs)
20:42:50 <TrueBrain> or do we extend in "master" too
20:43:00 <TrueBrain> as if we add forest, having tree placement in newgrf is even more unlikely
20:43:05 <Wolf01> I'm for extending with newgrfs
20:43:12 <andythenorth> content all the way
20:43:13 <TrueBrain> (lets put away if newgrfs are the correct solution for this btw)
20:43:27 <TrueBrain> currently there are two ways to extend OpenTTD: NewGRFs and Squirrel
20:43:29 <andythenorth> extensible, smaller core
20:43:34 <TrueBrain> rest of the world uses lua :D
20:43:49 <andythenorth> fewer conflicts, fewer tarpits
20:43:56 <andythenorth> more state = more tarpit
20:44:08 <andythenorth> more tarpit = slower development, or total stasis
20:44:08 <TrueBrain> I really like forest, but I do agree ... but do we want to become the project that rejects all new ideas? :D
20:44:30 <andythenorth> if only there was a really active patchpack :(
20:45:02 <TrueBrain> if only OpenTTD still had GPMI in its core ... :P (that allows binary extension via addons)
20:45:36 <TrueBrain> frosch123: do you agree, should we reject forest patch (refering him to the above URL)?
20:46:21 <frosch123> TrueBrain: i have not looked at the tree patch
20:46:38 <TrueBrain> in very simple words: it adds another tree-placement algorithm, that favors forests
20:46:41 <frosch123> but newgrf- or squirrel-based tree stuff is the argument, that is not going to happen
20:46:51 <TrueBrain> (YATP - Yet Another Tree Placement)
20:47:15 <TrueBrain> no, the argument is: should 'master' have YATP, or should it be by extension (either by an addon or by a patchpack)
20:47:57 <frosch123> i don't see the latter as a valid option
20:48:06 <TrueBrain> care to explain why?
20:48:11 <TrueBrain> (sounds more rude than intended :D)
20:48:17 <TrueBrain> could you elaborate? :D
20:48:17 <frosch123> there are many trees on the map
20:48:25 <frosch123> neither newgrf nor scripts are good with managing that
20:48:42 <andythenorth> do we really need so many trees?
20:48:42 <TrueBrain> isn't that irrelevant for the choice? HOW the latter should be solved, is a second talk?
20:48:46 <frosch123> so i am quite sure there will never be "addons" for tree placement
20:49:02 <TrueBrain> but if we choice when to add stuff to 'master' based on how difficult it is to do it in an addon, line becomes very blurry?
20:49:11 <TrueBrain> patchpacks are also addons, in some sense?
20:49:19 <frosch123> anyway, i ignore trees with invisiblity. so i have no opinion on the tree patch itself
20:49:20 <andythenorth> I would prefer to continue forcing choices by preferring content
20:49:24 <frosch123> if it is good, it should be added
20:49:35 *** Progman_ has joined #openttd
20:49:53 <andythenorth> I don't want to discuss an implementation, but tree tiles could be done with much less variety and nobody would notice
20:49:56 <frosch123> i assume it does not affect savegames
20:50:10 <frosch123> so changing/removing tree placement is fine any time
20:50:29 <TrueBrain> but do we want to bother with that in 'master', or is that something for patchpacks?
20:51:01 <andythenorth> patchpacks, possibly port a winning implementation if there is one
20:51:15 <andythenorth> growing more half-assed variants is not appealing
20:51:24 <andythenorth> and we don't seem to delete old dead implementations
20:51:25 <TrueBrain> well, you do want to prevent another PBS, yes :P
20:51:45 <TrueBrain> that last one, is the tricky part; once added to master, removing it is very difficult, as people start to depend on it
20:51:52 <andythenorth> or another New Economy
20:51:53 <TrueBrain> people still complain that the original AI is removed
20:52:06 <TrueBrain> but OpenTTD with TGP is .. ughhhhh, sucky
20:52:14 <TrueBrain> so there are really two sides to this coin :)
20:52:15 <frosch123> reading it, i would complain about the description :p
20:52:18 <andythenorth> 'people complain' doesn't actually mean very much
20:52:22 <frosch123> "'Original' plants trees uniformly scattered, 'Improved' plants them in groups, 'Forest' plants them in forests"
20:52:34 <frosch123> ^^ that says nothing
20:52:45 <frosch123> maybe it should say "large groups"?
