Config
Log for #openttd on 20th July 2021:
Times are UTC Toggle Colours
00:13:26  *** snail_UES_ is now known as Guest1446
00:13:26  *** snail_UES_ has joined #openttd
00:18:25  *** Guest1446 has quit IRC
00:42:36  *** HerzogDeXtEr has quit IRC
00:43:23  *** chris has joined #openttd
00:44:08  *** chris has left #openttd
02:30:13  *** debdog has joined #openttd
02:33:35  *** D-HUND has quit IRC
02:40:04  *** Wormnest has quit IRC
03:24:44  *** glx has quit IRC
05:03:07  *** Flygon has joined #openttd
05:21:05  *** snail_UES_ has quit IRC
05:48:02  *** andythenorth has joined #openttd
06:54:44  <DorpsGek> [OpenTTD/game-coordinator] TrueBrain opened pull request #35: Change: give TURN more time to connect (from 10 to 20 seconds) https://git.io/JWpJU
06:56:39  <DorpsGek> [OpenTTD/game-coordinator] TrueBrain merged pull request #35: Change: give TURN more time to connect (from 10 to 20 seconds) https://git.io/JWpJU
06:57:40  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9447: Feature: allow the use of TURN to connect client and server together https://git.io/JWpJp
07:19:31  <DorpsGek> [OpenTTD/OpenTTD] LordAro updated pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWbsL
07:37:01  *** andythenorth has quit IRC
07:37:50  *** sla_ro|master has joined #openttd
07:40:00  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWp3X
07:42:02  *** HerzogDeXtEr has joined #openttd
07:45:05  <DorpsGek> [OpenTTD/OpenTTD] LordAro updated pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWbsL
07:45:40  <TrueBrain> LordAro: no update to second commit message? :D (not sure you are just not agreeing or forgot or didnt see the comment :P)
07:46:00  <TrueBrain> owh, you did change it; right :)
07:46:13  <TrueBrain> a pointer to what "certain circumstances" are might be good to add?
07:46:32  <TrueBrain> not sure my suggestion is any good for that :P
07:46:47  <LordAro> afaik it's "all circumstances"
07:46:59  <TrueBrain> really? Interesting :)
07:47:08  <TrueBrain> mainly what confused me, is that there is code handling T_MIN
07:47:18  <TrueBrain> owh, but from T_MAX
07:47:21  <TrueBrain> yeah, I get what you mean
07:47:25  <TrueBrain> okay, so change it in that, and lets go :P
07:47:25  <LordAro> well, that said i'm only aware of one scenario where i've actually reached INT64_MIN
07:47:35  <LordAro> (money)
07:48:25  <DorpsGek> [OpenTTD/OpenTTD] LordAro updated pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWbsL
07:48:37  <TrueBrain> Basically the gist is that T_MIN != -T_MAX, which was an assumption of the code, I guess
07:48:43  <LordAro> mm
07:48:59  <LordAro> not ideal, but not a lot we can do about it really
07:49:06  <TrueBrain> yeah, now the changes also make a lot more sense in my head :)
07:49:14  <TrueBrain> it was comparing against T_MAX, instead of T_MIN, for negative cases
07:49:21  <TrueBrain> well, -T_MAX
07:49:25  *** WormnestAndroid has quit IRC
07:49:26  <LordAro> ah
07:49:37  *** WormnestAndroid has joined #openttd
07:49:57  <TrueBrain> "T_MAX - abs(other.m_value)) < abs(this->m_value)"
07:50:01  <TrueBrain> that is where it went wrong :)
07:51:00  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWpZ2
07:51:26  <TrueBrain> finally understanding what it fixes, makes me happy :D Stupid math :P
07:55:23  <LordAro> :)
07:55:47  <TrueBrain> I appreciate someone wrote those original commits, but here you see how important good commit messages can be :)
07:57:40  *** esselfe has quit IRC
07:57:53  <TrueBrain> guess one day we should add a ton of constexpr in that file too
07:57:56  <TrueBrain> might help compilers
07:59:42  <LordAro> i can do that now, easily enough
07:59:49  <LordAro> i've noticed an error with a comment anyway
08:02:45  <TrueBrain> :D
08:05:41  <DorpsGek> [OpenTTD/OpenTTD] LordAro dismissed a review for pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWpZ2
08:05:44  <DorpsGek> [OpenTTD/OpenTTD] LordAro updated pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWbsL
08:06:55  *** Progman has joined #openttd
08:07:25  <TrueBrain> I am wondering if we indeed should add a few static_asserts, just so we know the constexpr has been applied correctly, and nobody breaks it in the future .. or are there easier ways to detect everything in a class is constexpr? Not sure it matters ..
08:07:48  <LordAro> not that i'm aware of
08:08:06  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWplW
08:09:14  <LordAro> though perhaps we should have at least one -
08:09:32  <TrueBrain> up to you!
08:09:34  <LordAro> note: ‘constexpr OverflowSafeInt<T>::OverflowSafeInt(int64) [with T = int; int64 = long long int]’ is not usable as a ‘constexpr’ function because:
08:09:47  <LordAro> error: member ‘OverflowSafeInt<int>::m_value’ must be initialized by mem-initializer in ‘constexpr’ constructor
08:11:20  <TrueBrain> lol, I have no clue what that error is aiming at :P
08:11:37  <LordAro> needs to be an : m_value(foo)
08:11:44  <LordAro> rather than an assignment in the function body
08:11:57  <TrueBrain> makes sense, I guess
08:12:45  <TrueBrain> they fixed that with C++20
08:12:49  <TrueBrain> but ... yeah ...
