IRC logs for #openttd on OFTC at 2019-10-19
⏴ go to previous day
00:41:53 *** WormnestAndroid has quit IRC
00:41:54 *** WormnestAndroid has joined #openttd
04:41:33 *** WormnestAndroid has quit IRC
04:42:20 *** WormnestAndroid has joined #openttd
04:50:23 *** WormnestAndroid has quit IRC
04:50:48 *** WormnestAndroid has joined #openttd
08:03:30 *** arikover has joined #openttd
08:41:26 *** HerzogDeXtEr has joined #openttd
09:01:33 *** andythenorth has joined #openttd
09:19:17 *** Progman has joined #openttd
09:26:21 *** andythenorth has joined #openttd
10:04:21 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro approved pull request #7778: Doc: split compiling instructions from README, as they aren't requir… https://git.io/JeBE8
10:20:20 <andythenorth> that one's fake, but eh
10:35:00 *** sla_ro|master has joined #openttd
10:58:18 *** WormnestAndroid has quit IRC
10:58:32 *** WormnestAndroid has joined #openttd
12:00:40 <andythenorth> seems quite overwhelming?
12:02:55 <nielsm> "installing is fairly straightforward. <five long paragraphs of text followed by five subsections>"
12:03:55 <nielsm> installation instructions would perhaps be simpler if we could distribute ogfx directly in the package
12:04:40 <nielsm> or if the automatic bananas baseset download if a base graphics is missing, worked everywhere (supposedly it does not)
12:07:26 <andythenorth> above my pay grade?
12:17:24 <andythenorth> afaict, all of 'debugging_desyncs' is already in 'desync'
12:52:36 *** stefino has joined #openttd
12:53:28 <stefino> Hi. Sorry for the same question again but will be able to have different roadstops overlays in the future?
12:53:47 <andythenorth> there is no plan to do that
12:54:00 <andythenorth> almost everything for the future of OpenTTD is 'unknown'
12:54:56 <andythenorth> ^^ can haz docs review? LordAro? :)
12:55:30 <andythenorth> we have 30% of a gameplay manual in /docs :P
12:55:48 <andythenorth> well more of a technical manual actually :)
12:56:04 <stefino> yes yes I know :) it is only an idea because our roadtypes are divided by year and would be nice to have old, new ones , bus stops from village or city
12:56:54 <stefino> in case that is possible to have underlays for it
12:58:10 <stefino> I don't know how difficult is to code it but it will be good for us (for 4x zoom)
13:00:02 <stefino> little bit strange to have modern roadstop made of metal and glass in 1920 :D
13:11:19 <DorpsGek_III> [OpenTTD/website] andythenorth updated pull request #106: Create 2019-10-19-monthly-dev-post.md https://git.io/JeBzW
13:11:47 <DorpsGek_III> [OpenTTD/website] andythenorth updated pull request #106: Create 2019-10-19-monthly-dev-post.md https://git.io/JeBzW
13:12:08 <DorpsGek_III> [OpenTTD/website] andythenorth updated pull request #106: Create 2019-10-19-monthly-dev-post.md https://git.io/JeBzW
13:12:36 <DorpsGek_III> [OpenTTD/website] andythenorth updated pull request #106: Create 2019-10-19-monthly-dev-post.md https://git.io/JeBzW
13:14:09 <DorpsGek_III> [OpenTTD/website] andythenorth updated pull request #106: Create 2019-10-19-monthly-dev-post.md https://git.io/JeBzW
13:16:27 <DorpsGek_III> [OpenTTD/website] andythenorth updated pull request #106: Create 2019-10-19-monthly-dev-post.md https://git.io/JeBzW
13:35:24 *** Kitrana1 has joined #openttd
13:52:51 <DorpsGek_III> [OpenTTD/website] James103 commented on pull request #106: Create 2019-10-19-monthly-dev-post.md https://git.io/JeBzi
13:57:44 <andythenorth> does README.md need to work for people with no internet access?
13:57:53 <andythenorth> e.g. those playing via CD-ROM etc?
13:58:01 <nielsm> it's included in the binary distribution
13:58:23 * andythenorth finding many overlaps between README, other /docs, OpenTTD website
13:58:25 <andythenorth> and obviously, wiki
13:58:40 <andythenorth> and due to natural entropy, the information varies :)
13:58:56 <andythenorth> my preference is fewer places, more canonical, less maintenance
14:11:24 <glx> yeah it's a little annoying :)
14:11:58 <andythenorth> "git mv docs/HOWTO_compile_lang_files.txt docs/HOWTO_compile_lang_files.md"
14:12:03 <andythenorth> didn't keep history? :o
14:12:15 * andythenorth thought that was correct command :P
14:12:25 <glx> it worked for other files
14:12:45 <nielsm> make sure the rename is the only thing in the commit
14:12:47 <andythenorth> let me see if I can trick it
14:12:48 <LordAro> it's sometimes a little funny how it displays file moves
14:15:09 * andythenorth testing what it does now
14:15:35 <glx> 5 files changed instead of 6
14:17:24 <andythenorth> and docs/HOWTO_compile_lang_files.txt is gone
14:17:36 <andythenorth> even though my git stage showed it as added as a new file
14:18:18 <andythenorth> ok I see what I did
14:20:44 *** frosch123 has joined #openttd
14:21:11 <andythenorth> ok so again, that broke the history
14:21:41 <andythenorth> I must be doing git wrong
14:21:46 <andythenorth> what's the correct way to rename?
14:22:16 <andythenorth> git mv docs/HOWTO_compile_lang_files.txt docs/HOWTO_compile_lang_files.md is wrong
14:23:32 <glx> history is ok I think, as the rename is shown in the diff
14:25:28 <andythenorth> ok so that should be back in same state now :P
14:25:31 * andythenorth adventures in git
14:36:42 <andythenorth> canonical score 0.5 out of 1 :P
15:00:31 <glx> thats crazy, I can move it, then edit it, but once I squash the commits it returns to delete txt create md
15:03:57 <glx> but I'm sure there's a solution
15:08:36 <nielsm> yeah git doesn't have a concept of rename as such
15:08:44 <nielsm> you need to keep a commit that does nothing but rename
15:21:35 <glx> seems it's because the new files contains too many modifications
15:23:57 <nielsm> git guesses renames based on amount of identical file contents
15:24:13 <nielsm> so the safest is to keep the rename in a separate commit
15:25:18 <glx> andythenorth: I can split the commit in 2, so first rename, then edit for all files
15:31:28 <andythenorth> what happens when we squash it though?
15:35:03 <nielsm> it needs to stay separate commits forever
15:35:09 <nielsm> otherwise it doesn't count as a rename
15:45:07 <andythenorth> I'm drafting a proposal for what docs to put where
15:49:33 <nielsm> ideally should be made more pretty-layout-y
15:52:53 <andythenorth> do we have control of eints?
15:53:03 <andythenorth> I keep finding translation info around the place
15:53:08 <andythenorth> sometimes outdated
15:53:22 <andythenorth> and put the instructions in place there
15:55:25 <glx> ahah the diff containing both commits is still wrong, but each commit is ok
16:10:21 <frosch123> mind the "openttd" branch, it's not "master"
16:10:45 <andythenorth> yes that's helpful thanks
16:10:51 <andythenorth> is eints deployable?
16:10:59 <andythenorth> or is it yak-shaving?