20:52:47 <andythenorth> 'people complain' is easily circumvented by not reading tt-forums or the german forum :P
20:52:56 <TrueBrain> you nitpick on that in the PR, I was trying to see what we are at in the more high-level picture of this all :)
20:53:26 <TrueBrain> when are a group of trees a forest :)
20:53:37 <TrueBrain> for me the description reads fine btw :)
20:53:38 <andythenorth> when nobody hears them fall
20:54:09 <frosch123> anyway, everything the patch adds can be removed without breaking anything
20:54:41 <TrueBrain> okay; than I am going to merge it
20:54:44 *** Progman_ is now known as Progman
20:54:46 <TrueBrain> or do you want the description changed?
20:55:00 <frosch123> so, if anyone has tried it, and subjectively likes the appearance, i am +1 to it
20:55:14 <TrueBrain> ugh, this "trying" again ....
20:55:18 <TrueBrain> andythenorth: can you try that PR? :P :P
20:55:25 <andythenorth> give me 5-10 mins
20:56:05 <TrueBrain> okay ... FPS meter ... think everyone pissed over that enough ... I see no blockers ....
20:56:09 <frosch123> changing the text can be a separate pr
20:56:11 <TrueBrain> merging time? (going once ... goince twice ...)
20:56:20 <frosch123> "improved" is just as stupid :)
20:56:23 * andythenorth needs the new quad core mac
20:58:02 <TrueBrain> now to squash this FPS .....
20:58:03 <andythenorth> wow my fans span up :P
20:58:11 <TrueBrain> as 49 commits is a bit much
20:58:34 <andythenorth> I know Apple analogies are like cat sick
20:58:41 <TrueBrain> nielsm: only the first commit-message is sufficient? Or does it need more?
20:58:42 <andythenorth> but I read this today, re 'people will complain'
20:59:27 <TrueBrain> C++11 stuff .. oef .. that is a difficult one
20:59:47 <andythenorth> ok the forest patch is bad
20:59:56 <andythenorth> I need to comment on PR :P
21:00:04 <andythenorth> just thinking how to do it kindly
21:00:14 <andythenorth> it makes random diamond shaped groups of treees
21:01:26 <TrueBrain> and yes: BE NICE! :)
21:02:05 <andythenorth> - leaves no trees anywhere else
21:02:10 <TrueBrain> diamond-shapes are to be expected :)
21:02:15 <andythenorth> - conflates with forest industry
21:02:15 <TrueBrain> and that was the idea: forest-only
21:02:46 <andythenorth> - options makes no sense in world-gen menu 'none', 'original', 'improved', 'forest'
21:02:49 <TrueBrain> feels like it needs a settings between "improved" and "forest" :D
21:02:51 <andythenorth> UX: which do I pick?
21:03:05 <andythenorth> why don't we just fix on one good, and remove the bullshit?
21:03:06 <TrueBrain> yeah, as frosch123 said, "improved" is now silly
21:03:26 <TrueBrain> okay, make this into a nice reply, state a clear opinion and direction for the author to execute
21:03:41 <TrueBrain> (please don't leave a: you figure it out, reply; that is not nice for a first-time contributor)
21:03:59 <andythenorth> yeah, these replies demand the most care
21:05:17 <TrueBrain> frosch123: to bother you some more (sorry :D), adding 2 bytes to _map, your thoughts? How is memory etc these days?
21:06:21 <frosch123> you can also add 8 just fine :p
21:06:23 <TrueBrain> percentage-wise it is less and less :D
21:06:40 <frosch123> map array is neglible to the rest
21:06:50 <TrueBrain> "back in my day", it was a huge impact, as adding bytes often meant some players no longer could play big maps
21:06:59 <frosch123> esp. on clients with 32bpp
21:07:14 <TrueBrain> is there any "policy" in place for assignments? Or is a PR like peter1138's fine?
21:08:39 <Eddi|zuHause> but... map array has become this mythical catch-all for rejecting any major feature. you can't just expand it on a whim and undermine that :p
21:08:53 <frosch123> we try to put the same stuff into the same bits for different tile types
21:09:12 <frosch123> so, unless the patch contains switches on tiletype to read on different bits, it should be fine
21:09:22 <TrueBrain> *picks up what-ever Eddi|zuHause says, opens the door, puts it next to the trash*, anything else? :D
21:10:13 <TrueBrain> that only leaves me to wonder ... he added a m8 as 16bit .. while most (not all) are 8bit
21:10:22 <TrueBrain> better to add 2 8bit, or are we just going all-the-way here
21:10:58 <frosch123> i think we talked about that before, and i prefered 16bit
21:10:59 <TrueBrain> (only 6 bits are used, of the 16bit)
21:11:37 <frosch123> iirc peter had multiple versions of the patch, one with 2x8, and one with 1x16
21:11:43 <frosch123> and the favour was for 1x16
21:11:52 <TrueBrain> ack; that means someone just has to review the moving (and expanding) of railtypes
21:12:14 <TrueBrain> and don't share that favour, but I don't care enough (only care enough to mention it :P)
21:13:01 <frosch123> the idea was that reading a subset of bits is easier than merging multiple bytes into something bigger
21:13:27 <frosch123> we have some things, like objects, where we store ids over multiple bytes
21:13:40 <frosch123> like m2 and m7 bits 2 to 5 or something
21:13:55 <andythenorth> nielsm: re trees, it does just need someone to pick it up and drive it through
21:14:01 <andythenorth> many things have been declared impossible before
21:14:17 <andythenorth> 'performance reasons' is ~50% bullshit excuse :)
21:14:30 <TrueBrain> nielsm: but is the first commit message of your FPS sufficient? (I am about to squash the whole damn thing into a single commit, and want to be sure that is okay)
21:14:48 <nielsm> TrueBrain let me just check a bit
21:14:56 <TrueBrain> yeah ... maps not being 256x256 was impossible!
21:15:07 <TrueBrain> (and that was my first patch for OpenTTD .. .14 years ago :P)
21:15:39 <TrueBrain> tnx btw frosch123! Always helps to know in a PR what was already talked over :)
21:16:06 <Eddi|zuHause> most of my patches were 1-3 lines, of things other people were apparently not bothered enough to do them
21:16:30 <TrueBrain> I added some labels in Pull Request
21:16:32 * andythenorth looking for the 'most of savegame is trees' issue
21:16:34 <nielsm> TrueBrain, if you can make commit message the title of PR + first paragraph in PR initial post
21:16:44 <nielsm> i.e.: Frame rate and various game loop/graphics timing measurements and graphs. Accessible via the Help menu, and can print some stats in the console via the `fps` command.
21:16:54 <Eddi|zuHause> i think my biggest patch that got in (line-wise) was the screenshot thing
21:17:30 <TrueBrain> seriously nicely done for such a huge patch
21:17:32 <Eddi|zuHause> where it took the area that is currently visible, but applied the max zoom to it
21:17:36 <TrueBrain> how we say it in dutch: "petje af"
21:17:43 *** Thedarkb has joined #openttd
21:18:16 <TrueBrain> so that leaves NRT and std:vector
21:18:16 <Eddi|zuHause> which turned out not very difficult to do, just take the whole-map screenshot routine, and add some limits to it
21:18:59 <Eddi|zuHause> all of the infrastructure for it was already in place, just had to be combined the right way
21:19:07 <TrueBrain> frosch123: I assume you agree'd withi the std::vector stuff? (last quesiton of the evening, I promise)
21:19:46 <TrueBrain> nielsm: your comment in #6817, is a bit vague to me; are those two changes in there, or should they be added?
21:20:06 <nielsm> I think the std::vector patch requires C++11 but not sure
21:20:20 <TrueBrain> LordAro: related to #6817, do we need to enforce build systems to use C++11? With that PR, won't they simply fail?
21:20:54 <nielsm> TrueBrain the GUIBrideList is not being fixed iirc
21:21:05 <TrueBrain> can you make that more clear in your comment, so the author knows if he has any actions left?
21:21:26 <TrueBrain> (in general in pull requests, please make it very obvious to the author if it is an action on his side, or just a remark :D)
21:22:08 <TrueBrain> and as it is a first-time contributor, make it something nice :) "Would you mind to also fix ... as they are really important to fix ..." :D
21:22:55 <TrueBrain> ah, that introduces more of that 'auto' stuff ... 'auto' is dutch for 'car', so in my head I always go: BROOM BROOM
21:23:58 <frosch123> TrueBrain: it has some codestyle issues, but the principle is fine
21:24:02 <andythenorth> so the forest patch....is contentish trees ruled out or no?
21:24:05 <andythenorth> before I mislead Sylvain
21:24:21 <andythenorth> or it's a non-issue either way?