08:13:08  <LordAro> it's a "better" form of constructor anyway
08:13:26  <TrueBrain> :)
08:13:44  <LordAro> hmm, can't constexpr the builtin form of *= though
08:13:51  <LordAro> because of the `T out`
08:14:11  <LordAro> these static asserts are helping!
08:14:16  <TrueBrain> :D
08:14:24  <TrueBrain> owh, the built-in mul thingy
08:14:29  <TrueBrain> yeah, that wasn't in my godbolt .. :P
08:15:10  <LordAro> i'm not sure that's resolvable
08:15:27  <TrueBrain> I guess "__builtin_mul_overflow(this->m_value, factor, &this->m_value)" is not possible for a good reason?
08:15:41  <LordAro> JGR's second commit removes that
08:16:05  <TrueBrain> with no information why :P
08:16:07  <LordAro> i'm not sure what it fixes, even after reading the docs
08:19:12  <LordAro> but it does fix the issue that was reported to his repo ¯\_(ツ)_/¯
08:19:52  <TrueBrain> really curious what it fixes :D
08:20:38  <LordAro> probably starts writing to the output before fully reading the input
08:20:53  <LordAro> like strcpy and friends
08:21:00  <TrueBrain> would be my guess too, so I am testing that .. but so far .. nothing weird pops up
08:21:18  <TrueBrain> it is especially weird as the other variants don't need that
08:22:15  <LordAro> mm
08:22:52  <TrueBrain> maybe JGR remembers what was going on there :)
08:24:15  <TrueBrain> but godbolt testing shows no difference between the two approaches
08:26:43  <LordAro> well the static assert fails :p
08:26:58  <TrueBrain> "the" :P
08:27:25  <LordAro> static_assert(OverflowSafeInt32(INT32_MAX) * 2 == OverflowSafeInt32(INT32_MAX));
08:27:28  <LordAro> and also INT32_MIN
08:27:53  <TrueBrain> interesting
08:28:45  <TrueBrain> https://godbolt.org/z/v1WTeKd7z
08:28:47  <TrueBrain> that works fine :D
08:30:49  <LordAro> so it does
08:30:54  <LordAro> as does a constexpr variant of it
08:31:01  <TrueBrain> so what is the difference ...
08:32:55  <TrueBrain> ah! I think the issue is that "this->m_value" is invalid now
08:33:04  <TrueBrain> so you don't know if it has to become T_MIN or T_MAX
08:33:33  *** dP has joined #openttd
08:33:33  *** dP is now known as _dp_
08:34:42  <TrueBrain> yup
08:34:47  <TrueBrain> and that is also the solution I guess
08:34:51  <LordAro> ah
08:34:54  <TrueBrain> https://godbolt.org/z/1E5xb6znM
08:35:07  <TrueBrain> just do the check before the call
08:35:19  <LordAro> bit horrible.
08:36:00  <TrueBrain> https://godbolt.org/z/a866jPxba
08:36:03  <TrueBrain> better? :)
08:37:15  <LordAro> const bool is_positive = (this->m_value < 0) == (factor < 0);
08:37:17  <LordAro> is what i had :)
08:37:23  <LordAro> same difference
08:37:26  <TrueBrain> :D
08:37:39  <TrueBrain> nit: result_is_positive ;)
08:38:22  <LordAro> compiles \o/
08:41:21  <_dp_> > <TrueBrain> okay .. I am going to break the master-server JSON API .. no clue if other people use it .. guess their tool will break, and they will complain? Dunno :)
08:41:27  <_dp_> missed my complaints? :p
08:41:50  <_dp_> like, seriously, wtf, I should switch to parsing html, it seems more stable than that "api" :p
08:42:03  <TrueBrain> I hope you are trolling ;)
08:42:16  <_dp_> yeah, kinda
08:42:27  <TrueBrain> as if 1 change in .. 5 years is too much for you ..... ;)
08:42:49  <_dp_> but some kind of an announcement would be nice next time ;)
08:43:00  <TrueBrain> I did ;) You quoted it :D
08:43:08  <TrueBrain> we have 0 infrastructure to make such announcements, sadly enough :(
08:43:16  <_dp_> api did not exist for 5 years
08:43:19  <TrueBrain> we had the same issue with finger.openttd.org .. no platform :(
08:43:33  <_dp_> not on irc, some place it can be actually noticed :p
08:43:49  <TrueBrain> name 1
08:43:51  <_dp_> at least on discord #developer-updatess
08:44:25  <TrueBrain> I could next time, but I somehow doubt enough people read that :) It really is annoying, that we do not have something for this ...
08:44:49  <TrueBrain> in general, I would like a "notify me if this API changes" thing
08:45:00  <Zuu> Is a news article on the front page too big?
08:45:06  <_dp_> also you literally happened to pick the worst time breaking it the day after I went on vacation and stopped reading irc :p
08:45:20  <_dp_> well, I do read  #developer-updates
08:45:31  <_dp_> even though it's somewhat spammy with newgrf changes
08:46:04  <_dp_> front page is kind of pointless because you have to open it specifically
08:46:13  <LordAro> TrueBrain: << has no overflow protection
08:46:13  <_dp_> at least discord does notifications
08:46:13  <TrueBrain> Zuu: kinda .. I once considered a "developers news page", but .. it never went anywhere :)
08:46:20  <TrueBrain> these changes are so freaking rare ..