16:11:23 <frosch123> it requires a manual "git update"
16:11:38 <frosch123> but it is not tb-magic-automatic
16:12:14 * andythenorth making a list of what should be where
16:12:24 <andythenorth> probably some weeks / months of work for me to do it all :)
16:21:52 <andythenorth> strikes me it could have been (1) on the wiki or (2) moved into website
16:25:28 *** WormnestAndroid has quit IRC
16:28:31 *** WormnestAndroid has joined #openttd
16:36:33 *** WormnestAndroid has quit IRC
16:37:26 *** WormnestAndroid has joined #openttd
16:58:07 *** Wormnest has joined #openttd
17:21:21 *** Wormnest_ has joined #openttd
17:26:26 *** WormnestAndroid has quit IRC
17:45:40 *** WormnestAndroid has joined #openttd
17:49:57 *** arikover has joined #openttd
17:52:48 <andythenorth> is it known how to become a translator?
17:54:17 <glx> yes that's the procedure, send an email and wait
17:54:39 *** Wormnest__ has joined #openttd
17:55:16 <glx> but there are some ideas about changing this procedure
17:56:25 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro merged pull request #7778: Doc: split compiling instructions from README, as they aren't requir… https://git.io/Je4DK
17:56:48 <frosch123> andy discovering his own work :)
17:59:27 <frosch123> hmm, compiling wine reminds me of my gentoo days
18:11:06 *** ChanServ sets mode: +v tokai
18:16:16 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro approved pull request #7759: Codechange: Use std::vector for industry tile layouts https://git.io/JeBVz
18:16:28 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro merged pull request #7759: Codechange: Use std::vector for industry tile layouts https://git.io/Jecjs
18:33:31 *** Xaroth7 has joined #openttd
18:41:25 *** tokai|noir has joined #openttd
18:41:25 *** ChanServ sets mode: +v tokai|noir
19:18:49 <glx> seems correct but I'm not a nml person
19:19:20 *** gelignite has joined #openttd
19:19:45 <andythenorth> I used to have admin rights on nml
19:20:06 <andythenorth> but I gave them up so we could move them under official support
19:22:00 <FLHerne> When I run `make test`, I just get `Running test 001_action8`, `Running test 002_...`
19:22:06 <FLHerne> I assume that means they passed?
19:23:09 <andythenorth> they fail visibly if they fail
19:27:02 <andythenorth> wondering if there should be an nml team
19:27:20 <glx> that's what I was thinking
19:27:30 <andythenorth> TB made comments about the nml project config, can't remember what
19:29:13 <frosch123> none of the projects except main ottd have any rules configured
19:30:13 <glx> indeed there's no actual need for approval in nml
19:30:14 <frosch123> it's a lot of yak shaving and looking up what settings anctually mean
19:32:12 <andythenorth> I wonder if we know what we actually want
19:32:40 <andythenorth> the advantage on coop was that I was terrified of how much power I had :P
19:32:50 <andythenorth> so I rarely touched anything
19:33:10 <andythenorth> full project admin rights, full merge rights, no approval process, and quite lax issue tracking
19:33:23 <glx> we could create an nml team with admin rights on nml repo I guess
19:33:41 <frosch123> just don't allow andy to make merge commits :)
19:33:54 <glx> merge commits are disabled in the settings
19:34:31 <frosch123> my problem with gh is that i have no idea what someone is allowed to do, given some settings
19:34:36 <glx> but we also need a workflow checking commits and compilation
19:34:47 <frosch123> most settings are a mystery to me
19:35:39 <andythenorth> well, at the risk of suggesting a policy...
19:35:56 <andythenorth> could we cargo-cult the OpenTTD settings? :P
19:36:01 <glx> basically there are 2 types of peoples in the organisations, owners and members
19:36:26 <glx> owner are unlimited, members can have access depending on teams
19:39:52 <glx> oh there are also collaborators
19:42:08 <glx> to give access on a repo base for people not in the organisation
19:43:09 <andythenorth> all the access controls :)
19:46:41 <glx> hmm master is not protected for nml
19:46:55 <glx> we should probably enable that
19:47:03 <andythenorth> I think that was TB's main point
19:47:43 <frosch123> yeah, there is no way to do that for all the 20 repositories at once
19:48:10 <frosch123> wow, i just guessed the number, but we actually have exactly 20
19:48:30 * andythenorth wonders if there's a policy file approach
19:48:40 <glx> just added basic protection
19:49:10 <glx> ie no forced push to master and no deletion
19:49:26 <frosch123> anyway, if there was a separate nml team, would there also be separate teams for opengfx, sfx, msx, ... i.e. lots of empty teams?
19:50:46 <glx> nml team would just be to allow more power for some people but only on the nml repo
19:51:36 <glx> owner equivalent but without access to everything
19:52:13 <glx> like andy could manage all settings in nml
19:52:31 <glx> and add people to help him too
19:52:50 <glx> without relying on OpenTTD owners
19:53:18 <andythenorth> I wondered about inner and outer wheel
19:53:24 <frosch123> i don't think adding members is the bottleneck, it more about who can approve prs
19:53:27 <andythenorth> rather than specific teams
19:53:54 <andythenorth> there might be projects where we trade velocity against security
19:54:03 <andythenorth> or to explain differently
19:54:21 <andythenorth> when we moved from FS + forums + svn to github, core commits and activity increased
19:54:27 <andythenorth> because GH is much better than those ways
19:54:39 <andythenorth> but when we moved some projects from coop to GH, activity has died
19:54:51 <andythenorth> because GH has imposed a workflow that we have nobody interested in using
19:55:06 <glx> hmm nothing was imposed on nml
19:55:43 <glx> except having write access to the repo
19:56:59 <andythenorth> so can I just click the button?
19:57:08 <andythenorth> I am totally unclear who the maintainer is for nml
19:57:12 <andythenorth> is that like poke?
19:57:18 <andythenorth> does it mean it's me :(
19:57:23 <glx> if it's green you can click
19:57:46 <glx> maintainer for nml is currently OpenTTD owner members
20:01:08 <glx> well there's noone in maintainer role, as both teams just have 'write' permission only (none has 'maintain' or 'admin' access)
20:01:44 <andythenorth> I mean more like the politics of it
20:01:47 <glx> and of course owners are admin
20:02:34 <andythenorth> anyway, if there are no current objections, I am going to start approving my own PRs after a few days grace
20:03:25 <LordAro> approving your own PRs? :o
20:03:30 <frosch123> was the same fore coop and ottd-svn
20:04:57 <andythenorth> website should be on same basis really
20:05:13 <andythenorth> 'you break it, you own it' :P
20:05:26 <andythenorth> frosch123: do you not remember picking my bad commits out of hg nml? :P
20:05:59 <frosch123> now that you mention it :p
20:06:10 <andythenorth> I had sweaty palms :P
20:06:22 <frosch123> anyway, i have no idea what gh setting apply to the gui/pr,and what to direct pushes
20:06:52 <frosch123> but i was refering to 2016, when only albert was reviewing every now and then
20:07:14 <andythenorth> those were dark times :)
20:07:22 <andythenorth> there was one frog-shaped maintainer
20:07:28 <andythenorth> and I really thought you might just quit
20:08:01 <andythenorth> then I deleted, ahem 'closed', 500 FS issues, and peter took a run at FS patches
20:08:12 <frosch123> there were no infrastructure issues, so i could just ignore everything boring :)
20:08:49 <andythenorth> unrelated, should I just have ssh on bundles?
20:09:21 <andythenorth> I don't actually need to automate publishing the artefacts for every rev from CI
20:09:47 <andythenorth> manually uploading tags would be better than current situation, where all my grfs are outdated on bundles :)
20:10:05 <andythenorth> I can scp quite reliably, if I look up the man page on google :P
20:10:45 <andythenorth> how far in minecraft?
20:10:51 * andythenorth invents minecraft pedometer
20:15:53 <DorpsGek_III> [OpenTTD/nml] andythenorth merged pull request #48: Docs: (fixes #45) remove unused spritesets from roadtype example https://git.io/Je48T
20:15:55 <DorpsGek_III> - Doc: (fix) remove unused spritesets from roadtype example (#48) (by andythenorth)
20:16:21 <andythenorth> can't see a commit style checker for nml
20:22:01 <frosch123> just assume there is nothing
20:23:56 <frosch123> i can't even imagine how many clicks it would need to configure checks for all gh repos
20:24:22 <frosch123> compared to a bash loop on coop server
20:24:36 <frosch123> and we didn't even have hooks on coop :p
20:28:43 <andythenorth> do any of these GH settings map to gitconfig?
20:30:13 <frosch123> afaik there are 3 ways to get a commit message into a gh repo
20:30:45 <frosch123> direct push, pr approve, and manual message in pr squash
20:31:11 <frosch123> we only know one check method that only works for path 2
20:31:22 <frosch123> so, i would say, gh is not meant for detailed checks
20:42:30 <andythenorth> I noticed some of my older nml commits are invalid
20:42:34 <andythenorth> (commit messages
20:43:14 <frosch123> there are many invalid ones
20:43:34 <andythenorth> no kittens died?
20:43:41 <frosch123> that's no excuse to completely abandon style though :p
20:51:52 <TrueBrain> pfft, and sneaking in 2 commits into one .. lazy .. so lazy ..
20:59:44 <glx> hmm DorpsGek_III speaks to much for nml
21:08:51 <DorpsGek_III> [OpenTTD/nml] glx22 opened pull request #49: Change: [DorpsGek] don't push push notifications to IRC https://git.io/JeBo3
21:14:20 <glx> hmm really slow drawing in these screenshot it seems
21:14:49 <glx> but train ticks are worse
21:15:56 <glx> signals on each tile maybe
21:16:46 <andythenorth> the weird news FPS bug?
21:17:04 <glx> no it really seems it's the trains
21:20:36 <LordAro> does look like a lot of signals
21:24:00 <DorpsGek_III> [OpenTTD/nml] andythenorth merged pull request #49: Change: [DorpsGek] don't push push notifications to IRC https://git.io/JeBo3
21:24:20 <andythenorth> only one way to test it :)
21:24:24 <andythenorth> and that was merge :P
21:25:10 <DorpsGek_III> [OpenTTD/nml] andythenorth commented on issue #45: White pixel warnings when compiling example_roadtype_and_tramtype.nml https://git.io/Je8vj
21:25:10 <DorpsGek_III> [OpenTTD/nml] andythenorth closed issue #45: White pixel warnings when compiling example_roadtype_and_tramtype.nml https://git.io/Je8vj
21:31:38 <nielsm> should the dev post also mention that one or two new bugs that crept into 1.9.3?
21:31:40 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #7780: Feature: Configure minimum share trading years https://git.io/JeBoE
21:31:47 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro merged pull request #7780: Feature: Configure minimum share trading years https://git.io/Je4de
21:32:43 <andythenorth> now it's on the main repo, just edit :)
21:33:09 <glx> andythenorth: it was already doable when in your repo ;)
21:33:56 <andythenorth> what you lot can edit in my repo?
21:33:58 <andythenorth> without forking?
21:34:20 <glx> unless you uncheck a box when creating the PR
21:34:46 <glx> that's why I could rebase your txt->md PR
21:34:51 <nielsm> there's the station sign position thing preventing you from making new orders
21:35:00 <nielsm> I think there was also one with AIs?
21:35:06 <glx> there's also the AI bug yes
21:36:07 <glx> and this one is worse as it crashes the game
21:38:32 <andythenorth> are the two I thought were important
21:39:16 <nielsm> those are not 1.9.3 bugs though
21:40:18 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on issue #7770: Current-location indicator in orders list missing https://git.io/Je8Z9
21:40:30 <andythenorth> oh yes I see, that makes sense
21:40:43 <glx> this one is still in 1.9.3
21:41:28 <glx> we forgot the mark the fix as "to be backported"
21:41:58 <andythenorth> what's the overhead on releases?