21:24:50 <nielsm> also just before I was so busy looking at code I almost forgot my pizza in the oven :(
21:25:02 <nielsm> got it out while it was still edible
21:25:09 <andythenorth> I cook using timer on my phone
21:25:24 <andythenorth> primary benefit of smartphone
21:25:25 <nielsm> yeah I usually set a timer, forgot it this one time
21:25:34 <TrueBrain> frosch123: ack, tnx :)
21:25:49 <frosch123> i think noone read all the diffs :p
21:26:11 <TrueBrain> my oven just stops after the time I put in
21:26:14 <TrueBrain> avoids any mistakes :D
21:26:37 <andythenorth> mistake of forgetting to eat the food? o_O
21:27:10 <nielsm> ah yeah leaving commented-out code is a bad smell
21:27:57 <TrueBrain> so, it needs a review; that is fine :)
21:28:19 <TrueBrain> we closed 124 PRs already :o
21:28:52 <TrueBrain> happy the FPS stuff is in :D
21:29:12 <andythenorth> TrueBrain: damn yes
21:29:16 <andythenorth> I counted at the weekend
21:29:25 <andythenorth> just need binaries now :)
21:29:25 <TrueBrain> the number... is just there
21:29:32 <TrueBrain> YES! Working on it ..
21:29:43 <TrueBrain> tomorrow call with Google Cloud Platform .. I so hope they can offer something
21:29:49 <TrueBrain> would make shit so much easier (and faster)
21:30:01 <andythenorth> oof I counted commits manually (by counting pages)
21:30:11 <andythenorth> I should github better
21:35:25 <Alberth> better call microsoft nowadays :)
21:37:07 <andythenorth> in the tank game we say 'git gud' when players are lame
21:37:12 <andythenorth> I need to 'github gud' :P
21:40:33 <nielsm> okay I'm impressed how MS managed to hide Safe Mode in win10 behind a maze of buttons
21:40:50 <nielsm> it seems to prefer you reinstall windows before trying to boot to safe mode
21:41:28 <Alberth> instead of calling google
21:44:31 <TrueBrain> Azure will be fine too; but for kubernetes, GCP is better :)
21:47:06 <LordAro> TrueBrain: ideally the build system should be modified indeed
21:47:35 <TrueBrain> a bit more snippets of information would be greatly appreciated :D
21:47:38 <LordAro> otherwise someone will try compiling with something ancient and get some obscure error
21:47:50 <TrueBrain> modified in what sense?
21:47:54 <TrueBrain> just 'modified'? :D
21:47:59 *** sim-al2 has joined #openttd
21:48:20 <LordAro> some sort of test program probably
21:48:36 <TrueBrain> owh, you want to add a check if you are using a C++11 capable compiler!
21:48:41 <TrueBrain> see, that is more sensible :)
21:48:46 <LordAro> sorry, wasn't clear :p
21:48:52 <TrueBrain> no, it was very vague :)
21:49:00 <TrueBrain> I guess porting to cmake is easier :P
21:51:07 <LordAro> Alberth: i tried that, TB shouted at me :p
21:52:11 <TrueBrain> wait, who needs to be "corrected? :P
21:52:47 <TrueBrain> LordAro: in CMake you just say the lowest allowed compiler, and it will do its thing correctly :D Which is really nice
21:57:03 <TrueBrain> andythenorth: also a note on the forest patch: theyALL have EXACTLY the same size
21:58:03 <andythenorth> I am writing a comment
21:58:15 <andythenorth> kind of hard because I can never approve it
21:58:19 <andythenorth> unless I can? :P
21:58:36 <TrueBrain> otherwise by proxy! :P
22:01:31 <andythenorth> done that comment
22:24:58 <Eddi|zuHause> there are seriously chemicals called "Alite" and "Belite"?
22:27:19 <andythenorth> sounds like a FIRS economy
22:27:39 <Eddi|zuHause> apparently they're components in cement
22:28:05 <Eddi|zuHause> with different ratio of Ca to Si
22:28:50 <Eddi|zuHause> one is Ca3SiO5 and the other Ca2SiO4
22:28:57 <Eddi|zuHause> and i instantly forgot which is which
22:29:52 <frosch123> there is also cerite
22:31:23 <andythenorth> super extra detail
23:52:36 *** sim-al2 has joined #openttd
continue to next day ⏵
