IRC logs for #openttd on OFTC at 2020-04-26
⏴ go to previous day
00:00:14 <glx> I think the rounding is "tile" based
00:00:17 <mcbanhas> But would it be affected by the replacement though?
00:00:34 <glx> but I didn't check the code
00:01:19 <glx> the idea is to make sure the new train still fit in same stations as old train
00:02:23 <FLHerne> Replacement only occurs when a train enters a depot
00:02:23 <DorpsGek_III> [OpenTTD/bananas-frontend-web] frosch123 opened pull request #28: Fix: Upload entirely new packages failed. https://git.io/Jft8S
00:02:52 <mcbanhas> "Remove wagons if the replacement makes the train longer. Wagons closest to the front are removed first."
00:03:00 <mcbanhas> "Remove wagons if vehicle replacement makes the train longer. Wagons closest to the front are removed first."
00:03:19 <mcbanhas> I prefer present tense to conditional
00:04:02 <FLHerne> I can't say I like that one much either
00:04:22 <frosch123> how do those sentences differ in tense?
00:04:29 <FLHerne> I think it has the correct meaning at least
00:04:36 <frosch123> i only spot "the" vs "vehicle"
00:04:37 <FLHerne> frosch123: Compared to my version, I think
00:04:52 <mcbanhas> These don't, but FLHerne's version was in the conditional
00:05:40 <FLHerne> I /think/ the reason I don't like this one is that I tend to think of "autoreplace" as the entire process
00:06:09 <mcbanhas> You gotta put this into context though, you'll be reading this in the autoreplace window
00:06:21 <mcbanhas> so you know that this is part of the autoreplace process
00:07:00 <mcbanhas> It's actually the reason why my first version had "automatically" as the first word.
00:07:01 <FLHerne> Just that "if vehicle replacement makes the train longer" reads like the opposite of what it does
00:08:30 <mcbanhas> Then yeah, I guess you need conditional for that
00:08:32 <FLHerne> I know you mean the removal *compensates for* that, but at a glance it seems conflicting
00:08:43 <mcbanhas> It would have to be "would make the train longer"
00:12:57 <FLHerne> Actually, the *goal* is to prevent the train becoming longer
00:13:12 <FLHerne> Removing wagons is just a means
00:13:37 <mcbanhas> That's the point of conditional tense.
00:13:46 <glx> and longer meaning overflowing in next tile
00:13:59 <FLHerne> "Prevent the train becoming longer by removing wagons if necessary. Wagons closest to the front are removed first."
00:14:21 <FLHerne> I mean, not burying the lede ^
00:14:33 <FLHerne> (although I don't like that wording now I read it)
00:16:12 <glx> basically, with default vehicles, single unit and 6 wagons can become multi unit and 6 wagons without having to remove any
00:16:52 <FLHerne> I don't think it's worth trying to make that distinction in the tooltip though
00:17:45 <FLHerne> Players only care about rounded-up-to-a-tile length, which is why that mechanic exists in the first place
00:18:11 <mcbanhas> Wait, but if you replace a single unit with a multi unit and it does not remove wagons, wouldn't that be problematic if i.e, your station is shorter than the new train length?
00:18:46 <FLHerne> mcbanhas: In glx's example it goes from 3.5 to 4
00:19:26 <FLHerne> i.e. it'll allow the train to become longer so long as the new length is still within the same tile length
00:19:30 <mcbanhas> Yeah, and in that case, tile length is the only thing that matters.
00:20:16 <FLHerne> I think in case of signal sections on angled track a fractional-tile change could make a difference
00:20:17 <glx> it's so it still fits in stations, signal blocks, whatever
00:20:27 <FLHerne> But clearly no-one cares in practice
00:20:46 <glx> trains are shorter in diagonals :)
00:21:09 <mcbanhas> Anyone sane has a signal every other tile, either way.
00:23:17 <FLHerne> That's a good way to go bankrupt with infrastructure maintenance on
00:23:32 <DorpsGek_III> [OpenTTD/bananas-frontend-web] frosch123 opened pull request #29: Fix: Make the navigation bar wider to fit all the categories, and mov… https://git.io/Jft4i
00:23:42 <glx> and used to be a pain for bridges
00:24:22 <mcbanhas> What about "Automatically removes wagons to prevent train from becoming longer. Wagons closest to the front are removed first."
00:26:22 <FLHerne> Not quite sure about 'automatically'
00:26:49 <FLHerne> "Removes wagons if needed to prevent the train from becoming longer. Wagons..."
00:27:12 <FLHerne> No, "needed" has conflicting meanings there
00:27:36 <FLHerne> The wagons are the opposite of needed :P
00:27:40 <mcbanhas> Yeah that's why I prefer the "automatically"
00:28:51 <mcbanhas> I think I'm going with this one
00:29:02 <FLHerne> What if I just get rid of 'otherwise' from my earlier one?
00:29:15 <FLHerne> "Remove wagons if the train would become longer. Wagons..."
00:29:37 <FLHerne> Or even with your 'automatically'
00:30:03 <FLHerne> "Automatically remove wagons if the train would become longer. Wagons closest to the front are removed first."
00:30:43 <FLHerne> Yes, I like that last one
00:30:55 <mcbanhas> I kinda fancy the present continuous in "becoming longer"
00:31:26 <mcbanhas> (That just sounded riské)
00:31:42 <FLHerne> It's linguistically nice, but I don't think it's easy to read
00:32:37 <FLHerne> "to prevent train from becoming" is a lot of grammar
00:33:00 <FLHerne> And I still find it strange without a conditional, even if the 'automatically' sort of implies it
00:33:08 <mcbanhas> "if the train would become" is just as much
00:34:32 <FLHerne> That's not true, you've got an extra verb :P
00:35:02 <mcbanhas> We can dumbify the sentence.
00:35:08 <mcbanhas> "...if the train gets longer"
00:35:53 <FLHerne> Same as my earlier objection
00:36:37 <FLHerne> "the train gets longer" refers to a hypothetical intermediate stage
00:37:42 <FLHerne> That doesn't actually happen, because the feature does the opposite
00:38:26 * FLHerne will run out of paint for this bikeshed soon :P
00:39:19 <mcbanhas> I still prefer the present. "If the train becomes longer"
00:39:42 <mcbanhas> mainly because conditional is rarely employed in tooltips
00:41:26 <FLHerne> Without the conditional, I really do find it confusing
00:41:43 <FLHerne> Possibly we need another opinion as a tiebreaker or something :P
00:45:55 <mcbanhas> present makes sense, because it is what happens if wagon removal is not on
00:46:18 <mcbanhas> I'm just gonna go with present at this point
00:48:22 <mcbanhas> FLHerne, would you like to test my new english.lng?
00:49:41 <FLHerne> "because it is what happens if wagon removal is not on" is exactly my objection ;-)
00:49:57 <FLHerne> mcbanhas: Yes, would be interesting to try it out
00:51:58 <mcbanhas> Ok, I should let you know I made some major changes to tooltips. A lot of things have been condensed or shortened. Capitalization is mostly changed to sentence-style. Tooltips are organized the following way: 1 - Title with colons, 2 - description, 3 - additional shortcuts ("Ctrl+click does X")
00:52:32 <FLHerne> mcbanhas: I've seen the thread and github issue
00:52:42 <FLHerne> (although not looked at them in the last few days)
00:53:14 <mcbanhas> I'll just throw the .lng in a file upload thingy. One moment.
00:53:27 <mcbanhas> Oh wait, this is IRC
00:55:54 <mcbanhas> FLHerne, are you using webchat or something similar? DCC is not working
00:55:58 <FLHerne> Huh, no-one's sent me anything over DCC in years
00:56:17 <FLHerne> I got a DCC request and accepted it
00:56:17 <mcbanhas> Brings back memories huh
00:56:37 *** andythenorth has joined #openttd
00:57:05 <mcbanhas> prolly a firewall issue
00:57:11 <FLHerne> Oh, might be that I'm behind this znc bouncer
00:57:23 <mcbanhas> i'll put it on mediafire
00:57:49 <FLHerne> Hm, apparently it ought to work
01:00:02 <mcbanhas> If you see any stuff you don't like or find weird, take notes or screenshots and send them to me or put it in the thread
01:03:31 * FLHerne might sleep now, it's gone midnight
01:03:40 <FLHerne> Thanks for your work on this
01:06:28 <mcbanhas> Sure, thanks for your help
01:08:08 *** arikover has joined #openttd
05:10:36 <DorpsGek_III> [OpenTTD/OpenTTD] James103 opened issue #8098: The transfer value in "Transfer: £[number] / Income: £[number]" can be negative. https://git.io/Jftgp
06:33:38 <Flygon> Well, this's annoying. Moved OpenTTD to my new computer. And whilst everything loads and simulates a lot smoother, the mouse cursor is weirdly choppy.
06:33:44 <Flygon> I wonder what's up. :D
07:52:04 *** andythenorth has joined #openttd
08:20:48 *** Progman has joined #openttd
08:44:02 <Samu> i've been wondering what would happen if the other list has items sorted in another manner
09:10:02 <andythenorth> does anyone know if it's 'NML' or 'nml'?
09:10:09 <andythenorth> or is it a pointless question?
09:17:19 *** sla_ro|master has joined #openttd
09:31:50 <Samu> "couldn't detect real instance of class for non-static call", in english plz
09:32:06 <Samu> this._path_tiles = clone this._prev._path_tiles;
09:46:48 <TrueBrain> andythenorth: consistency is for the weak :P :P
09:47:04 <andythenorth> hobgoblin of tiny minds
09:49:14 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain approved pull request #28: Fix: Upload entirely new packages failed. https://git.io/JftrJ
09:49:23 <DorpsGek> andythenorth: glx was last seen in #openttd 9 hours, 25 minutes, and 41 seconds ago: <glx> and used to be a pain for bridges
09:49:46 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain approved pull request #29: Fix: Make the navigation bar wider to fit all the categories, and mov… https://git.io/JftrI
09:52:42 <andythenorth> FLHerne LordAro ^
09:52:54 <andythenorth> not perfection, but more of a release note than a bad list of commits
09:54:13 *** frosch123 has joined #openttd
09:56:14 <DorpsGek_III> [OpenTTD/nml] andythenorth opened issue #105: Document procedures for switches and random switches https://git.io/Jftrs
09:58:04 <DorpsGek_III> [OpenTTD/bananas-frontend-web] frosch123 merged pull request #28: Fix: Upload entirely new packages failed. https://git.io/Jft8S
09:58:18 <DorpsGek_III> [OpenTTD/bananas-frontend-web] frosch123 merged pull request #29: Fix: Make the navigation bar wider to fit all the categories, and mov… https://git.io/Jft4i
10:01:55 * andythenorth will be glad when all this docs yak shaving is done :)
10:05:30 <frosch123> there are no procedures in random_switches, are there?
10:07:18 <andythenorth> does that work? o_O
10:07:19 <TrueBrain> hmm .. "upload update" link doesn't seem to work for me .. it starts a completely new upload, instead of an update .. hmm
10:07:36 <frosch123> andythenorth: caller vs callee :)
10:08:02 <frosch123> switches and random_switches can be callee, but only switch can be caller
10:08:26 <TrueBrain> owh, ofc, I need to upload something before it sees that .. lol .. forgot about that ..
10:08:31 <andythenorth> sorry, I hadn't even thought the other route would work :)
10:09:16 * andythenorth adjusts GH issue
10:09:19 <frosch123> TrueBrain: yes, you need to click "validate" once, after uploading files
10:09:40 <TrueBrain> so uploading new vs uploading update has no real difference :D
10:09:50 <frosch123> i thought about adding a second "validate" button after the upload stuff, that is labeled "autofill fields"
10:10:03 <frosch123> TrueBrain: no, it's the same until the api supports it :)
10:10:51 <andythenorth> I was thinking of something like 'Note that a random switch can also be called as a procedure by a switch; see [switch page]'
10:10:59 <andythenorth> worthwhile, or junk?
10:11:28 <frosch123> andythenorth: no idea, i assumed procedures would be documented where they are used, at the caller side
10:11:33 <TrueBrain> we need someone with Javascript knowledge .. I want to be able to drag/drop files :D
10:11:49 <frosch123> stackoverflow probably knows :)
10:12:23 <frosch123> hmm, my changes are still not deployed?
10:12:26 <andythenorth> javascript murders productivity
10:13:10 <TrueBrain> ah, I see where "name" goes wrong .. I wonder how that come to be .. hmm .. interesting
10:13:45 <TrueBrain> basically, it never supported you not filling in a name :D Lets see ..
10:14:05 <frosch123> i guess merging two prs in succession breaks it :) the first pr is deployed, and the second one is blocked since deploy is already running?
10:14:54 <TrueBrain> "An error occurred (ValidationError) when calling the UpdateStack operation: Stack:arn:aws:cloudformation:***:***:stack/Live-Staging-BananasFrontendWeb/cfbb7080-8161-11ea-a132-02b8022d5b40 is in UPDATE_IN_PROGRESS state and can not be updated."
10:15:03 <TrueBrain> as shown by the GitHub Actions result :)
10:15:29 <TrueBrain> yes; merging a second PR before the first is deployed will always return that
10:15:35 <TrueBrain> I considered it unlikely enoug hto not care :P
10:15:40 <TrueBrain> just his the "re-run jobs"
10:15:43 <frosch123> well, we'll find another thing to fix :)
10:16:24 <frosch123> ok, pressed, let's see
10:16:29 <TrueBrain> at first I considered waiting for the cloudformation to be done, but that means things can break if you merge 3 PRs shortly after each other
10:16:39 <TrueBrain> unlikely events being unlikely
10:20:02 <Samu> memory usage is absurdly high
10:21:12 <frosch123> yay, worked, no more broken navigator
10:21:15 <Samu> it's sad, really, the number of ticks are so good though
10:21:31 <frosch123> also less weird empty space at top :p
10:22:42 <TrueBrain> okay, this is odd .. I upload files, I hit validate, files gone
10:23:29 <frosch123> before validate the js adds the files, validate asks the api for present files
10:23:40 <TrueBrain> yeah .. and poef .. gone :)
10:23:43 <TrueBrain> this was working a moment ago
10:24:35 <TrueBrain> owh, it is the JS client caching stuff
10:24:43 <TrueBrain> "you already uploaded this file so I am not realllllyyyy going to"
10:24:53 <TrueBrain> we should figure out how to fix that, tbh
10:25:24 <frosch123> it "resumes" uploads :)
10:25:53 <frosch123> we need an api call to add files then
10:26:17 <frosch123> i know the file uuid, so i can tell the api to check for missing files
10:27:55 <TrueBrain> well, tusd, as we use it, doesn't really allow "resuming" after it restarted
10:28:01 <TrueBrain> but the client still things it is the same tusd
10:28:08 <Samu> which one do you think is less cpu intensive? this._path_tiles.extend(this._prev._path_tiles); or this._path_tiles = clone this._prev._path_tiles;
10:28:12 <TrueBrain> but it is also because the file remained on disk
10:28:19 <TrueBrain> so I guess this works out in production fine
10:28:37 <TrueBrain> the javascript client keeps things in the localstore
10:28:45 <TrueBrain> which allows it to check for resumes
10:29:19 <TrueBrain> this is a bit the problem with this whole implementation of tusd .. so few nobs to turn ..
10:31:15 <TrueBrain> hmm .. 7zip doesn't like the tarballs we create now ..
10:31:17 <TrueBrain> not sure what is going on
10:32:42 <TrueBrain> ah, default format changed
10:35:41 <TrueBrain> I think, at least for now, that is best
10:36:11 <TrueBrain> removeFingerprintOnSuccess to true should also be enough, possibly
10:36:19 <TrueBrain> but as the docs say, we should set that to true anyway :D
10:36:34 <TrueBrain> lovely, these settings they strongly advise you to change .. :P
10:38:39 *** mcbanhas has joined #openttd
10:39:46 <TrueBrain> okay, we were creating PAX tarballs .. not sure what OpenTTD would have done with those :P
10:39:51 <TrueBrain> we are back to GNU tarballs .. lot better :)
10:40:19 <Samu> looks like extend is less memory intensive than clone
10:43:04 <frosch123> ottd expects "ustar", but also supports something more ancient
10:43:29 <TrueBrain> ustar? That is also very old ..
10:43:49 <frosch123> /* Check if we have the new tar-format (ustar) or the old one (a lot of zeros after 'link' field) */
10:44:16 <frosch123> apparently it's the "new tar-format" :)
10:46:33 <DorpsGek_III> [OpenTTD/OpenTTD] James103 opened issue #8099: UI elements (incl. news message) do not resize properly, causing graphical glitches https://git.io/JftoL
10:50:03 <TrueBrain> right .. I was going to fix something else .. I was like: shall I make a ticket, I was like, no I can remember this, and now I forgot about it
10:52:23 <TrueBrain> that is what we call a stack-corruption
11:00:55 <TrueBrain> oops, should have been a new PR. Owh well
11:01:45 <TrueBrain> right, next up: reloading in a save way ..
11:02:08 <TrueBrain> frosch123: I have been thinking to add an API endpoint that shows all the things you have pending for commit, so people understand it takes a bit of time before it lands live
11:02:16 <TrueBrain> and possibly a button to publish
11:02:29 <TrueBrain> but that is maybe for the wishlist
11:03:21 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain opened issue #45: Add an API endpoint to show the pending changes https://git.io/Jfto2
11:03:30 <TrueBrain> and all this, for as you point out, for ~200 users :P
11:04:54 <frosch123> can i nitpick on commit messages? :p
11:05:13 <frosch123> As GitHub doesn't track folders, only files, we need to do some manual cleanup. <- git, not github
11:06:45 <frosch123> does os.walk survive, if you delete folders while iterating?
11:09:33 <TrueBrain> you for sure can nitpick, will fix
11:09:53 <TrueBrain> I tested it; os.walk() does survive .. didn't even do what I wanted him to do .. in result I have to walk a few times
11:10:00 <TrueBrain> seems it doesn't really return an iterator
11:10:35 <TrueBrain> so it is a bit odd, tbh
11:14:06 <TrueBrain> no, I guess it makes sense
11:14:11 <TrueBrain> to find the folders it already listed the parent
11:14:15 <TrueBrain> no need to relist it again
11:14:18 <TrueBrain> so yeah, this should be good
11:15:24 <TrueBrain> frosch123: a free extra commit, and fixed the commit message :)
11:21:37 <TrueBrain> frosch123: from the API, do you want the full license, or just the name?
11:22:43 <TrueBrain> you wanted to be able to fetch the licenses from the API
11:22:48 <TrueBrain> so they are no longer hardcoded
11:22:59 <TrueBrain> I wonder what the API should return
11:23:26 <frosch123> ah, a dict of (enum-value, user-readable-text, deprecated-bool)
11:24:03 <TrueBrain> owh, "user-readable-text" is just "GPL v2"?
11:24:13 <TrueBrain> (your example is still .. pretty open to interpertation :D)
11:25:22 <TrueBrain> you are still leaving a lot of dots for me to connect, but okay :)
11:25:34 <TrueBrain> for licenses, there is no difference between "enum" value and user-readable
11:26:07 <TrueBrain> so I guess: {"name": "GPLv2", "deprecated": False} is sufficient?
11:26:24 <TrueBrain> but that is after upload :)
11:26:46 <frosch123> TrueBrain: yes, as said in #23, "GNU General Public license, version 2" is too long to be useful
11:28:10 <frosch123> yeah, but pick a response format that fits all of frontend's #23
11:28:34 <TrueBrain> the "ai" -> "AI" mapping sounds like a frontend thing more than an API?
11:29:02 <TrueBrain> hmm .. I could make "content-type" dynamic, I guess
11:29:24 <frosch123> well, yes, the frontend won't create the nav header dynamically :)
11:29:48 <frosch123> but i do not want to duplicate the supported branches
11:30:03 <frosch123> TrueBrain: because there are some thoughts in the order they are displayed
11:30:12 <frosch123> it's not alphabetic
11:30:18 <frosch123> (or is it by accident? :)
11:30:39 <TrueBrain> in that case, lets not add it to the API just yet
11:31:13 <TrueBrain> I updated the swagger API with 3 endpoints, /config series
11:31:18 <TrueBrain> can you take a look before I implement them?
11:31:57 <frosch123> the term "audience" is still funny to me :)
11:32:32 <TrueBrain> but .. the only wiggle room I had while still following OAuth specs :)
11:32:38 <frosch123> will users understand the meaning of "master"?
11:35:46 <frosch123> want to add #42 to /config/user-audience ?
11:36:24 <TrueBrain> hmm .. it won't be dynamic
11:36:35 <TrueBrain> so I think it belongs in /user
11:36:55 <frosch123> yeah, there's no guarantee that it will be the same for all audiences
11:37:13 <frosch123> just for gh it's a the same url for all users
11:38:05 <TrueBrain> owh, it is based on the application-id, ofc
11:38:13 <TrueBrain> yeah, lets add it to config in that case
11:40:10 <frosch123> however, does the frontend know how the user logged in?
11:41:12 <frosch123> currently the front-end does not need to know what audiences are available
11:42:18 <TrueBrain> well, no, currently it is forced to one :)
11:42:27 <TrueBrain> the frontend tells the API which audience to use
11:42:33 <TrueBrain> so yes, the frontend does know that :)
11:43:47 <frosch123> i can display all the buttons "Login to xyz" then
11:44:10 <TrueBrain> that was the plan; a page when you click Manager (and you are not logged in) which shows all the buttons you can use to login
11:44:16 <TrueBrain> so no longer hard-code the frontend :)
11:45:06 <frosch123> i was so used to the developer login page of the api, that i thought the api would decide which flow to use :)
12:04:25 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain opened pull request #46: Add: /config endpoints with info about user-audiences, licenses and branches https://git.io/JftPf
12:04:39 <TrueBrain> frosch123: will you make the changes to the frontend, or do you want me to do so? :) (just to avoid us both doing it :P)
12:05:24 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #46: Add: /config endpoints with info about user-audiences, licenses and branches https://git.io/JftPf
12:09:05 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #46: Add: /config endpoints with info about user-audiences, licenses and branches https://git.io/JftPf
12:09:14 <TrueBrain> bye OpenTTD login flow .. you will never exist ..
12:12:17 *** ChanServ sets mode: +v tokai
12:53:03 <frosch123> TrueBrain: i would do it, unless you already started now :)
12:53:50 <andythenorth> so nml release or rewrite changelog? o_O
12:54:00 <frosch123> i like the changelog
12:54:17 <frosch123> no idea whether anything else is blocking the release
12:55:31 <TrueBrain> Go for it frosch123 ; I am enjoying a lunch :D
12:56:11 <andythenorth> nml changelog has requested changes, so can't be merged
12:57:17 <andythenorth> merge is blocked though
12:58:23 <DorpsGek_III> [OpenTTD/nml] frosch123 dismissed a review for pull request #104: Doc: Changelog for 0.5.0 release https://git.io/Jftkw
12:58:30 <andythenorth> looks like github bug
12:58:39 <andythenorth> or just hard problems :P
13:02:23 <andythenorth> we release nml from master?
13:02:33 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 commented on pull request #46: Add: /config endpoints with info about user-audiences, licenses and branches https://git.io/JftXA
13:02:35 <andythenorth> or instructions are wrong, and I should create a 0.5 branch?
13:02:42 * andythenorth looking at the past
13:03:00 <frosch123> noone is going to maintain a branch
13:03:12 <frosch123> nml 0.2 was branched because 0.3 was incompatible
13:03:35 <frosch123> so if anything, one would create a 0.4 branch from whenever we broke stuff
13:03:52 <frosch123> but then, noone is going to add anything to 0.4, so the tags are good enough
13:04:02 <andythenorth> I have tagged 0.5.0
13:05:58 <andythenorth> README.md is wrong in the published zip
13:06:10 <andythenorth> Last updated: 2019-04-02
13:06:10 <andythenorth> Release version: 0.4.5
13:06:38 <frosch123> ah, we removed that silly date from ottd :)
13:07:13 <frosch123> make a pr to remoe that line from master
13:07:18 <frosch123> don't care for 0.5 :)
13:08:18 <frosch123> blathijs: nml release is done
13:08:50 <DorpsGek_III> [OpenTTD/nml] andythenorth opened pull request #106: Doc: remove hard-coded release date and version number from README https://git.io/Jft1m
13:09:12 <DorpsGek_III> [OpenTTD/nml] frosch123 approved pull request #106: Doc: remove hard-coded release date and version number from README https://git.io/Jft1Y
13:09:20 <andythenorth> I just have to a few more docs then I can be free of nml for a bit :)
13:09:38 <andythenorth> great tool, easy to work on the code, PITA to maintain the docs :)
13:10:04 <andythenorth> nml is like a happy dog that occasionally brings home rotting dead birds
13:10:47 <frosch123> isn't that a cat thing?
13:11:00 <frosch123> or is the difference that cats bring fresh ones?
13:11:24 <andythenorth> maybe a dog that rolls in cow pats
13:11:37 <andythenorth> great dog, but sometimes you have to clean the house and it smells for days
13:22:08 <TrueBrain> gratz guys in releasing NML :)
13:25:16 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #46: Add: /config endpoints with info about user-audiences, licenses and branches https://git.io/JftPf
13:28:46 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #46: Add: /config endpoints with info about user-audiences, licenses and branches https://git.io/JftPf
13:28:50 <TrueBrain> did not know 'assert' was a keyword
13:29:53 *** gelignite has joined #openttd
13:30:51 <frosch123> assert all(license in LICENSES for license in License) <- haha, a great line to illustrate case sensitivity
13:31:07 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 approved pull request #46: Add: /config endpoints with info about user-audiences, licenses and branches https://git.io/Jft1Q
13:32:33 <TrueBrain> frosch123: I very much enjoyed writing that line, I have to say :)
13:36:31 * andythenorth todo: finish NML docs, do a forum release announcement
13:39:47 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain closed issue #42: API endpoint to get URL for OAuth access review https://git.io/JftWr
13:39:47 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain closed issue #32: API endpoints to get list of licenses and branches https://git.io/Jfktr
13:39:51 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain merged pull request #46: Add: /config endpoints with info about user-audiences, licenses and branches https://git.io/JftPf
13:42:38 <mcbanhas> Guys, anyone knows if it's possible to change text alignment via a config file or something similar? I can't find anything of the sort in the .lng file documentation.
13:43:00 <mcbanhas> I would like to experiment with left-aligned tooltips
13:43:41 <frosch123> no, you need to compile the source for that
13:43:56 <frosch123> it would also change all tooltips
13:44:07 <mcbanhas> It's for testing purposes
13:44:16 <andythenorth> it's witty that our tooltips are long enough to require alignment :)
13:44:18 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #8098: The transfer value in "Transfer: £[number]" can be negative. https://git.io/Jftgp
13:45:53 <frosch123> misc_gui.cpp:735 remove the SA_CENTER
13:47:37 <DorpsGek_III> [OpenTTD/nml] andythenorth closed issue #43: Release nml 0.5.x to provide NotRoadTypes support for nml authors https://git.io/Je8v7
13:48:36 * andythenorth needs to find the link to NML Windows binaries
13:48:56 <andythenorth> many of whom won't be able to pip
13:49:14 <mcbanhas> Alright thanks, I'll have a look.
13:49:58 <mcbanhas> Essentially, I want to see how modern tooltip formatting would look in-game.
13:50:07 <andythenorth> is that a safe place to link to?
13:50:17 <frosch123> i expected it on the release page
13:50:18 <andythenorth> safe / sustainable
13:50:29 <frosch123> maybe LordAro or TrueBrain know how that is supposed to work
13:50:51 <frosch123> does it need manual adding? or was it supposed to add stuff on its own?
13:50:56 <andythenorth> I'd have to upload the artefact manually to releases
13:51:06 <andythenorth> I'm surprised if that's the intended design
13:51:51 <TrueBrain> andythenorth: never link to an action directly; they expire afte a few days
13:52:21 <andythenorth> ok, so for today, download the artefact and attach to the release manually?
13:52:34 <TrueBrain> so "pip install nml" will work, also on Windows
13:52:50 <TrueBrain> you have to ask LordAro about anything else :D
13:53:31 <TrueBrain> just know Actions can also publish files in releases, if that is helpful in any way
13:54:04 <andythenorth> maybe that's intended, or it was just an oversight
13:54:13 <andythenorth> I've done it manually for today
13:54:29 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #8099: UI elements (incl. news message) do not resize properly, causing graphical glitches https://git.io/JftoL
13:54:52 <TrueBrain> most people also attach the whls to the release page btw
13:55:40 <TrueBrain> BaNaNaS now consumes the most amount of resources on AWS :P Also in terms of deployment size :D
13:55:47 <TrueBrain> we might have overdone this a bit 15 years ago :P
13:57:43 <andythenorth> ouch time for €1 / download TrueBrain
13:57:50 <andythenorth> or flat pass, monthly
13:58:56 <TrueBrain> to pay for my time, sure :P
14:00:19 <frosch123> andythenorth: do you take bets how many people will reply?
14:00:58 <andythenorth> do we take bets that turns into the general NML questions thread?
14:17:44 <planetmaker> :) nice work, guys :) tyvm
14:18:42 <heffer> Updates for nml on Fedora built an pushed. Now they get to ferment in updates-testing for a week.
14:21:08 <Samu> michi_cc, caching via AIList consumes too much RAM after all
14:21:20 <Samu> just discovered a few hours ago
14:22:32 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #8100: Inconsistent target framerate (30 vs 33 fps) https://git.io/JftD2
14:22:58 <Samu> I'm currently testing with a priority queue that is reversed. it prioritizes worst paths first
14:23:25 <Samu> quickly reaches 12 GB RAM in a minute
14:24:02 <DorpsGek_III> [OpenTTD/OpenTTD] James103 commented on issue #8100: Inconsistent target framerate (30 vs 33 fps) https://git.io/JftD2
14:26:42 <michi_cc> Samu: I made a PR anyway, the code still seems overly expensive to me no matter if it helps your current thing.
14:27:19 <michi_cc> Somebody would need to approve them though, otherwise they just sit there.
14:27:48 <Samu> oh snap, this crashed 1.10.1
14:29:16 <Samu> Unhandled exception at 0x00007FF66CF44C66 in openttd.exe: 0xC0000005: Access violation writing location 0x000000E2C63FFFF0. occurred
14:31:03 <glx> it's on the download page
14:31:21 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #8100: Inconsistent target framerate (30 vs 33 fps) https://git.io/JftD2
14:31:27 <Samu> > openttd.exe!SQInstance::`scalar deleting destructor'(unsigned int) C++
14:32:07 <Samu> the only thing listed in the call stack
14:32:49 <glx> probably the usual memory allocation issue
14:33:14 <Samu> I guess it was caused by prioritizing worst paths first :p
14:33:42 <Samu> it zig-zags paths the most it can, it's like playing snake, covering the whole map without repeating
14:34:05 <Samu> too many open nodes at once?
14:34:11 <glx> crazy idea for a pathfinder
14:34:31 <Samu> yeah, i was testing the extreme
14:34:36 <Samu> see how much ram it would use
14:38:21 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #8100: Inconsistent target framerate (30 vs 33 fps) https://git.io/JftD2
14:39:09 <Samu> d:\a\1\s\src\3rdparty\squirrel\squirrel\sqstate.cpp - says I dont have this file?
14:39:26 <Samu> well, not in that path, but...
14:39:52 <glx> your source is in 1.10.1 tag ?
14:40:17 <Samu> Unhandled exception thrown: write access violation.
14:40:18 <Samu> o._unVal.pTable->**** was 0xA28E8AFFD8. occurred
14:40:37 <Samu> openttd.exe!SQSharedState::MarkObject(SQObjectPtr & o, SQCollectable * * chain) Line 245
14:40:37 <Samu> at d:\a\1\s\src\3rdparty\squirrel\squirrel\sqstate.cpp(245)
14:40:43 <glx> you need to "git switch 1.10.1"
14:41:07 <glx> yes that's the typical memory fail
14:41:10 <Samu> this time it was an array
14:41:32 <glx> garbage colletction issue as always
14:41:58 <Samu> it crashed in like 30 seconds, seems that the reversed priority is good at picking crashes
14:42:53 <glx> well a solution probably exists but...
14:43:00 <Samu> gonna test one more time, now with the original method, which is... i dunno what to call it
14:46:24 <Samu> > openttd.exe!SQSharedState::MarkObject(SQObjectPtr & o, SQCollectable * * chain) Line 245 C++
14:47:32 <nielsm> glx: I'm stumped on that slowness issue, any ideas?
14:51:10 <Samu> i get the usual 0.98x-0.99x
14:52:45 <nielsm> it sounds like something wrong with the system timer on his system, I'm not sure what could be happening otherwise
14:53:03 *** WormnestAndroid has quit IRC
14:53:05 <Samu> cpu running at low clocks?
14:53:12 <Samu> let me try power save mode here
14:53:16 *** WormnestAndroid has joined #openttd
14:53:24 <Samu> if a bulldozer could save power...
14:59:45 <Samu> I thought intel was good
15:00:18 <TrueBrain> okay ... I think the server should now auto-reload on changes to the BaNaNaS GitHub ..
15:04:59 <TrueBrain> the 1.10 title screen really is a nice title screen; too bad of the maglev, but otherwise .. lot of things to spot and see :)
15:05:36 <glx> Samu: it was, now it's amd ryzen
15:07:58 <TrueBrain> okay, really happy with that :)
15:08:02 <TrueBrain> frosch123: what else was on the list of things to do?
15:09:07 <Samu> wow, it almost looks like the finished product
15:12:46 <frosch123> write a nice paragraph about how to get access to old logins?
15:13:13 <frosch123> from the issues, the most noticeable ones (since they worked with old bananas) are "upload zip/tar" and "download link"
15:16:09 <TrueBrain> besides the paragraph about old logins, is there anything blocking us from release now?
15:16:58 <TrueBrain> bah; CDK fixed something that broke my configuration .. this happens sadly more than I would like that to happen ..
15:17:19 <planetmaker> right, so how to login indeed? :D
15:17:38 <frosch123> TrueBrain: no, just some kind of announcement :)
15:17:57 <TrueBrain> frosch123: good; guess we need to let some people test and see if they can break it
15:18:07 <TrueBrain> I have some infra things to check out, likecurrently the installer will break if we switch over :P
15:18:28 <frosch123> TrueBrain: i wasn't able to test the blacklist. it looked like devauth allows everything
15:18:39 <TrueBrain> euh ... yes ... I never wrote that part :D
15:18:57 <TrueBrain> it reads the blacklist, it just never checks against it
15:19:17 <frosch123> in that case: i also could update packages owned by others
15:19:19 <TrueBrain> so my list: unbreak CDK, write paragraph, prepare installer CDN, fix blacklist issue, validate we are migrating everything correctly, add upload zip/tar :)
15:19:26 <TrueBrain> that should be prevented
15:19:31 <TrueBrain> there is code that should disallow that
15:19:40 <TrueBrain> (a blacklisted package doesn't really exist; that is where it is different)
15:20:08 <TrueBrain> session["errors"].append("You do not have permission to upload a new version for this package.")
15:20:37 <glx> TrueBrain: people will of course break it, the question is "will sentry catch the breaks ?" ;)
15:20:54 <TrueBrain> glx: sentry is picking bugs up nicely atm, but yeah :D
15:21:04 <TrueBrain> I put some effort in sentry reporting correctly, so I hope so :P
15:21:56 <Samu> if i put stuff in bananas staging, will it go to finalized product too?
15:22:23 <planetmaker> that is a name I haven't seen in a decade
15:22:25 *** Yexo is now known as Guest23408
15:22:31 <TrueBrain> Samu: absolutely not. It is staging. Data on staging never moves to production
15:24:48 <Yexo> Yes, got busy with work, life. Didn´t make the time to keep up
15:25:01 <Yexo> Good to OpenTTD is still very much alive
15:25:08 <frosch123> TrueBrain: ok, ownership check works. just blacklist then
15:25:47 <TrueBrain> hi Yexo :) "alive" might be a bit overstating, but surviving :P :P :P
15:26:32 <Yexo> ¨Not completely abandoned¨ at least
15:27:47 <planetmaker> yes, I saw that URL
15:27:54 <planetmaker> and referred to that, yes
15:28:24 <frosch123> there are two login links on the frontpage?
15:29:02 <planetmaker> it has... one? "login via github"
15:29:12 <planetmaker> and TOS. And "upload updates or new content"
15:29:14 <frosch123> "manager" and "upload ..." also works
15:29:33 <planetmaker> I loged in via github fine, yes
15:30:42 <planetmaker> So that's not something which contains the real packages, so most of it is empty, yes?
15:30:57 <frosch123> no, it's the sandbox
15:31:03 <milek7> probably intended, but logout doesn't quite works
15:31:07 <frosch123> for testing, people are supposed to upload shit :)
15:31:29 <planetmaker> yeah, logout is... interesting. I am always logged-in again when I click either link
15:31:32 <milek7> clicking manager logs immadiately with same account
15:31:36 <frosch123> milek7: it works, but as long as you are logged in to github, it won't ask you again
15:31:51 <TrueBrain> it is on the todo to add an additional page there :)
15:31:56 <frosch123> milek7: but we want to insert another confirmation page inbetween
15:31:59 <glx> same with most oauth providers
15:34:39 <frosch123> ohoh, i tried to add an old scenario, and it says "junk at end of file"
15:35:42 <TrueBrain> there are a few out there yes, I believe 4
15:35:51 <TrueBrain> which aren't "master" scenarios :)
15:35:55 <TrueBrain> so gratz on finding one of them :P
15:36:08 <frosch123> TrueBrain: no, i took one from my personal test folder
15:36:22 <TrueBrain> mcbanhas: if you are willing to put in the effort for that, sure
15:36:28 <TrueBrain> frosch123: awh .. :P
15:36:38 <nielsm> mcbanhas: first needs the video drivers to be replaced with something hardware accelerated
15:36:49 <frosch123> TrueBrain: i cannot reproduce your "file is gone after delete+reupload"
15:37:38 <TrueBrain> not sure how you pieced that together, but what I did was: I uploaded some files, killed the API, restarted the API, uploaded the files again
15:37:55 <mcbanhas> Samu, you know it's free nowadays right?
15:37:58 <TrueBrain> from that point on, they were "uploaded", just never added
15:38:21 <frosch123> ok, i did not kill the api
15:38:25 <frosch123> just deleted and added them
15:38:44 <TrueBrain> you should have asked for steps to reproduce :P I haven't even tried the delete buttons :D
15:39:11 <frosch123> but if you kill the api, the upload token is invalid, isn't it?
15:39:20 <TrueBrain> basically, it happens when the API did not clean up the files, so "tusd" sees them again, and the tus-js-client knows in the localstorage the hash it used
15:39:23 <TrueBrain> that combination breaks it
15:39:31 <TrueBrain> yes, so you start a new session
15:39:33 <TrueBrain> you upload the files
15:39:36 <TrueBrain> and they simply never arrive
15:39:46 <TrueBrain> tus doesn't check metadata to be equal
15:39:51 <TrueBrain> it simply looks at filenames
15:40:29 <frosch123> i can even start a new upload with new token, and it also works
15:40:36 <TrueBrain> and for some reason, on the tusd-side, the hooks are not being triggered
15:41:25 <TrueBrain> the situation you have to create: when you kill the API, "ls data/" should still show files
15:41:43 <TrueBrain> after that, starting the API, and uploading the same file, hit validate, file is no longer there
15:41:49 <frosch123> well, i'll file this under low-prio then :)
15:41:58 <TrueBrain> no, you have to add the setting I showed you in the js client
15:42:01 <TrueBrain> the one they say: SET THIS TO TRUE
15:42:19 <TrueBrain> the "removeFingerprintOnSuccess "
15:42:37 <TrueBrain> they themself say their default is bullshit
15:44:16 <frosch123> ok, i'll just enable it
15:44:38 <TrueBrain> tusd has some implementation issues, to say the least
15:44:54 <TrueBrain> it really should just send the hooks, and it would have been fine .. it doesn't :(
15:45:18 <TrueBrain> but it is also the piece of software that first says: I AM DONE! And after there: HERE IS A CHUNK OF DATA
15:45:23 <TrueBrain> because the callbacks .. are implemented wrong
15:48:42 <TrueBrain> btw, frosch123 , another way the above problem could happen is if someone logged out and logged in. But the latest API solves that by removing all files on log out too. Otherwise that was the easier way to reproduce :D
15:49:09 <glx> "finished ! PS: forgot to send the data, here they are"
15:55:38 <Samu> can I do something to improve the speed of method 2 or 0?
15:55:55 <Samu> by speed, i mean, less ticks needed
15:56:35 <Samu> I'm still torn on this :(
15:58:55 <Samu> is it possible for the "continue" inside the while loop or foreach loop to apply to the loop outside of it?
15:59:31 <Samu> that would avoid the bool, and break to continue
16:00:18 <Samu> oh, there's no continue inside, my bad
16:01:12 <Samu> i wanted to "double continue"
16:02:51 <Samu> there's no such thing as a break + continue in 1 command?
16:06:38 <Samu> while { while { do_stuff; break_continue; } } ?
16:08:03 <milek7> but considier regrouping into functions, and return from them
16:15:57 <frosch123> sorry for the confusion, but on forums i have no 123 :)
16:16:07 <frosch123> maybe add a "sorry for the inconvenience"
16:16:11 <frosch123> the creator also did that
16:17:06 <Samu> nop, no improvement, same number of ticks
16:17:23 <frosch123> it's a bit complicated, maybe it's easier with a pm/email content in a box, where the author needs to fill in some <bla>
16:17:35 <frosch123> s/content/template/
16:19:07 <FLHerne> mcbanhas: Pixel art at OTTD scale looks really bad if you smooth-scale it
16:20:02 <FLHerne> All the details are hinted at with a handful of pixels
16:21:48 <mcbanhas> FLHerne, not with this type of smooth scaling methinks.
16:23:01 <FLHerne> I've tried it, it does
16:23:19 <milek7> account migration cannot be done automatically?
16:23:23 <FLHerne> Might work if you have one of these really high-DPI screens
16:24:10 <mcbanhas> That's strange, because the sprite scale shouldn't be much different from what you see in RA1 units. Though the buildings and terrain are more hi-res, yes.
16:24:13 <FLHerne> But with a normal display, if you do anti-aliasing the details all get blurred, and if you don't it distorts the shapes and makes arbitrary thigs too big
16:24:40 <mcbanhas> Yeah, but I don't think they're doing AA here.
16:25:11 <mcbanhas> I'm not a technical guy, but they call it pixel doubling in the changelog.
16:25:52 <FLHerne> mcbanhas: Pixel doubling gets you 2x steps, which is what we already have :P
16:26:16 <FLHerne> There are algorithms to do it in a more heuristic way than simply doubling every pixel, like HQ2x
16:26:17 <mcbanhas> It doesn't look like AA in the game though.
16:26:54 <milek7> whether hq2x actually looks better is debatable
16:26:56 <FLHerne> But those are mostly tuned for 2D Nintendo-like games and work very badly on OTTD sprites
16:27:01 <mcbanhas> I'd say try it yourself. But it's just surreal how well it scales.
16:27:33 <frosch123> TrueBrain: looks good. i prefer option one a lot, so let's hope people pick that one :)
16:28:00 <TrueBrain> lets add that to the text
16:28:25 <milek7> that openra fractional scaling is in some patch?
16:28:34 <TrueBrain> "This content-type + unique-id is blacklisted."
16:28:49 *** Progman has joined #openttd
16:29:24 <milek7> started dune2000 and it works in 2x steps like in ttd
16:29:52 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain opened pull request #47: Fix: don't allow uploading blacklisted content https://git.io/Jft9d
16:30:01 <TrueBrain> frosch123: you HTML-ify that migration page?
16:30:43 <mcbanhas> milek7, you gotta download the latest beta
16:30:49 <mcbanhas> you probably got the stable
16:31:09 <FLHerne> mcbanhas: Hard to judge from a low-resolution gif
16:31:32 <mcbanhas> Alright, I stand corrected then :)
16:31:56 <mcbanhas> Also, did you have a chance to try the new .lng file?
16:32:21 <FLHerne> It looks like it's doing something a bit smarter than a plain bilinear filter though
16:32:30 * FLHerne reads the linked blog posts
16:32:54 <TrueBrain> LordAro: what continent did you move to?
16:33:39 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #47: Fix: don't allow uploading blacklisted content https://git.io/Jft9d
16:34:11 <FLHerne> (those should be in the opposite order, sorry)
16:34:22 <LordAro> andythenorth: frosch123: TrueBrain: nml windows binary - the pypi package is still just the standard python "module", the purely standalone executable is, as you've found, separate, and needs manually uploading *somewhere* (automatically attaching stuff to the GH release page was more difficult than it was worth). But still, if you've got python installed, you don't need the standalone executable
16:35:42 <mcbanhas> So they're doing some trickery then?
16:35:46 <LordAro> TrueBrain: assert indeed a keyword (like print used to be in py2), but you still need to make sure you're not using it for control flow, as (like you might expect) they get removed if you run python optimisations. I noticed a few of these in nml...
16:36:12 <TrueBrain> owh, he is backlog-reading .. this can take a while :P
16:36:39 <LordAro> Samu: d:\a\1\s is the path where the "official" binaries are built, so the debug symbols you get for those will point to that path. you can (obviously) ignore that bit of the path and just look at the bit relative to src/
16:37:04 <TrueBrain> this is pretty funny :P
16:37:08 <TrueBrain> liek a rerun of a movie
16:38:02 <LordAro> and scrollback complete!
16:38:14 <LordAro> TrueBrain: yeah, i saw that action, decided i couldn't be bothered
16:38:36 <LordAro> TrueBrain: i did a cycle, so had to reread the last 5 hours :p
16:38:36 <milek7> looks average, maybe on higher dpi it is passable
16:38:41 <TrueBrain> so you have the tools to fully automate things, and you give up just before you reached the endline
16:38:59 <Yexo> Glad to see you (and others) keep NML alive
16:39:17 <LordAro> i'm mostly surprised that the pypi release actually worked
16:39:30 <TrueBrain> but really happy you managed to do that :)
16:39:32 <LordAro> Yexo: feel like finishing station support? :p
16:39:38 * TrueBrain feels less dependent on now :D
16:40:14 <Yexo> LordAro: perhaps. Just getting a feel back for how stuff works
16:40:25 <TrueBrain> your left thumb moves like this
16:40:30 <TrueBrain> and your right .. well .. similar
16:40:31 <Yexo> OpenTTD code itself hasn´t changed too much at least
16:40:45 <LordAro> quick, someone merge cmake and change everything
16:41:14 <TrueBrain> and I am waiting for my cert to be issues by AWS ..
16:41:30 <TrueBrain> with a bit I mean .. it often doesn't happen before the timeout happens
16:43:09 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #8097: Codechange: [Script] Improve copying a list into another empty list. https://git.io/JftHJ
16:46:37 <planetmaker> :) Yeah, creation of NML was an awesome thing you pulled-off, Yexo :)
16:47:13 <planetmaker> it's incredible of how much boost that gave to newgrf creation
16:47:35 <planetmaker> certainly > 50% market share :)
16:48:36 <glx> and today we officially broke all existing industry related .nml ;)
16:48:56 <glx> (well 0.4.5 still exists for them)
16:49:09 <Yexo> Glad to hear that, but that wouldn have happened without maintenance/updates
16:49:15 <Yexo> 0.5 is a breaking change?
16:49:18 <planetmaker> NML did that before that old versions for old code
16:49:41 <glx> syntax change for more cargo support
16:52:08 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #8091: Add: [Script] Native priority queue; useful e.g. for pathfinders. https://git.io/JftHc
17:09:51 <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick commented on pull request #8097: Codechange: [Script] Improve copying a list into another empty list. https://git.io/JftHH
17:13:33 <DorpsGek_III> [OpenTTD/nml] LordAro merged pull request #106: Doc: remove hard-coded release date and version number from README https://git.io/Jft1m
17:16:39 <DorpsGek_III> [OpenTTD/OpenTTD] glx22 commented on pull request #8097: Codechange: [Script] Improve copying a list into another empty list. https://git.io/JftQf
17:21:18 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro approved pull request #8097: Codechange: [Script] Improve copying a list into another empty list. https://git.io/JftQm
17:23:05 <andythenorth> LordAro maybe I should update the nml readme release instructions? ("Upload the windows binary to github manually")
17:23:35 <LordAro> andythenorth: possibly
17:24:44 <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick commented on pull request #8091: Add: [Script] Native priority queue; useful e.g. for pathfinders. https://git.io/JftQs
17:25:40 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 approved pull request #47: Fix: don't allow uploading blacklisted content https://git.io/JftQZ
17:33:19 <DorpsGek_III> [OpenTTD/bananas-frontend-web] frosch123 dismissed a review for pull request #30: PR of the day https://git.io/JftQu
17:33:31 <TrueBrain> frosch123: deployment will fail, but I will fix that after :)
17:33:44 <frosch123> i fixed the "auto-fill" commit
17:33:56 <frosch123> the button enable/disable failed
17:34:24 <TrueBrain> I don't have a real way to judge these commits, so I am just approving them :P
17:34:34 <TrueBrain> "it doesn't look wrong" :P
17:34:50 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain merged pull request #47: Fix: don't allow uploading blacklisted content https://git.io/Jft9d
17:35:28 <LordAro> dbg: [net] injecting: 000acfdb; 48; ff; 000000; 00000000; 00000000; 00000036; "zul" (CmdRenamePresident)
17:35:33 <LordAro> this is no desync only zul
17:36:59 <glx> <andythenorth> LordAro maybe I should update the nml readme release instructions? ("Upload the windows binary to github manually") <-- but the upload part can be handled via github API calls IIRC
17:37:27 <andythenorth> I am about to PR some tweaks 1 min
17:38:24 <DorpsGek_III> [OpenTTD/nml] andythenorth opened pull request #107: Doc: more acccurate release process in README https://git.io/JftQP
17:38:58 <andythenorth> if we can automate the Windows binary it would be better, but it's not the end of the world
17:40:05 <DorpsGek_III> [OpenTTD/nml] andythenorth updated pull request #107: Doc: more acccurate release process in README https://git.io/JftQP
17:40:18 <andythenorth> glx be interesting to try
17:40:30 * andythenorth has hopes of restoring what Coop Bundles did, but with GH Actions
17:40:35 <andythenorth> build + publish grfs
17:41:40 <LordAro> glx: i did look at that at the time, but decided not to bother
17:41:53 <LordAro> there was some weirdness about whether it actually wanted to create the release object as well
17:43:46 <frosch123> TrueBrain: you can do your magic :)
17:48:01 <TrueBrain> waiting for it to fail deployen, before I can upload the code that makes it succeed :D Waiting waiting waiting .....
17:49:48 <TrueBrain> hmm .. I forgot we also have scenarios .. what to do with those .. I guess they are also for the installer, so lets do that
17:51:02 <TrueBrain> I was checking what files we host
17:51:08 <TrueBrain> and I found the "scenarios" folder again
17:51:11 <TrueBrain> totally forgot about that tbh
17:51:20 <frosch123> isn't that a thing of ottd < 0.4 ?
17:51:40 <TrueBrain> there is a 0.4.6, 0.4.7, 0.4.8, 0.5.0 version
17:51:53 <frosch123> glx: does windows installer stil offer to install scenarios?
17:51:54 <TrueBrain> no clue if anything ever uses it,tbh
17:52:10 <TrueBrain> it does not frosch123
17:53:21 <TrueBrain> lets check the logs if they are ever accessed ..
17:53:33 <TrueBrain> in the last 90 days, 0 times
17:55:07 <TrueBrain> frosch123: login looks a lot better now, awesome
17:56:00 <TrueBrain> okay, given nobody ever uses those files, I will let them be forgotten in the way they were 10 minutes ago :P
17:56:21 <Eddi|zuHause> archeologists in 100 years will discover them
17:57:08 <LordAro> dbg: [net] sync check: 000ad0c5; 00; mismatch expected {b65c131c, 005de115}, got {8549d8e1, 9077ce26}
18:01:22 <FLHerne> michi_cc: With the append-to-empty patch, what happens if the other list is modified afterward?
18:01:47 <michi_cc> FLHerne: Nothing. It's a copy.
18:01:51 <FLHerne> (can that happen? I'm not familiar with Squirrel's containers, just a thought)
18:02:08 <michi_cc> ScriptList is our container, not Squirrel's.
18:02:41 <michi_cc> The contents is entirely managed on the C++ side.
18:09:39 <TrueBrain> okay .. installer is prep'd on new infra, bananas I double-checked that we have all the URLs, and is running on staging correctly ..
18:09:45 <DorpsGek_III> [OpenTTD/OpenTTD] michicc merged pull request #8097: Codechange: [Script] Improve copying a list into another empty list. https://git.io/Jftl6
18:10:25 <TrueBrain> to me it sounds like tomorrow I can launch it in production on AWS, wire the whole shit up, and redirect traffic .. but I have to say: there are so many things going on with binaries .. it became a real maze over the last 15 years .. I hope I have everything in view :D
18:10:28 <TrueBrain> worst case .. something breaks :P
18:12:07 <TrueBrain> is there any reason not to push the new BaNaNaS live tomorrow? ( LordAro , nielsm , frosch123 , glx , who-ever-has-input-sorry-if-I-forgot-to-mention-you)
18:12:40 <DorpsGek_III> [OpenTTD/OpenTTD] michicc commented on pull request #8091: Add: [Script] Native priority queue; useful e.g. for pathfinders. https://git.io/Jft7g
18:12:55 <glx> probably the only way to really test it
18:13:14 <TrueBrain> and pretty sure there will be a screw-up somewhere or what-ever
18:16:03 <LordAro> can it be put back if everything falls over?
18:17:28 <TrueBrain> after that .. well, still yes, but in degraded capacity
18:17:54 <TrueBrain> personally I am more worried about integrity .. not sure anyone really tested if what they upload also shows up ingame
18:18:01 <TrueBrain> for me that is hard to judge, other than: it works :D
18:36:00 <LordAro> has anyone else tried to reproduce the desync in #8093 ? i've followed the docs in desync.md (not done it before), but no mismatch
18:36:10 <LordAro> (the one above was due to me starting the wrong savegame)
18:58:29 <frosch123> TrueBrain: i have a question left: when some dude sends some email to transfer accounts, how do i match the email to some account?
18:59:02 <frosch123> other than that, i have by bets on the first 3 complains, but no real blockers :)
18:59:37 <glx> usually the email doesn't match IIRC
19:00:46 <TrueBrain> frosch123: let me pick up the first of the emails, I guess .. but basically: login to the LDAP :)
19:00:52 <glx> and the user doesn't remember username or password either
19:03:26 * glx is looking at actions/upload-release-asset for nml, I think the easiest way is to change the workflow trigger
19:06:11 <frosch123> LordAro: what did you do exactly, if you replay everything, and get no mismatch, then that is good
19:09:29 <frosch123> TrueBrain: did someone already write a script to add users to the bananas data?
19:11:25 <LordAro> frosch123: i ran with the full log (nearly, had to remove the titlegame from the log), and then truncated it down until i was past the point where the client was disconnected due to desync
19:12:12 <frosch123> that means your replay matches the server, which is a good sign, as it means that you have enough data
19:12:25 <frosch123> next check when the desyncing player joined
19:12:38 <frosch123> and replay starting with the same just before/after the join
19:13:51 <frosch123> hmm, or did you mean that with "truncate"?
19:13:55 <frosch123> truncate from font?
19:17:21 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 opened issue #48: Validation warning when no readme.txt and/or changelog.txt was uploaded https://git.io/Jftdg
19:18:08 <frosch123> blathijs: i only delivered the message :) no work from me involved
19:18:48 <LordAro> frosch123: yeah, done that
19:18:56 <LordAro> truncated to the individual save: lines
19:19:12 <LordAro> "truncate from font" ?
19:19:22 <blathijs> LordAro: andythenorth: Congrats on the nml release, I'll be packaging it this week :-)
19:20:17 <LordAro> blathijs: still not really me, i just did the release process :p
19:20:29 <frosch123> LordAro: two options left: (1) replay from start and let it make a save at exactly the join tick. (2) compare the savegames with the savegame-diff tools, sometimes they differ without the desync being detected
19:21:12 <frosch123> after that i would claim the user was using a patched client
19:21:40 <LordAro> how can i force a save on join tick?
19:21:51 <LordAro> (and where are savegame diff tools?)
19:24:04 <frosch123> LordAro: for making a save: openttd.cpp:1364
19:24:34 <frosch123> aren't the diff tools linked from desync.md?
19:24:55 <frosch123> (they are really shitty, and probably should be rewritten in python using the new bananas savegame code)
19:26:30 <frosch123> LordAro: oh, option (3), the desync depends on the OS/compiler
19:28:03 <Samu> just experimented with walking backwards to check for conflicting bridges, ticks went from 307 to 374, it regressed :(
19:28:03 <frosch123> stuff like signed/unsigned enums
19:28:20 <Samu> the advantage is that it retains compatibility with graph.aystar-6
19:28:46 <LordAro> that's why my terminal didn't do anything
19:29:45 <frosch123> "ctrl+q" is a close second to "ESC q!"
19:34:16 *** snail_UES_ has joined #openttd
19:37:48 <blathijs> The Debian package checks that URL for new versions, but I can easily change that to check github instead, of course.
19:38:20 <blathijs> Also, the README still says "Release version: 0.4.5"
19:38:30 <Samu> michi_cc, have you worked on implementing graph aystar back in openttd? I was actually wondering what kind of performance there was to gain
19:38:54 <blathijs> Oh never mind the latter, git master has the 0.4.5 reference removed already I see.
19:39:08 <LordAro> blathijs: not sure about devzone, probably needs a planetmaker to fix
19:39:51 <blathijs> Samu: What do you mean by "back"? Was the A* pathfinder removed at some point?
19:40:17 <frosch123> blathijs: remove the devzone url :) it won't sustain
19:40:47 <Samu> reading the docs in devzone indicates it was in openttd before
19:41:09 <LordAro> probably related to the original AI
19:42:25 <Samu> it was an old version, version 1
19:42:30 <blathijs> Samu: I implemented the first A* pathfinder in OpenTTD many years ago for vehicles (called NPF for New Pathfinde, 2004 or something I think), and I think aystart might have been a generalization of that for use with AIs? I think TrueBrain originally built the Aystar AI implementation, maybe I also worked on that.
19:44:56 <Samu> make it version 6 compatible
19:45:41 <TrueBrain> the original AI did not have A* :P
19:46:09 <glx> the original AI was silly (but better than TT)
19:46:59 <TrueBrain> blathijs: indeed, something like that. I think in the end we both worked together to make a generalized version of A* in OpenTTD, but the scripting language never had direct access to that, as it would mean the game could stall for too long. So the Squirrel version is a near-copy of the OpenTTD variant (well, which is just A*)
19:47:12 <TrueBrain> it was always a very delibrate choice not to let AI use the built-in pathfinders
19:47:19 <TrueBrain> the AI would stall the game constantly
19:47:26 <TrueBrain> was considered very unacceptable ;)
19:47:58 <LordAro> that would be a good reason :p
19:48:06 <TrueBrain> I tried to tell that in the PR :)
19:48:14 <TrueBrain> you need additional ways to promise an AI doesn't kill performance
19:48:19 <TrueBrain> in Squirrel, they are bound to ticks
19:49:02 <TrueBrain> adding NoAI was already a bit of: wtf will performance do
19:49:10 <TrueBrain> so back then it was more an issue then it would be now
19:52:21 <Samu> I was imagining... in squirrel the pathfinder would provide the callback functions to graph.aystar, kinda like it does now, but a* would be governed by openttd, would that kill performance?
19:53:05 <mcbanhas> Guys how do I mark an article as a draft on the wiki? The {{Draft article}} tag doesn't seem to be working on this version of mediawiki.
19:53:43 <Yexo> Currently the AI can be stopped at any point during pathfinding by the squirrel interpreter. However it is not allowed to stop a script while in a callback from native code
19:53:43 <LordAro> "{{Draft article}}" sounds like something that's been added by whichever mediawiki instance you were using
19:53:55 <LordAro> but the wiki is very very old
19:54:14 <Yexo> Which means that the native pathfinder code would have to include code that can stop pause your AI until the next tick
19:54:28 <mcbanhas> So can I just save a page that's incomplete?
19:57:12 <andythenorth> mcbanhas just stick DRAFT somewhere near the top
19:57:22 <andythenorth> if no better way is found
19:59:21 <LordAro> frosch123: ok, so the desync disconnect occurred at 000ad0ef. I compared the dumps of 000ad100 from save that duck provided, and the one that i just generated - no differences
19:59:56 <LordAro> wait hang on, i need to start the game from 000ad053
20:00:22 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain opened pull request #49: Minor fixes and zip/tar/tar.gz support https://git.io/JftF4
20:00:25 <TrueBrain> there you go frosch123 , specially for you :)
20:01:04 <TrueBrain> you need to press auto-fill after uploading for frontend-web for now
20:01:31 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #49: Minor fixes and zip/tar/tar.gz support https://git.io/JftF4
20:01:32 <TrueBrain> and the mandatory: did you forget black again?!
20:01:44 <TrueBrain> so happy with GitHub Actions :D
20:04:03 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #49: Minor fixes and zip/tar/tar.gz support https://git.io/JftF4
20:08:05 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain opened pull request #31: Fix: minor tuning on the migration text https://git.io/JftFz
20:08:19 <TrueBrain> where did I type that ...
20:08:44 <frosch123> last commit message
20:09:02 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #49: Minor fixes and zip/tar/tar.gz support https://git.io/JftF4
20:09:42 <TrueBrain> it is funny how the interface of TarInfo and ZipInfo are JUST not the same
20:09:58 <LordAro> TrueBrain: there's a lot of the python stdlib like that :p
20:10:01 <mcbanhas> Just so you know what I've been up to. It still neds a lot of work and all the examples still need to be put in.
20:10:28 <mcbanhas> But hopefully it will see a lot more polish in the next few days.
20:10:44 <LordAro> frosch123: i'm getting the impression that saving the game on a specific tick value (date_fract) doesn't work so well - the desync stuff definitely appears to be assuming that i'm starting on 0 (rather than 9)
20:10:54 <mcbanhas> I know it's just a document, but hopefully having this will make proofreading for UI texts much easier.
20:18:06 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 commented on pull request #49: Minor fixes and zip/tar/tar.gz support https://git.io/JftF6
20:19:12 <frosch123> LordAro: why would it care about the tick?
20:19:40 <LordAro> frosch123: i'm not sure, but commands.log is getting skipped
20:20:07 <LordAro> i'm just reversing 9 ticks anyway, nothing else apparently happened in between
20:21:07 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain commented on pull request #49: Minor fixes and zip/tar/tar.gz support https://git.io/JftFX
20:21:42 <frosch123> LordAro: assuming the desync is caused by pathfinder caches, the commands do not matter
20:21:53 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #49: Minor fixes and zip/tar/tar.gz support https://git.io/JftF4
20:23:44 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 approved pull request #49: Minor fixes and zip/tar/tar.gz support https://git.io/JftFS
20:27:07 *** mcbanhas_ has joined #openttd
20:29:33 *** gelignite_ has joined #openttd
20:35:10 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain merged pull request #49: Minor fixes and zip/tar/tar.gz support https://git.io/JftF4
20:45:32 <LordAro> dbg: [net] sync check: 000ad0ef; 00; mismatch expected {33188c8c, 1d7b69f5}, got {007a3060, 2361519c}
20:45:39 <LordAro> and i think this is genuine this time :)
20:46:22 <LordAro> no idea what it means, but...
20:49:15 <LordAro> and yeah, got a difference in savegame dumps at ad0e0
20:51:16 <Samu> are you hunting desyncs?
20:54:28 <frosch123> what chunk? vehicles?
20:56:22 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain updated pull request #31: Fix: minor tuning on the migration text https://git.io/JftFz
20:56:25 <TrueBrain> frosch123: feel free to improve the wording of both commits :)
20:56:50 <LordAro> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20:56:54 <LordAro> wait, these are identical
20:58:44 <frosch123> map array is unusual for a desync
21:05:12 <LordAro> i'll write up what i've got in the bug
21:05:20 <LordAro> but feel free to recheck my work :p
21:10:35 <DorpsGek_III> [OpenTTD/bananas-frontend-web] frosch123 approved pull request #31: Fix: minor tuning on the migration text https://git.io/JftNL
21:14:26 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain merged pull request #31: Fix: minor tuning on the migration text https://git.io/JftFz
21:15:30 <TrueBrain> Except for a web download url it seems to at least do what the current version does
21:16:23 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on issue #8093: Regular server desynchronisations from clients https://git.io/JfkmV
21:25:54 <TrueBrain> you broke something I am sure I fixed already :P Let see
21:26:28 <TrueBrain> hmm .. you pressed delete on something :)
21:27:29 <TrueBrain> I did not check removing of files, clearly :D
21:29:00 <mcbanhas_> https://i.imgur.com/T3vhgZA.png frosch123 if I take down the SA_CENTER parameter like you told me, it unfortunately does not compensate for the top or right margins if the text is small. Which values do I have to tweak to make this possible?
21:30:30 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain opened pull request #50: Fix: there is no .info file for extracted files https://git.io/JftAv
21:30:40 <TrueBrain> frosch123: it is a silly mistake :D
21:32:31 *** mcbanhas has joined #openttd
21:34:21 <mcbanhas> These are the tooltips from CK2. Tooltips are left aligned in this game, but always reserve a margin for smaller instances to appear centered.
21:37:01 <frosch123> mcbanhas: use SA_VERT_CENTER then
21:38:45 <LordAro> (or SA_HOR_CENTER :p )
21:39:21 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 commented on pull request #50: Fix: there is no .info file for extracted files https://git.io/JftAC
21:42:08 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #50: Fix: there is no .info file for extracted files https://git.io/JftAv
21:42:10 <TrueBrain> frosch123: during testing that never happened, but it can't hurt to prevent it :D
21:45:27 <DorpsGek_III> [OpenTTD/nml] LordAro approved pull request #102: Fix #100: follow PEP440 when generating version https://git.io/JftAV
21:49:35 <DorpsGek_III> [OpenTTD/nml] LordAro approved pull request #107: Doc: more acccurate release process in README https://git.io/JftAD
21:50:41 *** gnu_jj_ has joined #openttd
21:51:19 *** gnu_jj__ has joined #openttd
21:55:33 <DorpsGek_III> [OpenTTD/OpenTTD] JGRennison commented on issue #8093: Regular server desynchronisations from clients https://git.io/JfkmV
21:59:25 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 approved pull request #50: Fix: there is no .info file for extracted files https://git.io/JftxT
22:01:11 <mcbanhas> https://i.imgur.com/2Telfwz.png frosch123 this still doesn't fix it entirely. The box is essentially still behaving as if the text was still centered and leaves a way too big margin on the right. The max it should ideally go should be the length of the longest line.
22:02:33 <mcbanhas> You can see also on the small tooltip, the right margin is a bit bigger.
22:02:33 <frosch123> mcbanhas: ottd's layout capabilities are very limited
22:02:42 <frosch123> it firs decides for a width of the box, then breaks the lines
22:03:54 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain merged pull request #50: Fix: there is no .info file for extracted files https://git.io/JftAv
22:04:19 <mcbanhas> I don't suppose that has an easy fix.
22:04:56 <mcbanhas> Would it require a serious overhaul or anything of the kind?
22:05:12 <frosch123> LordAro: aw, i was betting for ships/rv pathfinder caches, but it looks like it's trains again :/
22:05:18 <glx> but not very easy to test
22:06:00 <LordAro> glx: that's why there's a nml-test package on test.pypi.org :p
22:06:24 <LordAro> actually, i can give you a token for it, if you like
22:06:34 <LordAro> you can add the secret to your repo
22:06:55 <glx> means I need to merge it in my repo to trigger a release
22:07:34 <LordAro> or make a separate branch that also removes the master-only release behaviour
22:07:59 <glx> the trigger is now on release, not push
22:08:23 <LordAro> well, you can change that :p
22:08:42 <glx> but I need it to easily get the upload url
22:08:45 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 closed issue #22: Allow uploading .zip/.tar instead of individual files https://git.io/Jfktf
22:10:39 <mcbanhas> Wait a moment, can I adjust the minimum margin manually? I see it is 1px on the left side. Is this controlled by line 730?
22:11:31 <TrueBrain> glx : why zip the executable? That feels weird :p
22:11:49 <frosch123> mcbanhas: sure, it's jsut code
22:12:15 <glx> TrueBrain: so it's always nmlc.exe once extracted
22:12:30 <mcbanhas> not even a pro-gamer
22:12:56 <TrueBrain> Feels weird .. maybe add license and readme in zip too?
22:13:31 <TrueBrain> Otherwise it looks nice :)
22:13:58 <glx> yeah just hard to test as always with github actions ;)
22:17:37 <TrueBrain> Especially if you start using non-git-hooks indeed
22:17:50 <TrueBrain> There are simulation tools for it .. never tried them
22:18:12 <TrueBrain> I try to stay within the git domain :D
22:19:33 <andythenorth> oof it's nearly Monday :|
22:19:39 <andythenorth> usually I am fine with Monday
22:20:36 * andythenorth has not had a day off since Jan 2nd, except public holidays
22:23:13 <frosch123> TrueBrain: i guess i can break it again, be deleting the ".tar". but i guess we do not care
22:28:48 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 opened issue #51: Uploading an update discards all meta data except "version" https://git.io/Jftxb
22:33:50 <TrueBrain> frosch123 : define break? It should give a 400 on the API
22:34:06 <andythenorth> hmm NRT bug report in forums :|
22:34:17 <andythenorth> only 15 types available
22:34:54 <andythenorth> oof, docs say more
22:35:09 * andythenorth tactically ignores this for today :D
22:38:13 <frosch123> TrueBrain: ok, all good then :)
22:40:02 <TrueBrain> It is the front end that is using obsolete info :p
22:43:24 <TrueBrain> hmm .. frosch123 .. can a version have another name than the main package?
22:43:37 <TrueBrain> feels a bit weird :P
22:43:42 <andythenorth> wait wait wait wait wtf
22:43:46 <TrueBrain> by spec, you currently should be able
22:43:48 <frosch123> we allowed editing it :)
22:43:49 <andythenorth> yexo just commented in GH
22:44:12 <DorpsGek> andythenorth: yexo was last seen in #openttd 2 hours, 49 minutes, and 58 seconds ago: <Yexo> Which means that the native pathfinder code would have to include code that can stop pause your AI until the next tick
22:44:33 <frosch123> andythenorth: you missed something :p
22:45:05 <andythenorth> what next Dalestan?
22:45:19 <frosch123> dalestan was spotted on factorio forums
22:45:31 <frosch123> he explained to planetmaker how to play the game properly, or something
22:45:58 <DorpsGek> LordAro: bjarni was last seen in #openttd 8 years, 29 weeks, 1 day, 20 hours, 26 minutes, and 51 seconds ago: <Bjarni> heh
22:46:15 <DorpsGek> TrueBrain: Celestar was last seen in #openttd 2 years, 15 weeks, 4 days, 7 hours, 50 minutes, and 1 second ago: <Celestar> lo
22:46:22 <DorpsGek> TrueBrain: I have not seen MiHaMiX.
22:46:30 <DorpsGek> frosch123: I have not seen mihamek.
22:46:38 <frosch123> how many other spellings?
22:46:56 <DorpsGek> TrueBrain: ludde was last seen in #openttd 7 years, 41 weeks, 4 days, 13 hours, 43 minutes, and 34 seconds ago: <ludde> hej
22:47:10 <LordAro> how far back does the DorpsGek history go?
22:48:03 <TrueBrain> at least 8 years and 29 weeks :P
22:48:15 <FLHerne> According to recent empirical evidence, at least 8 years, 29 weeks, 1 day, 20 hours, 29 minutes and a few seconds :P
22:48:48 <TrueBrain> LordAro: #openttd.20070403.log
22:49:08 <TrueBrain> who is a good bot? DorpsGek is a good bot! WHO WANTS A COOKIE?! DorpsGek should visit www.amazon.com
22:49:27 <DorpsGek> frosch123: I have not seen MiHaMeK.
22:49:35 <frosch123> lies, he has been here in 2008
22:51:13 <frosch123> LordAro: those logs start 2006-02-25
22:51:47 <LordAro> it parses logs to see when last seen?
22:51:51 <LordAro> that seems... inefficient
22:52:19 <frosch123> no idea, but i do it like that :p
22:52:53 <TrueBrain> it did record MiHaMeK in another channel ..
22:53:23 <DorpsGek_III> [OpenTTD/OpenTTD] JGRennison commented on issue #8093: Regular server desynchronisations from clients https://git.io/JfkmV
22:54:13 <TrueBrain> weird, there is an entry of him in the database
22:54:20 <DorpsGek> TrueBrain: I have not seen MiHaMeK.
22:54:27 <TrueBrain> I .. am looking .. at the entry ...
22:54:33 <DorpsGek> frosch123: I haven't seen anyone matching miha*.
22:54:43 <TrueBrain> #openttd,MiHaMiX,1199645917.23,<MiHaMiX> hmm
22:54:50 <DorpsGek> frosch123: miho was last seen in #openttd 5 years, 25 weeks, 0 days, 23 hours, 21 minutes, and 5 seconds ago: <miho> yeah, I'll give it try
22:55:12 <DorpsGek> frosch123: WHamilton was last seen in #openttd 1 year, 39 weeks, 0 days, 18 hours, 15 minutes, and 55 seconds ago: <WHamilton> forgot i closed it earlier, thanks
22:55:31 <frosch123> well, now i feel like samu
22:56:49 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain opened pull request #52: Fix #51: set the name / description / url / tags when uploading a new version https://git.io/Jftpg
22:57:24 <TrueBrain> I am happy so far the bugs you keep finding are easy
22:59:39 <frosch123> is there a difference in raw_data and session for new packages?
22:59:52 <TrueBrain> raw_data is the processed session info
22:59:56 <TrueBrain> so it went through the validator
23:00:05 <TrueBrain> just to catch any remaining glitches
23:01:36 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 approved pull request #52: Fix #51: set the name / description / url / tags when uploading a new version https://git.io/Jftpy
23:02:45 <frosch123> so, news post tomorrow :)
23:04:38 <frosch123> LordAro: wasn't there some samu/ai discussion about attaching wagons to trains in depots? was anything changed?
23:05:00 <LordAro> frosch123: i don't believe i ended up changing anything, i just changed documentation
23:05:15 <nielsm> could it be related to auto-refit on purchase?
23:08:44 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain merged pull request #52: Fix #51: set the name / description / url / tags when uploading a new version https://git.io/Jftpg
23:08:45 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain closed issue #51: Uploading an update discards all meta data except "version" https://git.io/Jftxb
23:15:12 <frosch123> nielsm: the GroupStatistics stuff in the build+refit looks fishy for sure
23:15:48 <frosch123> the OrderBackup in there may even cause a desync
23:15:55 <Samu> what are we talking about? I was mentioned
23:16:14 *** snail_UES_ has joined #openttd
23:16:59 *** snail_UES_ has joined #openttd
23:17:37 <LordAro> nielsm: frosch123: put something in the issue, to make sure you're not overlapping with JGR
23:20:04 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on issue #8093: Regular server desynchronisations from clients https://git.io/JfkmV
23:20:24 <DorpsGek_III> [OpenTTD/OpenTTD] frosch123 commented on issue #8093: Regular server desynchronisations from clients https://git.io/JfkmV
23:31:35 <andythenorth> I suspect those clown shoes are mine
23:33:36 <TrueBrain> counting from 0, I assume? :D
23:33:50 <TrueBrain> so that is right; 16, 0 .. 15 :)
23:34:17 <andythenorth> "Roadtypes and Tramtypes share a pool of 64 IDs. Either feature can use as many IDs as needed (i.e. it is not limited to 32 per type), but both features combined are limited to 64 in total"
23:36:57 <frosch123> andythenorth: looks like 64 is also a lie
23:37:44 <frosch123> i guess one id is used for "no road/tram" present
23:41:00 <andythenorth> nfo and nml wiki pages are wrong then :)
23:41:03 * andythenorth will fix, not now
continue to next day ⏵