08:46:20  <LordAro> frontpage is too user-focused
08:46:29  <TrueBrain> LordAro: lol ... who uses << anyway :P
08:47:10  <TrueBrain> it is also the reason nowhere officially we tell about these APIs
08:47:21  <TrueBrain> as .. well .. I have no way of reaching the users :(
08:47:38  <TrueBrain> I once considered versioning the APIs ... then I realised what a hell-hole it is to support all the old versions ..
08:47:56  <TrueBrain> either way, now I know _dp_ is a user, I will poke you next time, no worries
08:49:01  <_dp_> ty
08:49:03  <Zuu> anyway I see the point and taking something down or break an API is when you get to know if someone really care about the service provided. :-)
08:49:21  <LordAro> if an API is undocumented, that's basically a notification that it can and will change without notice
08:49:42  <TrueBrain> LordAro: that is basically the route I opt'd for :P Not ideal .. but again, I can't find anything better :D
08:49:42  <_dp_> though in general it would still be a good practice to announce changes beforehand
08:49:47  <_dp_> that's kind of the whole point of api
08:49:54  <LordAro> not really
08:50:03  <LordAro> the whole point of the api is for the frontend, no?
08:50:15  <LordAro> if it's useable by other people, then great
08:50:30  <_dp_> I had an impression it was made for external tools
08:50:36  <DorpsGek> [OpenTTD/OpenTTD] LordAro dismissed a review for pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWplW
08:50:39  <DorpsGek> [OpenTTD/OpenTTD] LordAro updated pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWbsL
08:51:02  <TrueBrain> the only well documented API we have, is the BaNaNaS-api
08:51:07  <TrueBrain> it even has a swagger-file for it
08:51:13  <TrueBrain> but that is because it does have external tools using it too
08:51:20  <TrueBrain> so changes to that API will be processed with a lot more care
08:51:27  <TrueBrain> for the others .. I do my best to keep them stable
08:51:34  <TrueBrain> just in this very very rare case, I couldn't
08:52:49  <TrueBrain> LordAro: maybe for the "nice" put the *= thingy in the other commit?
08:53:11  <LordAro> sure
08:53:17  <LordAro> that's going to be a rebase pain :p
08:53:23  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWpwt
08:54:04  <TrueBrain> ugh, got a new keyboard .... the keys are a lot louder ... it is already annoying me :P Lol ...
08:57:07  <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWprk
08:57:13  <TrueBrain> LordAro: kinda curious if constexpr also makes it faster :D
08:57:40  <TrueBrain> haha, lol @ comment :D I know that pain/feeling :D
09:00:14  <DorpsGek> [OpenTTD/OpenTTD] LordAro updated pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWbsL
09:10:19  *** virtualrandomnumber has joined #openttd
09:10:34  *** virtualrandomnumber has quit IRC
09:39:19  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWpyr
09:42:07  <DorpsGek> [OpenTTD/OpenTTD] LordAro closed issue #8284: Negating or taking absolute values of maximally-negative OverflowSafeInt may result in undefined behaviour https://git.io/JJz6P
09:42:10  <DorpsGek> [OpenTTD/OpenTTD] LordAro closed pull request #8285: Rewrite `OverflowSafeInt` to avoid undefined behaviour https://git.io/JJgcH
09:42:13  <DorpsGek> [OpenTTD/OpenTTD] LordAro merged pull request #9451: Fix potential undefined behaviour when using OverflowSafeInt https://git.io/JWbsL
10:03:39  *** Samu has joined #openttd
10:06:50  <LordAro> https://nvd.nist.gov/vuln/detail/CVE-2021-33623
10:06:53  <LordAro> sigh.
10:08:10  <TrueBrain> May 28 .. I hope everyone updated by now :P
10:08:44  <LordAro> TrueBrain: actions repo hasn't been :p
10:09:01  <TrueBrain> https://github.com/Xmader/musescore-downloader/issues/5 <- in case you are bored and want to be horrified .. it goes well for a while, but .. in the end ... something snaps :P
10:09:23  <TrueBrain> LordAro: yeah, I am lazy ... especially as it is not usable by the average joe :P But I will see if I can patch it up soon :)
10:09:32  <TrueBrain> well, you can do it too ofc, but yeah :)
10:09:38  *** Zuu has quit IRC
10:11:47  <LordAro> i could!
10:13:24  <TrueBrain> ugh, I forgot how annoying updating NPM was
10:13:27  <TrueBrain> this is why I never do it :P
10:13:55  <TrueBrain> gues I could enable dependabot for it
10:16:01  *** sla_ro|master has quit IRC
10:17:28  <TrueBrain> will look at it later today if you haven't by that time :)
10:17:56  *** sla_ro|master has joined #openttd
10:22:49  <DorpsGek> [OpenTTD/actions] LordAro opened pull request #29: Update: package-lock.json https://git.io/JWpbl
10:29:20  <TrueBrain> There are more lock files in that repo :D
10:29:28  <TrueBrain> One in every subfolder
10:29:33  <LordAro> oh no
10:31:09  <peter1138> +9945-1823... Wow.
10:31:17  <LordAro> format change, apparently
10:31:30  <LordAro> i don't think there are actually that many new dependencies :p
10:31:51  <TrueBrain> Npm7 instead of npm6 :)
10:32:21  <LordAro> https://github.com/OpenTTD/website/security/dependabot/requirements.txt/aiohttp/open there's also this one
10:32:32  <LordAro> but we're already on 3.7.4.post0, which is newer than 3.7.4
10:32:35  <LordAro> silly github
10:32:37  <TrueBrain> In theory we should also check if package.json has anything to update .. but that is even more annoying
10:32:47  <TrueBrain> Yeah .. I muted it on all the other repos
10:38:18  <DorpsGek> [OpenTTD/actions] LordAro updated pull request #29: Update: package-lock.json https://git.io/JWpbl
10:40:11  <DorpsGek> [OpenTTD/website] LordAro opened pull request #220: Update: Gemfile.lock https://git.io/JWppB
10:41:57  <TrueBrain> not sure why, but the way you updated the package-lock files violate the version constraint given in de package.json :D
10:41:59  <TrueBrain> that is very weird
10:42:11  <LordAro> oh?
10:42:26  <TrueBrain> or am I misunderstanding what ^1.2.7 means ..
10:42:51  <LordAro> which one are you looking at?
10:43:00  <TrueBrain> anyway, after updating you have to run `npm run all` to make sure all the src files are generated with the new dependencies :P
10:43:08  <TrueBrain> actions/core, for example, in annotation-check
10:43:39  <TrueBrain> ah, ^ allows minor releases to update
10:43:46  <TrueBrain> okay, so this is the expected outcome :D
10:44:12  <TrueBrain> I thought it only allowed patches .. but that is fine
10:44:15  <TrueBrain> shouldn't be a problem :)
10:47:36  <LordAro> hmm
10:47:56  <LordAro> i've run `npm run all` and now a load of stuff in various dist/index.js has been changed
10:48:01  <TrueBrain> yup
10:48:12  <TrueBrain> it now contains the actual new code ;)
10:48:25  <TrueBrain> only updating the package-lock.json doesn't do anything, basically
10:48:31  <TrueBrain> rerendering the js file does
10:48:33  <LordAro> right
10:48:35  <TrueBrain> it is annoying and confusing
10:48:46  <TrueBrain> we might want to rewrite that repo to use composite steps instead
10:48:50  <TrueBrain> which is basically just bash :)
10:49:09  <LordAro> heh
10:49:14  <DorpsGek> [OpenTTD/website] TrueBrain approved pull request #220: Update: Gemfile.lock https://git.io/JWpjp
10:49:23  <DorpsGek> [OpenTTD/website] TrueBrain merged pull request #220: Update: Gemfile.lock https://git.io/JWppB
10:49:44  <TrueBrain> as so often I forget to run "npm run all", and am annoyed that it is not working :D
10:55:26  <DorpsGek> [OpenTTD/actions] LordAro updated pull request #29: Update: package-lock.json https://git.io/JWpbl
10:55:44  <LordAro> +15968 -4321
10:55:49  <LordAro> #justnpmthings
10:59:07  <TrueBrain> The more I use npm, the more I hate it
10:59:13  <LordAro> yup
11:04:00  <TrueBrain> I am just scared to approve that PR, because it is impossible to tell if anything breaks :D Lol
11:04:13  <LordAro> yup!
11:04:28  <DorpsGek> [OpenTTD/actions] TrueBrain approved pull request #29: Update: package-lock.json https://git.io/JWhU1
11:04:37  <DorpsGek> [OpenTTD/actions] TrueBrain merged pull request #29: Update: package-lock.json https://git.io/JWpbl
11:04:42  <TrueBrain> fucking YOLO it
11:04:54  <LordAro> lol
11:04:56  <TrueBrain> and now I am so mad, that I am going to see which ones I can change into non-npm variants :P
11:05:03  <LordAro> :D
11:05:22  <TrueBrain> this is just insane
11:05:46  <TrueBrain> I mean, it is 60 lines of Javascript
11:06:00  <TrueBrain> compiles to 9546 lines of Javascript
11:11:20  <DorpsGek> [OpenTTD/actions] TrueBrain opened pull request #30: Change: use GitHub composite actions over NPM-based where possible https://git.io/JWhka
11:11:20  <TrueBrain> I do need to test it
11:11:22  <TrueBrain> but just as an example
11:11:33  <TrueBrain> +11 −2,051
11:11:35  <TrueBrain> absolutely FUCK npm
11:13:58  <LordAro> haha
11:16:27  <LordAro> TrueBrain: if we're being pedants, 50 lines of Typescript*
11:16:49  <TrueBrain> ha, fair point
11:21:22  <peter1138> Hmm, soft gherkins are somewhat unpleasant.
11:21:59  <TrueBrain> I had to look up wtf gherkins were :P
11:40:56  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #30: Change: use GitHub composite actions over NPM-based where possible https://git.io/JWhka
11:40:56  <TrueBrain> 3 done .. few more to go :P
11:41:11  <TrueBrain> +138 −15,352
11:41:12  <TrueBrain> so far so good :D
11:43:14  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #30: Change: use GitHub composite actions over NPM-based for checkout* actions https://git.io/JWhka
11:43:26  <TrueBrain> okay, lets start with these 3 first .. if you like to review that LordAro , much appreciated :D
11:43:58  <TrueBrain> oof, hmm .. the checkout-pull-request I should test
11:44:02  <TrueBrain> as it is broken .. but how to test it .. euhm
11:45:22  <TrueBrain> ha, found a way :D
11:48:50  <DorpsGek> [OpenTTD/actions] LordAro requested changes for pull request #30: Change: use GitHub composite actions over NPM-based for checkout* actions https://git.io/JWhsZ
11:50:14  *** esselfe has joined #openttd
11:50:30  <TrueBrain> not the only bug in that action .. damn, I did a sloppy job there :D
11:55:17  <TrueBrain> attempt #2 :)
11:55:18  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #30: Change: use GitHub composite actions over NPM-based for checkout* actions https://git.io/JWhka
11:55:45  <TrueBrain> I now tested it :D
11:55:54  <LordAro> there are some advantages to it being written in not-bash :p
11:56:02  <LordAro> not many, but some
11:56:27  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #30: Change: use GitHub composite actions over NPM-based for checkout* actions https://git.io/JWhka
11:56:33  <TrueBrain> the reason I did it in Typescript, as it is a lot easier to write it
11:56:36  <TrueBrain> but ... it is not worth it :D
11:57:07  <TrueBrain> can't really test checkout-dispatch .. but I am sure we fix it if it is broken :P
11:58:00  <DorpsGek> [OpenTTD/actions] LordAro approved pull request #30: Change: use GitHub composite actions over NPM-based for checkout* actions https://git.io/JWhZe
11:58:12  *** snail_UES_ has joined #openttd
11:58:22  <DorpsGek> [OpenTTD/actions] TrueBrain merged pull request #30: Change: use GitHub composite actions over NPM-based for checkout* actions https://git.io/JWhka
12:00:18  <TrueBrain> the next ones are even harder to test :D Will be fun :P
12:41:07  *** glx has joined #openttd
12:41:07  *** ChanServ sets mode: +v glx
12:52:21  *** gelignite has joined #openttd
13:26:12  <DorpsGek> [OpenTTD/actions] TrueBrain opened pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWh1I
13:26:17  <TrueBrain> 3 more down .. 4 left to go :P
13:27:45  <glx> IIRC npm was the only option when all these actions were created
13:27:47  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWh1I
13:27:54  <TrueBrain> yup
13:28:01  <TrueBrain> and I am not regretting making it burn in a big fire
13:28:30  <TrueBrain> hmm ... secrets are misbehaving ..
13:29:58  <TrueBrain> guess I cannot access secrets .. would make sense :P
13:31:32  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWh1I
13:35:46  *** Progman has quit IRC
13:41:51  <TrueBrain> right, that should be it :D
13:50:31  *** nielsm has joined #openttd
14:22:50  <LordAro> :)
14:34:21  <TrueBrain> okay, lets see if I can crack at least 2 of the remaining 4 .. should be easy
14:35:25  <TrueBrain> ah, except that I cannot call other actions, I forgot about that .. that was the stupid thing called composite actions
14:35:26  <TrueBrain> eeuuuhhh
14:35:29  <TrueBrain> curl I guess?
14:37:15  <Samu> some AIs can't really handle a 2 Billion loan
14:37:33  <Samu> and i was thinking that would help them
15:19:52  *** sla_ro|master has quit IRC
15:56:20  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWh1I
15:56:27  <DorpsGek> [OpenTTD/actions] LordAro approved pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWjuZ
15:56:31  <TrueBrain> sorryyyyyyy
15:56:32  <DorpsGek> [OpenTTD/actions] TrueBrain dismissed a review for pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWjuZ
15:56:35  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWh1I
15:56:38  <LordAro> lol
15:57:01  <LordAro> no activity for 2 hours, and you push 7s before i approve
15:57:14  <TrueBrain> I know right :(
15:57:24  <TrueBrain> and I didn't even want to do it like this!
15:57:53  <TrueBrain> there ... all fixed again
15:57:55  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWh1I
16:00:05  <TrueBrain> https://github.com/OpenTTD/actions/compare/4891852..19264022c5bbac5b56186a1d3e4d21226d3fcf20 for the diff
16:06:10  <DorpsGek> [OpenTTD/actions] LordAro approved pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWj2q
16:06:12  <LordAro> it's not like i'm reviewing things *that* closely
16:06:35  <TrueBrain> at least I am testing :P
16:06:45  <DorpsGek> [OpenTTD/actions] TrueBrain merged pull request #31: Change: use GitHub composite actions over NPM-based for docker* actions https://git.io/JWh1I
16:07:13  <LordAro> quite :)
16:07:33  <DorpsGek> [OpenTTD/actions] TrueBrain opened pull request #32: Change: use GitHub composite actions over NPM-based for deployments* actions https://git.io/JWj2i
16:21:03  <peter1138> Is it a sensible temperature to go on a bike ride? (No)
16:23:00  <LordAro> it's nicer when you're moving
16:23:13  <TrueBrain> okay .. so 2 more actions to go ... lets see ...
16:23:16  <LordAro> sweating while moving is better than sweating into your chair, anyway
16:28:40  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 commented on pull request #9447: Feature: allow the use of TURN to connect client and server together https://git.io/JWjKX
16:30:42  <Samu> strange, KrakenAI doesn't crash on "excessive CPU usage in valuator function" though the error shows up
16:30:56  <Samu> what kind of trickery is this ai using
16:31:38  <TrueBrain> Rubidium: I understand that you want to see an enum .. so do I .. but fuck adding enums to our codebase :P It is such an annoying job ........ I was hoping nobody would notice :D
16:32:09  <LordAro> what's difficult about it?
16:32:23  *** virtualrandomnumber has joined #openttd
16:32:25  <TrueBrain> location location location
16:32:41  <TrueBrain> after that: casting
16:32:44  <TrueBrain> after that ...
16:32:48  <TrueBrain> using numbers is just easier :D
16:33:04  <LordAro> if you're doing it right you shouldn't need casting :p
16:33:11  <TrueBrain> have you seen our codebase?
16:33:19  <LordAro> shh
16:33:21  <TrueBrain> :)
16:35:31  *** virtualrandomnumber has quit IRC
16:35:57  <Samu> the crash is in superlib-40, this is odd
16:36:06  <Samu> a fake crash
16:37:11  <glx> main issue is enum containing not directly related data, like flags mixed with numbers
16:39:51  <TrueBrain> main issue for you maybe :P
16:40:12  <TrueBrain> it is an issue, but not my main >:D
16:40:14  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 approved pull request #9450: Fix #8335: Race condition in music mixer https://git.io/JWjPo
16:40:39  <TrueBrain> ugh, there is no way I can test deploy-aws
16:40:44  <TrueBrain> hmm ... what to do about that ...
16:40:57  <LordAro> merge it!
16:41:23  <TrueBrain> I don't even know if it is a valid action .. even testing that is tricky
16:41:27  <TrueBrain> but okay .. it will indeed be a YOLO :P
16:41:38  <LordAro> then do something that makes use of the action :p
16:42:31  <peter1138> Urgh at old program that uses an MS Access database. Over a network share. Over a WAN...
16:43:01  <LordAro> burn it
16:44:11  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #32: Change: use GitHub composite actions over NPM-based for deployments* actions https://git.io/JWj2i
16:44:12  <TrueBrain> okay .. 1 action left, but that is for another day
16:45:35  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #32: Change: use GitHub composite actions over NPM-based for deployments* actions https://git.io/JWj2i
16:46:18  <TrueBrain> I mostly do not know if those "export" statements in deploy-aws work
16:46:43  <TrueBrain> I think they should work
16:47:48  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9447: Feature: allow the use of TURN to connect client and server together https://git.io/JWj18
16:51:53  <TrueBrain> owh, shit, I added the stuff in the same commit again didn't I? Ugh ...
16:52:25  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #32: Change: use GitHub composite actions over NPM-based for deployments* actions https://git.io/JWj2i
16:52:41  <TrueBrain> sorry about that ... I will first await this PR before creating the next :D
16:55:05  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9447: Feature: allow the use of TURN to connect client and server together https://git.io/JWDeR
16:59:17  <peter1138> 17:23 < LordAro> it's nicer when you're moving
16:59:22  <peter1138> True but... 29 degrees :/
17:01:04  <peter1138> Maybe it'll cool a bit.
17:01:53  <Samu> ah... it's a try/catch
17:02:51  <Samu> this trick makes the AI not crash when it's supposed to crash, isn't that cheating?
17:03:01  <Samu> circunvents a crash
17:05:09  <Samu> https://i.imgur.com/XYbSyj6.png
17:05:35  <Samu> the error is shown, but the AI doesn't stop, it continues, it's continuously doing this
17:05:49  *** virtualrandomnumber has joined #openttd
17:06:02  <LordAro> it does seem like that should be uncatchable
17:06:09  *** virtualrandomnumber has quit IRC
17:09:02  <peter1138> Why is using too much CPU an exception?
17:09:20  <Samu> he's using valuators inside valuators
17:09:21  <glx> because it hangs the whole game
17:10:02  <Samu> ended up crashing in a valuator from a library superlib-40
17:10:08  <glx> and squirrel uses exceptions to communicate with openttd
17:10:11  <Samu> but it preceeds his own valuator
17:11:51  <Samu> the map is too big for the valuators
17:12:20  <peter1138> Oh right, and valuators can't be suspended?
17:13:04  <LordAro> /* Kill the script when the valuator call takes way too long. Triggered by nesting valuators, which then take billions of iterations. */
17:13:48  <glx> valuating is a native function, it can't be suspended
17:14:01  <LordAro> does squirrel have any concept of an uncatchable error?
17:14:07  <LordAro> /exception
17:14:13  <LordAro> i suspect not
17:22:48  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 approved pull request #9447: Feature: allow the use of TURN to connect client and server together https://git.io/JWjQX
17:23:16  <TrueBrain> pam pam pammmmmmm
17:23:22  <TrueBrain> tnx a lot Rubidium for all these reviews; much appreciated!
17:23:44  <glx> now we need the real test, players
17:24:46  <TrueBrain> hmm .. I guess I first want to merge "actions" PR, so I can test if they work when deploying a new Game Coordinator :)
17:25:18  <TrueBrain> I tested Deploy AWS mostly, so that is a good :D
17:25:45  *** Wormnest has joined #openttd
17:26:12  <TrueBrain> so can I tempt LordAro to review https://github.com/OpenTTD/actions/pull/32 ... :D
17:30:33  * LordAro looks
17:31:52  <DorpsGek> [OpenTTD/actions] LordAro approved pull request #32: Change: use GitHub composite actions over NPM-based for deployments* actions https://git.io/JWj5Y
17:32:01  <DorpsGek> [OpenTTD/actions] TrueBrain merged pull request #32: Change: use GitHub composite actions over NPM-based for deployments* actions https://git.io/JWj2i
17:32:22  <DorpsGek> [OpenTTD/actions] TrueBrain opened pull request #33: Change: use GitHub composite actions over NPM-based for deploy-aws action https://git.io/JWj5c
17:32:24  <TrueBrain> and last one of the day :P
17:33:08  *** andythenorth has joined #openttd
17:33:42  *** tokai|noir has joined #openttd
17:33:42  *** ChanServ sets mode: +v tokai|noir
17:34:00  <TrueBrain> wait, I missed an "echo"
17:34:30  <DorpsGek> [OpenTTD/actions] TrueBrain updated pull request #33: Change: use GitHub composite actions over NPM-based for deploy-aws action https://git.io/JWj5c
17:34:31  <TrueBrain> there we go .. should be good for review too now :)
17:37:31  <TrueBrain> https://blog.mozilla.org/security/2021/07/20/stopping-ftp-support-in-firefox-90/ <- noooooooooooooooooooooooooooooooooooooooooooooo
17:37:36  <TrueBrain> wait ... people were still using FTP?
17:37:37  <TrueBrain> owh
17:37:38  <TrueBrain> nevermind
17:37:39  <TrueBrain> :P
17:38:26  <glx> using ftp without filezilla ?
17:40:36  *** tokai has quit IRC
17:40:50  <TrueBrain> +44 −11,294
17:40:54  <TrueBrain> that just cracks me up
17:40:55  <TrueBrain> fuck npm
17:40:59  <Rubidium> TrueBrain: yes!
17:42:03  <Rubidium> for some reason uploading large files is still easier with FTP than a web browser
17:42:27  <TrueBrain> I don't even remember when the last time was when I wanted to connect over FTP
17:42:28  <TrueBrain> SFTP, sure
17:42:30  <TrueBrain> but FTP ..
17:42:39  <TrueBrain> but, to everyone his own :)
17:43:13  <nielsm> hey! I occasionally use FTP to transfer files to my retro pc's!
17:45:12  <Rubidium> TrueBrain: well... the customer also runs Silverlight on the website, so I guess think again. I would have liked SFTP a lot more too, though this is FTP without even SSL
17:46:03  <glx> isn't silverlight supposed to be dead ?
17:46:18  <Rubidium> nah, it's still supported
17:46:29  <Rubidium> let me open the DHCP configuration
17:46:53  <Rubidium> ... for another 84 days
17:49:18  <andythenorth> yo
17:50:45  <TrueBrain> I see what you did there Rubidium :P
17:50:59  <TrueBrain> hmm ... with what can I bribe LordAro one last time for a review .. hmmmmm :D
17:53:30  <Samu> CoronaAI is such a small AI code-wise, and it works brilliantly
17:54:05  <Samu> 1 bus in every town, reusing the town's own roads
17:54:17  <Samu> no pathfinding needed
17:56:36  <Samu> profitable
17:57:28  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #9447: Feature: allow the use of TURN to connect client and server together https://git.io/JWDeR
17:59:25  <andythenorth> so YouTube https://www.youtube.com/watch?v=abqMLqHwqpo
17:59:35  <andythenorth> definitely SFW
18:07:52  *** Flygon has quit IRC
18:16:00  <Rubidium> andythenorth: depends on your job I guess ;)
18:18:18  *** EER has joined #openttd
18:32:33  <DorpsGek> [OpenTTD/game-coordinator] TrueBrain opened pull request #36: Change: [Actions] also deploy TURN server upon merging / releasing https://git.io/Jleff
18:32:48  <TrueBrain> right, while waiting on the last review .. lets at least start testing the other changes, so I do not forget :D
18:33:18  <DorpsGek> [OpenTTD/actions] TrueBrain created new tag: v2.1.0 https://git.io/Jlefm
18:35:46  <DorpsGek> [OpenTTD/game-coordinator] TrueBrain closed issue #33: TURN server doesn't auto-deploy https://git.io/JWAv7
18:35:49  <DorpsGek> [OpenTTD/game-coordinator] TrueBrain merged pull request #36: Change: [Actions] also deploy TURN server upon merging / releasing https://git.io/Jleff
18:35:59  <TrueBrain> let's see what explodes :)
18:36:39  <TrueBrain> so far so good
18:37:06  <TrueBrain> can't believe it, but they all still seem to work fine :P
18:37:14  <TrueBrain> with 50k less lines :D
18:37:48  <DorpsGek> [OpenTTD/game-coordinator] TrueBrain created new tag: 1.0.1 https://git.io/JleJv
18:51:49  <DorpsGek> [OpenTTD/OpenTTD] DorpsGek pushed 1 commits to master https://git.io/JleTG
18:51:50  <DorpsGek>   - Update: Translations from eints (by translators)
18:51:57  <TrueBrain> right .. STUN is operation .. so is TURN .. this seems to be deployed and ready for use
18:52:09  <TrueBrain> there are 2 TURN instances in rotation (turn-1 and turn-2) .. I can always scale up if it is needed
18:54:26  <TrueBrain> we might want to make a "reconnect" thing in OpenTTD, that if a connection is lost it tries to reconnect on its own
18:54:28  <TrueBrain> but .. future-work :)
18:55:53  <DorpsGek> [OpenTTD/aws-infra] TrueBrain pushed 2 commits to main https://git.io/JleTH
18:55:54  <DorpsGek>   - Fix: use a single Version key for the TURN servers (by TrueBrain)
18:55:55  <DorpsGek>   - Add: deploy a Test task, so GitHub Action "deploy-aws" can be tested (by TrueBrain)
18:59:14  <TrueBrain> @calc 64 * 0.015
18:59:14  <DorpsGek> TrueBrain: 0.96
18:59:18  <TrueBrain> okay .. so the game-coordinator leaks about 1MB of RAM every 90 minutes ..
18:59:52  <TrueBrain> it runs with 32MB RAM free after startup, so that gives it ~50 hours before it runs out of RAM
19:05:35  <glx> still no clues about the leak ?
19:05:55  <TrueBrain> enough clues, but haven't tried reproducing it :)
19:06:09  <TrueBrain> as nobody is using the game-coordinator, a lot of things are NOT the issue :D
19:06:59  <glx> not too many nightlies users yeah
19:07:36  <glx> so I guess it may be in the intercommunication with master server
19:08:12  <TrueBrain> or a background task
19:09:39  <TrueBrain> it is insanely consistent, which is a bit odd
19:11:03  <glx> like if a timer was increasing memory usage on each interval
19:12:26  *** gelignite has quit IRC
19:13:03  *** Progman has joined #openttd
19:16:38  <TrueBrain> very well possible it is aioredis, as they are running something in a thread, which is unusual .. but I just started a server locally, and we will see what it does memory-wise :)
19:19:48  <andythenorth> 10-day-old curry has a real depth of flavour
19:19:56  <DorpsGek> [OpenTTD/aws-infra] TrueBrain pushed 1 commits to main https://git.io/JletK
19:19:57  <DorpsGek>   - Fix: coding formatting error found by "black" (by TrueBrain)
19:25:28  <TrueBrain> ah ... found one thing .. servers via master-server get more and more direct-ips in the game-coordinator
19:25:39  <TrueBrain> the master-server fails to notice it already announced the IP
19:25:44  <TrueBrain> and the game-coordinator doesn't use a set :P
19:26:08  <glx> ah that could be a valid cause
19:26:36  <TrueBrain> and we have a pretty set amount of servers online .. so that could also explain the very predictable jumps, I guess
19:26:51  <TrueBrain> @calc 1000000 / 90 * 15 / 500
19:26:51  <DorpsGek> TrueBrain: 333.3333333333333
19:27:03  <TrueBrain> it kinda fits
19:27:12  <TrueBrain> something to fix tomorrow at least, as it is wrong for sure :D
19:29:46  <TrueBrain> https://www.sqlite.org/fasterthanfs.html
19:29:47  <TrueBrain> lol
19:30:13  <TrueBrain> rather old article .. but somehow made it on Hacker News
19:31:23  <andythenorth> Apple use quite a bit of sqlite inside their own applications
19:31:39  <andythenorth> we tried using it as a cache at work
19:44:19  <TrueBrain> yippie, other magic of the GitHub actions repo we have is also doing its job correctly :D
19:47:08  <glx> <TrueBrain> https://www.sqlite.org/fasterthanfs.html <-- windows I/O are so slow
19:47:18  <TrueBrain> I remember git in the old days :P
19:47:47  <glx> and it's worse for mingw, because extra layer
19:48:47  <glx> yeah I remember git was very slow on windows
19:56:06  <TrueBrain> right ... poked Discord to see if people can test out the new network goodies
19:56:09  <TrueBrain> hopefully that works
19:57:18  <TrueBrain> we might want to do something similar on our website in a week orso
19:57:23  <TrueBrain> just to get some attention and some testing done :)
19:57:31  <TrueBrain> or just, you know, release 1.12.0-beta1
19:57:34  <TrueBrain> that would do the trick too :P
19:58:41  <glx> nightly may be blocking for some people, as it's easy to have a different version than others
19:58:42  <Rubidium> though probably not as much as 2.0.0-beta1 ;)
19:59:06  <TrueBrain> for 2.0.0 I would like to add 2 more things ...
19:59:10  <TrueBrain> not sure I have time to make that a reality :D
19:59:19  <TrueBrain> both not HUGE, but do need a bit of time :)
19:59:55  <TrueBrain> (presets and reworking of the main menu UI :P)
20:00:48  *** iSoSyS has joined #openttd
20:26:57  *** sla_ro|master has joined #openttd
20:34:16  *** nielsm has quit IRC
20:39:16  *** andythenorth_ has joined #openttd
20:45:14  *** EER has quit IRC
20:45:31  *** andythenorth has quit IRC
21:10:56  *** Samu has quit IRC
21:17:02  *** Tirili has joined #openttd
21:34:52  *** andythenorth_ has quit IRC
21:38:14  *** HerzogDeXtEr has quit IRC
21:54:56  *** andythenorth has joined #openttd
22:02:15  *** sla_ro|master has quit IRC
22:12:34  *** andythenorth has quit IRC
22:13:11  *** andythenorth has joined #openttd
22:16:19  *** andythenorth has quit IRC
23:04:38  *** Speedy` has quit IRC
23:39:48  *** Gustavo6046_ has joined #openttd
23:39:55  *** Gustavo6046 has quit IRC
23:39:55  *** Gustavo6046_ is now known as Gustavo6046
23:48:18  *** tokai has joined #openttd
23:48:18  *** ChanServ sets mode: +v tokai
23:55:06  *** tokai|noir has quit IRC
23:56:02  *** Gustavo6046_ has joined #openttd

Powered by YARRSTE version: svn-trunk