Minetest logo

IRC log for #minetest-dev, 2020-04-11

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
01:35 kilbith joined #minetest-dev
01:38 nephele_ joined #minetest-dev
02:17 ANAND joined #minetest-dev
05:30 search_social joined #minetest-dev
05:44 Icedream joined #minetest-dev
06:15 bigfoot547 joined #minetest-dev
07:02 bigfoot547 joined #minetest-dev
07:05 nerzhul joined #minetest-dev
07:06 QwertyDragon_ joined #minetest-dev
07:07 nerzhul hello, i will do a clone reimport from github on our gitlab repository to have proper registry (because the project on gitlab.com is old and there is a notice that the container registry cannot set the expiration policy and the project took 23 GB). Nothing special expected, i will do it in another project in the group then drop the old one and rename where ready
07:08 nerzhul ah fuck i cannot because gitlab know is already exists
07:14 nerzhul if someone wants to take a look on https://github.com/minetest/minetest/pull/9069
07:14 nerzhul #9069
07:14 ShadowBot https://github.com/minetest/minetest/issues/9069 -- Remove sound menu and show proper messages if trying to change sound volume when sound if disabled/unsupported by Wuzzy2
07:15 QwertyDragon joined #minetest-dev
07:23 erlehmann_ joined #minetest-dev
07:33 calcul0n joined #minetest-dev
07:56 calcul0n joined #minetest-dev
08:03 nerzhul i don't know if we have a regression but i'm unable to connect on localhost
08:03 nerzhul i should use my network card ip
08:39 ShadowNinja joined #minetest-dev
09:19 Beton joined #minetest-dev
09:27 nerzhul just finished #9633 PR, appart for the file move, as as thought i needed to cleanup a old 0.3 era factory code :)
09:27 ShadowBot https://github.com/minetest/minetest/issues/9633 -- Move SAO to dedicated files & drop the old now useless factory by nerzhul
10:03 mizux joined #minetest-dev
10:44 sfan5 nerzhul: I've heard of connecting to local IP not working from several people
10:44 sfan5 but never happened for me
10:44 sfan5 or hm
10:45 sfan5 maybe it's just the fact that localhost resolves to ::1 while MT doesn't listen on ipv6 by default
10:53 nerzhul sfan5: i tried with 127.0.0.1 too and same problem
10:54 nerzhul sfan5: a CI improvement, i think we should switch to it :) https://github.com/minetest/minetest/pull/9641
10:54 nerzhul i started with the luacheck at first
10:56 nerzhul #9641
10:56 ShadowBot https://github.com/minetest/minetest/issues/9641 -- