IRC logs for #openttd on OFTC at 2020-10-31
⏴ go to previous day
00:59:43 *** WormnestAndroid has quit IRC
01:00:24 *** WormnestAndroid has joined #openttd
01:43:02 *** Speeder__ has joined #openttd
01:58:59 *** ChanServ sets mode: +v tokai
02:04:36 *** WormnestAndroid has quit IRC
02:05:30 *** WormnestAndroid has joined #openttd
07:38:44 *** sla_ro|master has joined #openttd
08:03:46 *** longtomjr has joined #openttd
08:21:25 *** sla_ro|master has joined #openttd
09:11:15 *** WormnestAndroid has quit IRC
09:11:27 *** WormnestAndroid has joined #openttd
09:19:40 *** HerzogDeXtEr has joined #openttd
09:49:26 *** Gustavo6046 has joined #openttd
10:03:38 *** andythenorth has joined #openttd
10:11:01 *** jottyfan has joined #openttd
10:34:59 *** Progman has joined #openttd
11:11:28 <andythenorth> so 1 week to draw a ship
11:11:39 <andythenorth> whereas I can draw 2 trains per day
11:12:44 <andythenorth> Sam has 42 ships I need to draw
11:13:13 <andythenorth> whereas Horse has 500 trains I've drawn
11:13:18 <andythenorth> which takes longer? o_O
11:23:20 <longtomjr> Dunno sounds complicated
13:07:20 *** frosch123 has joined #openttd
13:24:21 *** WormnestAndroid has quit IRC
13:26:33 *** WormnestAndroid has joined #openttd
13:48:59 <TrueBrain> Cannot connect to host github.com:443 ssl:True [SSLCertVerificationError: (1, "[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: IP address mismatch, certificate is not valid for '140.82.118.3'.
13:51:27 <TrueBrain> this is the exact same code as running for bananas-api :P
13:58:45 <LordAro> github.com has several IPs, one of them got an incorrect certificate?
13:58:59 <TrueBrain> I do not think it is github.com :)
13:59:04 <TrueBrain> bananas-api works, with the same code
14:00:16 <TrueBrain> it works with aiohttp 3.6.2, not with aiohttp 3.7.0
14:00:18 *** Speedy` has joined #openttd
14:00:23 <TrueBrain> so it would have broken bananas-api too in 2 days :P
14:01:02 <TrueBrain> owh, no, now it works ...
14:01:14 <TrueBrain> rriiiigggghhhtttt .... walking away now ....
14:11:08 <LordAro> to me, that error message would suggest github.com being at fault
14:11:36 <LordAro> or possibly some local certificate issue
14:12:12 <frosch123> or the secret service intercepting stuff
14:12:42 <TrueBrain> LordAro: I normally assume I am at fault in these cases :)
14:12:50 <TrueBrain> so possibly my cert-store was not up-to-date or what-ever
14:13:00 <TrueBrain> it works now .. shrug
14:22:51 <TrueBrain> okay, so you can login now, even with multiple backends .. sweet :)
14:28:34 <TrueBrain> ah .. it was aiohttp 3.7.0 after all
14:28:40 <TrueBrain> I just now installed 3.7.1, didn't notice that
14:28:43 <TrueBrain> which fixes the issue:
14:28:47 <TrueBrain> "Fix a variable-shadowing bug causing ThreadedResolver.resolve to return the resolved IP as the hostname in each record, which prevented validation of HTTPS connections."
14:28:54 <TrueBrain> so it wasn't github.com LordAro :D
15:43:05 <TrueBrain> and I can edit pages ... that was easy :P
15:43:10 <TrueBrain> now for all the weird edge-cases ....
15:43:49 *** Wormnest has joined #openttd
15:43:50 <TrueBrain> creating new pages even works without effort :D I love those bonuses
15:43:57 <FLHerne> Is is documented anywhere that the `GRFID` variable is available in NML?
15:44:16 <FLHerne> It seems to be an implementation detail, and I can't find any mention of it
15:44:34 <FLHerne> But it would be annoying if it went away and someone complained
16:05:58 <andythenorth> but not as a named var
16:09:51 <TrueBrain> but otherwise you can edit pages, create pages, etc
16:10:16 <TrueBrain> there is also a preview
16:10:23 <TrueBrain> just not the way we want it to be in the end
16:10:26 <TrueBrain> but that is Javascript :P
16:11:11 <TrueBrain> code-wise it is a bit of a mess, but I knew that already :)
16:11:20 <TrueBrain> I have to rewrite where code lives .. deduplicate some other code ..
16:12:26 <frosch123> ah, the oauth redirect goes to localhost :)
16:12:44 <TrueBrain> well, fix the URL and you will be fine :P
16:12:56 <TrueBrain> or use developer :P
16:13:10 <frosch123> no idea what "developer" is, sounds like work
16:13:36 <TrueBrain> the ... other button on the Login page? :)
16:13:38 <frosch123> do you want to commit with a group account again? or query oauth for the email and commit as the user?
16:13:52 <frosch123> TrueBrain: yes, i saw it, did not press it :)
16:14:09 <TrueBrain> commit .. I dunno .. I am fine with both
16:14:19 <TrueBrain> if we do want to query email, we need to ask a bigger scope from oauth
16:14:30 <TrueBrain> using a group account might be easier
16:14:52 <TrueBrain> or we can do what you do for migration too: Username <username@users.openttd.org>
16:14:53 <frosch123> preview is weird. it shows === instead of h2
16:15:40 <TrueBrain> will have to check out what is going on there :P
16:16:21 <TrueBrain> yeah, I rather not increase the scope of the OAuth honestly ..
16:17:46 <frosch123> well, unlike bananas, here we have actual history pages
16:18:00 <TrueBrain> as said, we can just do Username <username@users.openttd.org> ofc
16:18:06 <TrueBrain> which is fine for history purposes?
16:18:11 <frosch123> ah, the edit page has no commit message though
16:18:25 <TrueBrain> is that ever used? Like .. at all?
16:18:29 <TrueBrain> I never know what to write in there
16:18:33 <TrueBrain> doesn't feel very useful to me
16:18:55 <frosch123> true :) though moving to opinionatedwiki :)
16:20:06 <TrueBrain> come to think of it, maybe we should also change how BaNaNaS commits, and use the anonymous github.com email for users
16:20:50 <TrueBrain> but it does link your GitHub account to the commit
16:20:55 <frosch123> then i prefer username@users.openttd.org
16:20:59 <TrueBrain> without having to deal with emails etc
16:21:18 <frosch123> yeah, but the hashes are github specific
16:21:19 <TrueBrain> so it would be TrueBrain <truebrain@users.noreply.github.com>
16:21:48 <frosch123> oh, that is a thing? i only saw frosch_1jij48zus8hjmnhsmjhasj@github.com
16:21:49 <TrueBrain> at least, I think that links the commits to my account
16:25:00 <TrueBrain> so that might be the nicest way to do this
16:26:42 <frosch123> yes, did not know it worked like that :)
16:27:03 <TrueBrain> and I think we can change BaNaNaS too, to make a commit per user like this
16:27:09 <TrueBrain> might be better than the librarian
16:27:21 <TrueBrain> (well, librarian can be the committer, we could just change the author)
16:27:36 <frosch123> originally bananas made a group commit for everything. then you split it into commit per package
16:28:37 <TrueBrain> so far there hasn't been a change between 2 users at the same time I think :P
16:29:05 <TrueBrain> but wiki first, BaNaNaS after
16:29:24 <TrueBrain> for wiki, I was thinking: on Save page, commit, and push only after 5 minutes of no commits, or something
16:30:05 <TrueBrain> so if someone is changing multiple pages, he has some time to go from one to the other
16:30:05 <frosch123> when a user edits 5 pages in that time, are they squashed into one commit?
16:30:24 <frosch123> and when they edit one page 5 times?
16:30:32 <TrueBrain> just commit on "Save page"
16:30:44 <frosch123> people do not know how "preview" works :)
16:30:53 <TrueBrain> true .. but otherwise the history is a bit meh
16:31:04 <TrueBrain> we can hold off on commits, but what if 2 people edit the same page
16:31:45 <TrueBrain> so from a history perspective, commit per edit is better
16:31:50 <TrueBrain> just possibly more bloat
16:32:08 <frosch123> hacktoberfest only counts prs? :p
16:32:31 <TrueBrain> well, we can always change it
16:32:38 <TrueBrain> commit per edit is the easiest to implement :P
16:33:56 <TrueBrain> I also never understood the "minor" edit checkbox on mediawiki
16:34:03 <TrueBrain> it is so much extra work they want you to do for making changes
16:34:11 <TrueBrain> like .. you figure it out if this was minor or not, don't let me figure that out
16:34:54 <frosch123> there are also reviews?
16:36:11 <andythenorth> I have used it at least 5 times
16:36:46 <TrueBrain> frosch123: I have no idea what you meant with "there are also reviews?" :) I am missing context :(
16:38:57 <frosch123> i thought sometimed pages have "changes waiting for review" and "reviewed commits"
16:39:04 <frosch123> but i cannot find an example
16:39:25 <frosch123> offical wiki. i think it's some kind of page protection on edit wars
16:39:28 <TrueBrain> you can "lock" pages
16:39:36 <TrueBrain> yeah, they have something for that in place
16:39:41 <TrueBrain> do we want something like that?
16:40:14 <TrueBrain> it is not difficult: instead of pushing to master, it would push to a branch and make a PR :P
16:41:25 <frosch123> no, i doubt andy would be up for reviewing them
16:42:23 <andythenorth> I have to spend the next 2 years drawing ships
16:42:57 <frosch123> "to get access to editing this page, please provide a unique ship sprite"
16:52:45 <FLHerne> What does the `traininfo_y_offset` variable mean?
16:52:55 * FLHerne just saw it and was curious
16:53:06 <FLHerne> "Used to correctly position the depot view of trains" isn't really informative
16:53:45 <FLHerne> Does that apply to all grfs, or just the current one?
16:54:34 <FLHerne> ...how can `climate` change during the game?
16:54:54 <frosch123> there also used to be a cheat
16:55:10 <FLHerne> frosch123: That's not really informative either :p
16:55:18 <frosch123> FLHerne: anyway, that var now applies to a single grf only, but it used to be global
16:55:38 <FLHerne> I can see how global would be a mess
16:56:02 <frosch123> well, positioning sprites in depot vs. on track is "opinionated" :)
16:56:28 <frosch123> mixing wagnos from different sets into the same consist is equally troublesome
17:03:14 <andythenorth> FLHerne I'm wrong, not implied
17:06:09 <Wolf01> <andythenorth> I have to spend the next 2 years drawing ships <- can't you just automate?
17:08:06 <frosch123> spend the next 2 years automating drawing?
17:08:24 <TrueBrain> did Wolf01 just really suggest that art can be automated? :P
17:08:34 <andythenorth> art can slightly be automated
17:08:59 <frosch123> you just can't tell people that is was automated :)
17:24:07 <FLHerne> andythenorth: Once you've drawn a few, you can start copy-pasting bits?
17:26:57 <andythenorth> FLHerne somewhat yes
17:27:10 <andythenorth> originally I planned to draw 8 hulls and use them for everything
17:27:16 <andythenorth> but that makes it all look the same
17:28:24 <FLHerne> Yeah, Squid has that problem to some extent
18:12:52 <TrueBrain> When your ships are done
18:15:44 <andythenorth> I need colours for hull, funnel, and cabin
18:15:49 <andythenorth> user configurable
18:15:53 <andythenorth> not happening eh
18:19:16 <TrueBrain> Unlikely, is the right word :p
18:29:38 <longtomjr> Just make it infinite refittable liveries for every cargo andythenorth
18:31:20 <longtomjr> it might not be the most fun user experience though, maybe just make one of the pieces configurable via refit, and the other 2 from cc
18:31:24 <frosch123> andythenorth: you can make that easily with spritestacks
18:42:31 <longtomjr> Not identical, similar though
18:43:03 <Eddi|zuHause> yes-ish, but it would need to be seen in context (like, dozens of these moving on a map)
18:43:39 <longtomjr> Different enough I think, maybe make one with 2 chimneys? (also other angles?)
18:43:50 <longtomjr> the hull is different, I think that helps
18:45:00 <andythenorth> ships are problematic in all respects
18:45:18 <longtomjr> just make trains that swim
18:45:51 <longtomjr> What is the different roles of the 2 ships?
18:46:00 <longtomjr> or just the same, but more and faster?
18:46:07 <Eddi|zuHause> @longtomjr we did that already, it's called "wetrails"
18:46:24 <andythenorth> bigger and smaller
18:46:39 <andythenorth> ships are needed in 6 different sizes
18:46:41 <longtomjr> Yep, I know, but I am imagining a IH train just chugging through the ocean.
18:46:54 <andythenorth> the requirement for different sizes poses challenges
18:47:08 <andythenorth> IRL all ships look the same, from 400grt to 4000grt
18:47:14 <longtomjr> andythenorth, I think it is fine if they look similar for different sizes.
18:47:22 <longtomjr> more chimneys for bigger ships
18:47:29 <andythenorth> they have to look similar-ish because they are the same type
18:47:34 <Eddi|zuHause> longtomjr: you could do that in TTO, halftile-water didn't crash trains, so you could send them down the slope to water level and they would go on
18:47:41 <andythenorth> there are about 15 types of ship, and 3-6 sizes of each
18:48:38 <longtomjr> What is the jump between the shackleton and longstone?
18:48:58 <longtomjr> cargo capacity wise
18:49:49 <longtomjr> Do you have other angles we can look at as well?
18:51:45 <andythenorth> other angles aren't drawn, takes a week to draw each ship
18:51:55 <andythenorth> 720 is the biggest
18:52:05 <andythenorth> the scale gets implausible after that
18:52:45 <longtomjr> I would remove the masts from the largest one I think.
18:55:01 <longtomjr> Is that the correct term for those pointy bits?
18:56:23 <andythenorth> they're masts and cranes
18:56:36 <andythenorth> they're needed to tie all the sizes of this type together visually
18:56:43 <andythenorth> and distinguish from other types
18:57:37 <longtomjr> Aah ok, I think having the extra one helps then, Maybe add some radio equipment at the top. (I assume that the bigger one is also more modern?
18:58:11 <longtomjr> Anyways, I am off for the evening.
18:59:02 <longtomjr> Good luck with the ships, excited for your new ship set!
19:00:52 <Eddi|zuHause> andythenorth: i'm thinking you could add some distinguishing highlights by shading the cargo holds, instead of having them a uniform colour the whole length
19:07:31 <frosch123> TrueBrain: i added a "def prepare(self):\npass" to storage/local.py
19:08:28 <andythenorth> Eddi|zuHause I tend to agree
19:08:35 <andythenorth> it's not so amenable to copy-paste
19:08:41 <TrueBrain> Ah, yes, I never tested local only :D
19:12:30 <frosch123> i like how my cpu fan tells me when it's done with loading the metadata :)
19:13:01 <TrueBrain> owh, and editing won't work currently; I have to release a new wikitexthtml version :)
19:13:19 <frosch123> i only edit via python :p
19:13:36 <TrueBrain> more: just so you know ;)
19:16:23 <TrueBrain> and I mostly disable the metadata scan btw .. as it is just too darn slow to make development useful :P
19:16:38 <TrueBrain> in _ _ main _ _ .py, line 75; you can just disable it
19:18:11 <TrueBrain> owh, and frosch123 , if you enable "github" as storage, remember it will wipe any commit/change in "data" :P (like it does with BaNaNaS)
19:18:20 <TrueBrain> so dont use "github" storage if you are working on your migration :D
19:19:12 <frosch123> what does the metadata scan do?
19:19:18 <TrueBrain> translations and categories
19:20:03 <TrueBrain> pushed a fix for the prepare() (and bumped wikitexthtml so editing works :P)
19:20:29 <TrueBrain> so without the metadata scan, you simply don't see if a page has other translations, and no page is in any category :)
19:34:47 *** HerzogDeXtEr has joined #openttd
20:01:02 <frosch123> down from 28k7 to 25k6 errors. now i also can post percentages like andy :)
20:01:58 <andythenorth> always move the goalpost when you get to 90% though
20:07:00 <TrueBrain> 10% improvement, not bad
20:09:47 <LordAro> for some reason, i have just fixed grfcodec's compilation on mingw
20:10:42 <FLHerne> ^ pushed more as an RFC than anything else
20:10:51 <FLHerne> Oh, I forgot to add the updated test results
20:14:08 <FLHerne> ...and I screwed up using this `gh` CLI tool and pushed the branch to the OTTD repo and not mine, so now I can't force-push...
20:15:16 <LordAro> FLHerne: sure? only master should be protected
20:17:09 <FLHerne> Back to just using git and the web interface, I understand that :p
20:17:13 <TrueBrain> LordAro: so who-ever made the NML one already copied the wrong one ;)
20:17:20 <FLHerne> Also, the tests might pass this time
20:17:21 <TrueBrain> I am trying to stop this tainting of the waters :P
20:20:47 <DorpsGek_III> [OpenTTD/nml] LordAro opened pull request #167: Fix: Dorpsgek config file had OTTD-specific things in it https://git.io/JTH6l
20:21:18 <TrueBrain> I could bitch about the commit message, I guess
20:21:40 <TrueBrain> and about a silly newline .. what is that extra newline doing there .. do others also have that? Hmm ..
20:21:55 <LordAro> i thought i only spelled the nml branchname wrong
20:21:56 <TrueBrain> no, they do not .. so yeah, I can bitch about that :D
20:23:51 <DorpsGek_III> [OpenTTD/nml] TrueBrain approved pull request #167: Fix: Dorpsgek config file had OTTD-specific things in it https://git.io/JTH6a
20:24:45 <LordAro> hmm, am i supposed to have to wait for the codeql run to finish before mering?
20:24:57 <LordAro> it is so much slower than everything else :p
20:24:58 <TrueBrain> you expect it to find anything more/less? :P
20:25:08 <LordAro> no, i mean i can't merge until it's finished
20:25:11 <LordAro> it's a required stage
20:25:16 <TrueBrain> so .. disable the required part?
20:25:22 <TrueBrain> I didn't enable it on any repository :)
20:25:26 <TrueBrain> exactly for this reason
20:25:32 <TrueBrain> it is fine if you want to wait for it, it is fine if you want to skip it :D
20:25:59 <TrueBrain> it isn't required .. lol .. so why can't you merge? Odd
20:26:15 <LordAro> webpage hadn't fully reloaded
20:26:18 <frosch123> FLHerne: no nml file hosted on devzone contains "GRFID"
20:26:23 <LordAro> it was greyed out until i moused over
20:26:30 <TrueBrain> it changed here too out of nothing
20:26:34 <TrueBrain> so something must have changed :P
20:26:34 <DorpsGek_III> [OpenTTD/nml] LordAro merged pull request #167: Fix: Dorpsgek config file had OTTD-specific things in it https://git.io/JTH6l
20:28:01 <TrueBrain> FLHerne: it is a stupid idea, so why keep on doing it? :)
20:28:05 <FLHerne> frosch123: Ok, I can delete it
20:28:59 <FLHerne> TrueBrain: Well, if the release actions would do it, it wouldn't be stupid
20:29:16 <TrueBrain> I dislike release actions that make commits honestly
20:29:23 <TrueBrain> it makes things somewhat unpredictable
20:29:42 <TrueBrain> LordAro: I think you misunderstood me
20:29:47 <TrueBrain> adding of the / was a good idea
20:30:03 <TrueBrain> as I am pretty sure the intend is not to ignore any "objs" folder anywhre
20:30:06 <TrueBrain> but one specific one
20:30:09 <LordAro> i looked at the nml gitignore, and copied that
20:30:23 <TrueBrain> again, copying from bad examples :P
20:30:26 <TrueBrain> you are doing well tonight :D
20:30:34 <TrueBrain> we changed mostly all other .gitignores to include a /
20:31:02 <TrueBrain> specific before generic :D
20:34:31 <TrueBrain> it has to be bad for the people making that addon, for the shit they are getting .. but on the other hand, read up om licenses really helps avoiding these kind of things :)
20:34:54 <TrueBrain> it is a nice warning for all, to somewhat understand what licensing is (and isn't) :)
20:39:42 <TrueBrain> owh, and deleting messages from an issue because you disagree with it .. well, that always means the pitchforks come out .. angry people doing stupid shit, etc :P
20:39:53 <TrueBrain> well, that was a good popcorn moment
20:42:55 <LordAro> TrueBrain: what's the distinction between GPL & AGPL licences?
20:43:06 <TrueBrain> webservices, API usage
20:43:19 <TrueBrain> with AGPL, if you host the software, you also have to release modifications
20:43:21 <TrueBrain> with GPL, you do not
20:43:31 <TrueBrain> basically, it is a fuck-AWS license
20:44:03 <LordAro> so not particularly relevant to this situation?
20:44:29 <FLHerne> Is there any sense in which you can 'host' a browser extension anyway?
20:44:32 <TrueBrain> I don't know the plugin honestly :)
20:44:42 <TrueBrain> but it seems you can replace AGPL with GPL in the stories
20:44:45 <FLHerne> I suppose you could have virtual machines, but who'd use that
20:45:13 <TrueBrain> the main issue the author created: if you accepted contributions, all contributors have to agree for a license change
20:45:19 <TrueBrain> well .. you have to do your best to reach all of them
20:45:34 <TrueBrain> in the end it is okay if you can't reach them all, etc .. but it is a process of months
20:45:40 <TrueBrain> not because-I-am-mad :P
20:46:54 <TrueBrain> btw, it is a reason to relicense bananas-api to AGPL .. if someone would run a clone of bananas-api, and modified it to be more awesome, but only makes the API available .. that is fine under GPL
20:47:38 <TrueBrain> the "Application Service Provider" hole :P I like that term :D
20:49:09 <TrueBrain> a clear example of big companies not respecting the intend of a license, and only following the exact letter .. sad world ... where is the Star Trek era? Is it there yet?
20:51:05 <frosch123> LordAro: to quote my company lawyer: you can use gpl stuff internally, but not as part of the product. don't even touch anything with agpl :)
20:53:37 <milek7_> hmm.. is there some way to have PR-like workflow in subversion?
20:53:54 <LordAro> we use svn branches at work quite a lot
20:54:24 <LordAro> there's not really any 'infrastructure' that enables anything like pull requests though
20:54:30 <frosch123> milek7_: phabricator
20:54:36 <LordAro> it's "run build against branch, merge if it passes"
20:54:49 <frosch123> is the only tool i heard of that handles git. but a git-svn bridge is probably beter
20:54:50 <LordAro> does phabricator support svn?
20:55:12 <LordAro> we have a reviewboard install at work, but it's basically unused
20:55:13 <frosch123> maybe they dropped it :p
20:55:24 <frosch123> but iirc that was the main feature of phabricator
20:55:33 <LordAro> and not particularly nice to work with - essentially have to add commits manually
20:56:29 <LordAro> i've only used phabricator very briefly about 7 years ago, and that was to migrate everything off it :p
20:56:59 <TrueBrain> Just stop using subversion :p
20:57:21 <milek7_> svn is fine, git sucks for large files
20:57:31 <frosch123> i should use "moderately high-performance" some time
20:58:50 <TrueBrain> Large files in general don't belong in code repositories :p
20:59:03 <LordAro> TrueBrain: it's being worked on
20:59:15 <milek7_> well, content has to be stored *somewhere*
20:59:15 <LordAro> but it requires reworking every single developement workflow
20:59:26 <TrueBrain> (Says he who is pushing wiki data in GitHub, hihi)
20:59:58 <TrueBrain> LordAro: I know :) helped doing it more than once in my life :)
21:00:19 <TrueBrain> Developers always enjoyed the new after a while ... never during
21:02:00 <LordAro> i have managed to convert our main monorepo to git - there's only one tool that can cope with it, and somehow requires ~250G RAM to actually convert it
21:02:12 <LordAro> the resulting .git dir is 9G
21:02:22 *** tokai|noir has joined #openttd
21:02:22 *** ChanServ sets mode: +v tokai|noir
21:03:38 <LordAro> given the amount of shit that's gone into that repo over the last 15 years (which itself was a CVS conversion), it's quite an impressive compression ratio
21:05:51 <TrueBrain> Well, I just archived everything before a year of the conversion .. helped a lot
21:06:09 <TrueBrain> And we accessed the archive like 3 or 4 times ... ever
21:06:19 <LordAro> it's certainly an option
21:06:20 <TrueBrain> And that was only because we were curious
21:06:46 <TrueBrain> Keeping an infinite history is a silly human emotion
21:10:20 <frosch123> that depends on who wrote the history :) sometimes it's very helpful to figure out what some code is supposed to do, or why it does it. but sometimes the result is just "they did not know either"
21:10:41 <LordAro> or if they did, they do not know now
21:13:57 <frosch123> hmm, something about "writing a script that filters a list of errors for the unexpected errors" sounds weird
21:15:25 <TrueBrain> So you are at that point :D
21:15:45 <TrueBrain> It is why I made it valid YAML btw, hoping that would help you a bit :)
21:16:41 *** gelignite has joined #openttd
21:16:49 <frosch123> i used the api on the old wiki to get a list of pages/categories/templates/files that are missing, to filter them from the "missing on new wiki" list :)
21:33:02 <frosch123> do you replace " " with "_"?
21:34:23 <frosch123> many files you report as missing have "_", but the sources say " "
21:35:31 <TrueBrain> I should have removed those replacements all
21:36:38 <frosch123> "es/Manual/Base Set/Road Vehicles/Camiones de Correo" reports "File:Wizzowow_Mail_Truck.png"
21:36:46 <TrueBrain> there is 1 replacement left, that changes "_" into " "
21:37:00 <TrueBrain> owh, but that is not for the link
21:38:18 <frosch123> hmm, oh, but the en/ also went missing
21:38:25 <frosch123> something is very strange here
21:38:43 <TrueBrain> " |imagen = Wizzowow_Mail_Truck.png"
21:38:53 <TrueBrain> so the error is correct
21:39:19 <TrueBrain> but my source might be outdated with your source, ofc :)
21:39:22 <frosch123> oh, so the issue is that i did not rerun the check after some fix?
21:39:53 <TrueBrain> well, the file doesn't mutate on its own, so that is a question for your bash history :)
21:39:53 <frosch123> yeah, wrong input file
21:40:27 <TrueBrain> pfew, happy it wasn't me :P I had so many replaces for the weirdest mediawiki shit :)
21:40:39 <TrueBrain> well, 50% in 1 day ..
21:40:44 <TrueBrain> keep that up, and you will be done ... never
21:42:22 <TrueBrain> I think tomorrow I will spend some time fixing those {{SERVER}} thingies, if you haven't tackled that already
21:43:15 <TrueBrain> And check out the image errors :)
21:46:12 <frosch123> pff, the pl translator translated yet another historic page
21:47:25 <frosch123> TrueBrain: pushed an update.
21:47:30 <frosch123> i am not going to look at SERVER
21:47:42 <frosch123> but i'll look at more missing images
22:02:58 <TrueBrain> I see you had fun :p
22:03:49 *** matt21347 has joined #openttd
22:16:31 <frosch123> LordAro: does changing "strncpy(dest, src, n);" to "strncpy(dest, src, n-1);" in safestrncpy help?
22:16:44 <frosch123> (may require moving the silly n>0 check
22:17:30 <frosch123> otherwise -Wstringop-truncation seems to be about "don't use strncpy", but that is also the reason "safestrncpy" exists...
22:19:39 <LordAro> frosch123: it does for the first one, but for the second it just results in "output may be truncated copying 1 byte from a string of length 2"
22:20:53 <frosch123> but MAXDRIVE is 3...
22:22:29 <LordAro> it's difficult to tell from all the inlining - the warning is only triggered with -O2
22:22:42 <LordAro> could it be the `dest[n-1] = 0` causing it?
22:24:56 <LordAro> ...uninlining the function manually also makes the warning go away
22:25:49 <frosch123> maybe MAXDRIVE is defined in some header?
22:26:03 <frosch123> so it is defined as 3 or as 2 in different places?
22:26:45 <frosch123> grfcomm.cpp includes path.h with some ifdef
22:26:53 <frosch123> does that include mingw?
22:28:21 <LordAro> nope, definitely 3 - reading preprocessed source
22:28:39 <frosch123> that #if around "path.h" is bonkers
22:28:46 <frosch123> it uses MAXDRIVE and friends unconditionally
22:28:54 <frosch123> so it assumes they are defined somewhere else?
22:29:17 <LordAro> well, they're definied in path.h
22:30:29 <frosch123> did you read the preprocessor source of grfcomm.cpp?
22:30:35 <frosch123> where is MAXDRIVE defined?
22:30:54 <LordAro> but it's definitely getting expanded to 3
22:31:02 <frosch123> but there is in an #if :p
22:31:54 <LordAro> WIN32 is very definitely defined on MINGW :p
22:32:54 <frosch123> that #if is still stupid. why would it not include path.h when unconditionally using macros and functions defined there?
22:33:45 <LordAro> you're trying to rationalise decisions made for a poorly-maintained 20 year old C++ program :p
22:34:27 <frosch123> oh, this file is core grfcodec. it's C, not C++. and written by an assembler guy
22:35:07 <frosch123> patchman uses a lot more spaces than dalestan :)
22:35:32 <LordAro> mm, you'd think dalestan was severely limited on disk space
22:36:30 <frosch123> you know why C uses "==" for comparison and "=" for assignment? :)
22:36:42 <TrueBrain> funny .. when editing templates which include the template itself, the aggressive caching of mediawiki causes it to not update after edit :D
22:37:09 <frosch123> TrueBrain: yes, that's a big annoyance. i always have to make a second change
22:37:33 <TrueBrain> well, TrueWiki doesn't have that issue :P
22:37:39 <frosch123> TrueBrain: i think it's a weird way to prevent infinite recursions
22:37:47 <TrueBrain> it does in preview btw .. which is an interesting bug on its own :)
22:37:56 <TrueBrain> I just think they forgot to invalidate their cache :P
22:42:33 <frosch123> pff. 1000 Talk and User pages :p
22:43:33 <frosch123> TrueBrain: btw. this was no option with gollum. but do you want .mediawiki pages for uploads?
22:43:55 <TrueBrain> is there ever anything interesting there?
22:43:57 <frosch123> or do you treat File: as pure uploaded files, without descriptions or whatever
22:44:17 <frosch123> there are translation links :p
22:44:27 <frosch123> no idea whether that counts as "interesting"
22:44:28 <TrueBrain> it sounds really useless to me, honestly
22:44:53 <TrueBrain> I mean .. do you really expect that you click an image and read something more than you did on the Page you saw the image?
22:45:25 <TrueBrain> so personally I would expect File to be pure uploads ... but I am not a wiki expert, and I might be totally wrong
22:45:32 <TrueBrain> LordAro: convince us otherwise!
22:45:50 <frosch123> TrueBrain: on wikipedia it contains the license of the image, etc
22:46:00 <frosch123> but our editors do not understand licenses
22:46:12 <TrueBrain> haha, no, that is very true :)
22:46:17 <TrueBrain> and most images are screenshots of OpenTTD
22:46:22 <TrueBrain> so that license is pretty obvious
22:46:45 <TrueBrain> unless you have an example on our current wiki that shows it is useful .. I would just not bother honestly
22:46:52 <TrueBrain> on the other hand, it is a small effort to port those over too
22:47:02 <TrueBrain> but ... opinionated wiki, remember :)
22:48:29 <TrueBrain> I think I removed all fullurl/localurl/SERVER entries
22:48:35 <TrueBrain> but I would need a new export from you to confirm :)
22:48:46 <frosch123> ... somehow that ":)" just reminded me about planetmaker always using ":-)" ... and now there is this term "opinionated smiley" in my head :/
22:48:59 <TrueBrain> hahaha :D I get what you mean :)
22:51:23 <TrueBrain> searching the wiki itself shows a few more instance of SERVER, but I think those are not exported :)
22:51:59 <TrueBrain> " Further instructions can be found in [//wiki.openttd.org/?title=Cross-compiling_for_Windows&oldid=90126 older version of this article]." <- wtf is this for bullshit?
22:52:59 <TrueBrain> BBBYYYEEEEEEEE, I said to that part of the article
22:54:02 <TrueBrain> frosch123: would you be able to modify the "hash" part too, btw?
22:54:07 <TrueBrain> as in, #title stuff?
22:54:21 <TrueBrain> as TrueWiki simply slugifies them, where mediawiki did ... something
22:54:39 <frosch123> yes, but i would have to undo whatever mediawiki does
22:54:46 <TrueBrain> I think that is not needed
22:54:49 <TrueBrain> pip install python-slugify
22:54:59 <TrueBrain> and I think if you just slugify what-ever is now behind the hash
22:55:02 <TrueBrain> it work for 90% of the pages
22:55:15 <TrueBrain> possibly urldecode it first
22:57:41 <frosch123> i started a new export, but i still do not plan to make it incremental :p
22:59:13 <LordAro> TrueBrain: it is quite annoying when i encounter various things that assume .png (or whatever) is an image and tries to load the html page as an image
23:00:07 <TrueBrain> "various things" is a bit unspecific, and allows me to read your sentence in several ways :P
23:08:31 <TrueBrain> frosch123: I am removing hard-links to wiki.openttd.org now .. we have many of those .. often very silly
23:09:15 <frosch123> that sounds like something the conversion could do
23:09:49 <TrueBrain> not really, as I have seen so far
23:10:19 <TrueBrain> well, you would be adding a case for each of them almost :P
23:11:36 <TrueBrain> yeah, but those are the rare ones
23:11:46 <TrueBrain> this is more what I find :P
23:12:19 <frosch123> ah, i think those are from before the 2010 wiki update :p
23:14:11 <frosch123> want to add a check to truewiki? "do not link to the wiki using external links"
23:14:21 <TrueBrain> I think that would be a good feature :)
23:14:49 <frosch123> hmm, is it possible to put relative urls into external links?
23:14:56 <TrueBrain> also lots of people that did not know :File: links to the file
23:15:12 <TrueBrain> not by their own specs, anyway
23:22:11 *** WormnestAndroid has quit IRC
23:22:52 *** WormnestAndroid has joined #openttd
23:28:45 <frosch123> just that one? how often does it appear? :p
23:30:09 <TrueBrain> didn't feel like editing that 37 times :P
23:30:38 <frosch123> added it to the list :)
23:31:06 <TrueBrain> right, I most likely missed a few wiki.openttd.org, but I will check that again when-ever that sed has ran :D
23:31:49 <TrueBrain> was looking at the diff, because .. I was bored I guess?
23:31:59 <TrueBrain> -|Image = Grfcrawler vehicles.png
23:31:59 <TrueBrain> -|Name1 = [[en/Community/NewGRF/2cc TrainSet|2cc - Trains of the World]]
23:32:03 <TrueBrain> +|Image =en/Grfcrawler vehicles.png|Name1 = [[en/Community/NewGRF/2cc TrainSet|2cc - Trains of the World]]
23:32:08 <TrueBrain> it is not wrong or broken, but is this intended?
23:32:47 <frosch123> well, wikitextparser includes the leading and trailing whitespace in the parameter
23:32:57 <frosch123> and i replace the parameter value
23:32:59 <TrueBrain> yeah ... I have a lot of strip() :D
23:33:18 <frosch123> that 7k reduction from earlier was adding a strip() :)
23:33:44 <frosch123> i did not know a page named "Grfcrawler vehicles.png\n" :p
23:34:07 <TrueBrain> well, maybe check if there was a \n in the original, and add one with the replacement?
23:34:17 <TrueBrain> might be nice to not completely mangle the mediawiki files :D
23:36:23 <TrueBrain> hmm .. some other errors we currently get we can easily fix in post
23:36:35 <TrueBrain> 's/center|none/center/g' is one of those :P
23:38:37 <TrueBrain> and for good measure, 's/none|center/none/g' .. at least, I think ..
23:38:42 <TrueBrain> hard to know what mediawiki does in the end :P
23:39:34 <TrueBrain> yeah, I have that right
23:39:40 <TrueBrain> frosch123: can you already integrate those 2 seds in your export?
23:40:35 <frosch123> hmm, do you mean \| ?
23:40:40 <TrueBrain> also the variation with "center" -> "right"
23:40:48 <TrueBrain> [[Image:Helicopter.png|right|none|frame|]]
23:41:09 <TrueBrain> that "none" doesn't do anything, basically
23:41:10 <frosch123> well, i can never remember what chars need escaping and which need unescaping :p
23:41:20 <TrueBrain> owh, sorry, I did not try the sed
23:41:25 <TrueBrain> possibly you need to escape shit, yes :D
23:42:26 <TrueBrain> no, do not escape it! DO NOT ESCAPE IT :P
23:42:28 <frosch123> basic regex and extended regex are difficult :)
23:42:30 <TrueBrain> that went .. horribly wrong :P
23:42:37 <frosch123> add/remove \ until it works
23:42:39 <TrueBrain> sed -i 's/center|none/center/g' Page/ //*.mediawiki
23:42:42 <TrueBrain> that does what I expect :)
23:42:50 <TrueBrain> you can also fix it via wikitextparser btw
23:42:56 <TrueBrain> check if they are both in there, and remove the last one
23:43:01 <TrueBrain> but that "last one" might be tricky
23:43:02 <frosch123> well, i do it in python anyway, no sed
23:43:21 <TrueBrain> you use "re" or "regex" library?
23:43:36 <frosch123> but also wikitextparser
23:43:43 <frosch123> only re where wtp fails
23:44:09 <TrueBrain> well, if you could look to implement this in your export, would be nice; otherwise we can do it post
23:44:12 <TrueBrain> it is not the worst problem
23:44:19 <TrueBrain> just an easy way to fix N of these errors :P
23:45:06 <frosch123> my original plan was to apply the conversions on the full history, so images/links also work on older revisions. but for some changes that is weird :)
23:45:39 <TrueBrain> owh, like this change ..
23:45:49 <TrueBrain> yeah, some changes just need to be a commit on top of the export tbh
23:45:53 <TrueBrain> "Fixing broken shit" :P
23:45:55 <frosch123> well, also the "remove edit link"
23:46:05 <TrueBrain> yeah, those are all post-fixes
23:46:20 <frosch123> well, i guess my script needs to learn post-fixes then :)
23:46:23 <TrueBrain> either click 37 times in the mediawiki interface, or a separate commit :P
23:46:32 <TrueBrain> your script will be huge :P
23:46:52 <TrueBrain> okay, I think I fixed most of the things I can fix
23:47:34 <frosch123> the scripts are smaller than their parameter files :)
23:49:28 <TrueBrain> frosch123: I do assume you don't plan to export the whole of the history; just of those files you marked as interesting? :D
23:50:04 <TrueBrain> as in, you move the files to their new location from commit 1, right?
23:50:10 <TrueBrain> (and trash those we don't care about)
23:51:04 <frosch123> yes, deleted files do not exist in the history
23:52:28 <TrueBrain> hmm .. making a "top 10" list of wiki pages accessed is tricky, because of the many ways you can access a single page :P
23:52:31 <frosch123> i also retroactively remove all redirects :)
23:52:48 <TrueBrain> yeah, so the history is already not really what the page was
23:52:54 <TrueBrain> it is already modified to work with TrueWiki :)
23:53:12 <frosch123> exactly, but the changes are only syntax, not content
23:53:25 <TrueBrain> yeah, content we can do in post, each their own commit or what-ever
23:54:22 <TrueBrain> oef, scanning 90 days for wiki access turns out to be a lot of data to process :D
23:55:41 <frosch123> is the wiki used? or is everything from us and the pl translators?
23:56:07 <TrueBrain> it is used a lot :)
23:56:08 <frosch123> i guess api.php is pretty high on the list :p
23:56:31 <TrueBrain> hmm .. api.php is not in the list of today
23:56:35 <TrueBrain> which is a bit odd :P
23:57:40 <TrueBrain> ah, they are unique
23:57:45 <TrueBrain> so they don't show up in any top NN
23:58:38 <frosch123> yeah, i do get's with urlparam
continue to next day ⏵