21:42:13 <nielsm> getting players to update
21:42:27 <andythenorth> we have work projects that can ship several times in an hour
21:42:28 <nielsm> multiplayer servers lagging behind splitting the player base
21:42:40 <andythenorth> others take 1-2 weeks, mostly due to customer comms
21:43:11 <LordAro> not sure what to do with the other order related issues
21:43:31 <LordAro> closing them as "no suggestions here, only issues" probably, but seems overly harsh :)
21:43:37 * LordAro lets andythenorth do it
21:43:48 <andythenorth> there are 96 issues
21:43:55 <andythenorth> which is about 20-30 too many imho
21:44:00 <andythenorth> I will close some later in the year
21:44:21 <andythenorth> right now I have the layout of ~all our docs in my head :P
21:44:23 <nielsm> if we had a way to release client-side-only patchlevel versions we could fix the little making-orders-to-stations-not-below-sign bug
21:44:24 <andythenorth> can't hold much else
21:46:13 <andythenorth> I have eints checked out so I can add words
21:47:27 <glx> don't forget to work from openttd branch
21:47:42 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #7767: Large numbers of trains and signals cause low FPS https://git.io/Je8mF
21:47:59 <glx> and target this branch if you open a PR
21:48:04 <andythenorth> there's only master in the eints repo now
21:48:23 <glx> no there's an openttd branch
21:48:41 <DorpsGek_III> [OpenTTD/OpenTTD] BlackPlague666 commented on issue #7767: Large numbers of trains and signals cause low FPS https://git.io/Je8mF
21:48:46 <glx> master is the generic eints
21:48:47 <LordAro> eints doesn't autodeploy, afaik
21:49:11 <andythenorth> 'when git branch doesn't list the branches'
21:50:46 <andythenorth> hmm, can't run eints, dropping that, too many things in the air :)
21:53:43 <frosch123> nielsm: the majority of players are singleplayer
21:53:55 <frosch123> so if you want to fix something, just 1.9.4, or wait a month
21:54:08 <glx> the majority of servers are empty ;)
21:58:14 <andythenorth> I would sooner force server updates than have a hard-crash bug in prod
21:58:24 <andythenorth> causes failure demand :P
21:59:18 <andythenorth> frosch123: is it known / rememberable / discoverable how to make windows nmlc.exe?
21:59:57 <milek7> i think it is unnecessary
22:00:32 <frosch123> andythenorth: i have been trying to make a proper docker for weeks now
22:00:39 <frosch123> i have no luck with wine
22:01:01 <frosch123> it would be little work, if it works
22:01:22 <frosch123> but i cannot even install msys in wine
22:01:36 <frosch123> the installer fails at 60% with whatever
22:01:44 <andythenorth> can we do it on Azure Pipelines?
22:01:55 <andythenorth> just get a Windows worker node?
22:02:07 <frosch123> probably, but i cannot test that locally
22:02:19 <andythenorth> and the cycle time on Azure is horrible
22:02:35 <glx> I just need the steps and I can try something
22:02:55 <andythenorth> can we at least put the instructions in nml repo?
22:02:59 <frosch123> glx: install msys2 and python3
22:03:05 <andythenorth> then whoever can be bothered :)
22:03:31 <frosch123> msys2 has a python3 package
22:03:40 <frosch123> when you have that and pip, you can pip pyinstaller
22:03:51 <DorpsGek_III> [OpenTTD/nml] andythenorth commented on issue #42: Document how to release nmlc, including building a windows binary https://git.io/Je8vQ
22:04:07 <glx> but if we can do it without requiring msys2 it's better for github/azure VMs
22:04:48 <glx> seems I have python 2.6.4 installed on my system
22:08:19 <LordAro> glx: clearly should test with all versions
22:08:40 <glx> that's doable with actions
22:09:02 <frosch123> oh, nml has cx_freeze specific code
22:09:03 <glx> many versions are available on the VMs
22:09:26 <frosch123> so, better to use cx_freeze directly
22:16:51 <frosch123> anyway, if cx_freeze is installed and knows the compiler and stuff, it's just a "cxfreeze nmlc", and then zipping the results
22:20:02 <DorpsGek_III> [OpenTTD/nml] andythenorth commented on issue #42: Document how to release nmlc, including building a windows binary https://git.io/Je8vQ
22:21:30 <glx> ok I have python3 installed
22:25:42 <andythenorth> cxfreeze nmlc gets me...something
22:25:48 * andythenorth reading the docs
22:27:01 <andythenorth> got some .so files
22:27:04 <andythenorth> shared objects? :P
22:27:11 <glx> of course "python .\setup.py install" fails as I don't run with admin rights
22:28:16 <andythenorth> I really dislike the common need to sudo setup.py :P
22:28:59 <frosch123> you can check the old bundles what files are needed
22:29:04 <LordAro> shouldn't need to install it, just bundle
22:29:10 <frosch123> iirc it always missed to copy some dlls
22:29:44 <glx> oh I should read the makefile
22:30:48 <glx> hmm and I don't really want to install
22:31:19 <andythenorth> I'm doing this in a virtualenv, but that's one extra complication :P
22:31:31 <andythenorth> avoids worries about installing crap though
22:32:11 <glx> and I don't see bundle rule in makefile
22:32:44 * andythenorth considers making a graphviz of local python
22:32:46 <frosch123> cxfreeze puts all stuff into a directory
22:32:59 <frosch123> or it used to do so :p
22:33:14 *** Thedarkb1-X40 has joined #openttd
22:33:34 <frosch123> contents should be the files listed in the paste above
22:33:39 <LordAro> glx: setup.py should have a bundle "rule"
22:33:54 <LordAro> you shouldn't need to install it directly while building it...
22:34:05 * andythenorth has 6 buildout pythons, 1 apple python, and a double-click installed python when buildout stopped working
22:35:46 <LordAro> a bundle rule, not the bundle rule :p
22:36:00 <LordAro> at least run --help-commands...
22:36:01 <frosch123> i guess the first one builds the c extension module
22:37:39 <andythenorth> I am getting a type error
22:37:55 <andythenorth> TypeError: unorderable types: NoneType() >= str()
22:38:11 <andythenorth> from python setup.py build -c mingw32
22:38:26 <andythenorth> this is sometimes an artefact of python 2 -> 3 comparisons
22:38:59 <LordAro> i can imagine that wouldn't work well if you're not currently on windows
22:39:25 <glx> ok "python.exe .\setup.py build" did something
22:39:33 <andythenorth> cx-freeze claims to be cross platform
22:40:04 <LordAro> cython's a bit nicer to use
22:40:09 <LordAro> but if you're already using it..
22:40:35 <frosch123> andythenorth: no, they always only build for the host system
22:40:42 <frosch123> that's why devzone used wine
22:41:25 <andythenorth> I could break out virtualbox :P
22:41:32 <andythenorth> it just got slower and slower recently :P
22:48:13 <LordAro> there we go, setup.py bdist creates a nice tar archive
22:48:30 * LordAro might have also rewritten setup.py in the process
22:49:31 <frosch123> is it complete though'?
22:50:06 <LordAro> well it didn't create an executable, because i was missing cx_freeze
22:50:20 <LordAro> but i don't think the setup.py was written in a way that would've done that anyway
22:51:15 <glx> I see a nmlc.exe in the zip
22:51:47 <frosch123> is there also a python.dll, and a nml_lz77.pyd ?
22:51:58 <LordAro> python.dll is always dynamically linked
22:52:24 <LordAro> windows dll path rubbishness means that it'll be copied with the extension
22:52:41 <frosch123> afaik the "pyd" are also all "dll"s
22:52:54 <frosch123> just following the python dlopen spec
22:53:04 <frosch123> or whatever "dlopen" is called on windows
22:56:01 <LordAro> ooh, that's a bit more interesting
22:56:52 * andythenorth wonders where to document nml releasing
22:58:45 <frosch123> LordAro: that's a lot of files
23:00:22 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on pull request #7782: Codechange: Improve NetworkAddress usage https://git.io/JeBK5
23:04:36 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #7782: Codechange: Improve NetworkAddress usage https://git.io/JeBKA
23:05:28 <frosch123> LordAro: nielsm: src/network/core is also used in the masterserver
23:05:45 <frosch123> if you change the api, please make sure that also the masterserver compiles
23:08:00 <glx> of course I need an admin powershell for pip
23:08:34 <LordAro> only to install globally
23:08:38 <LordAro> --user or whatever should be fine
23:08:50 <glx> I'm installing cx-Freeze :)
23:10:27 <glx> error: file 'C:\Users\Loïc\AppData\Local\Temp\pip-install-xanxs8q0\cx-Freeze\cxfreeze-postinstall' does not exist
23:10:27 <glx> ----------------------------------------
23:10:53 <glx> and how am I supposed to install a broken package ?
23:11:16 <LordAro> by googling the error message :p
23:11:25 <LordAro> indicates that cx_freeze 6 fixed the issue?
23:11:35 <andythenorth> standard python packaging :P
23:14:18 <glx> hmm but it is cx_freeze 6
23:14:35 <glx> maybe it's because I use python 3.8
23:17:55 <LordAro> compiles into a package on windows (probably), doesn't bother on Linux
23:18:15 <glx> ok using the alternate method it's installed
23:18:59 <LordAro> am happy to PR that, but someone should probably test it on Windows first..
23:21:52 * andythenorth 'mistakes were made'
23:22:34 <andythenorth> I have no idea how to update my local branch to match the remote PR
23:22:48 <andythenorth> so I dropped the commit, and cp the 2 remote commits :P
23:23:55 <andythenorth> I mean, it worked, but it's probably clown shoes
23:25:30 <glx> usually I git pull, git reset then git pull again
23:26:24 <andythenorth> I was suprised that strgen is still built
23:26:53 <glx> of course it is, the game needs it
23:27:19 * andythenorth is endlessly surprised
23:27:24 <andythenorth> every day is exciting
23:27:36 <glx> you know, all the "generating XXX language"
23:28:12 <andythenorth> ok so now I need to amend c21e0953694b4b0cc5d52ccddfb0cc1bbd611522
23:28:18 * andythenorth reading git docs
23:29:14 <glx> do the changes, git add/rm whatever, git commit --amend
23:29:26 <glx> then git rebase --continue
23:30:38 <glx> note that git mv already does add/rm IIRC
23:47:11 *** GroovyNoodle has joined #openttd
23:55:33 <glx> LordAro: your version seems to work
23:57:42 <frosch123> andythenorth: noai. and nogo.openttd.org are useful
23:57:56 <frosch123> but in 11 years i only remember one looknig at ottd doxygen
23:58:08 <andythenorth> I am not proposing killing it :)
23:58:16 <andythenorth> just whether it should be linked from github docs
23:58:29 <andythenorth> we have this scattered collection of dev docs :)
23:58:36 <andythenorth> project is quite confusing to approach :)
23:58:43 <glx> a link doesn't cost anything
23:58:53 <andythenorth> it will under new EU rules
23:59:00 <andythenorth> 0.02 Eurocents or something
23:59:03 <frosch123> glx: pip instal pillow
23:59:32 <frosch123> oh, wait, i always confuse ply and pil
23:59:36 <frosch123> ply is still called ply
continue to next day ⏵
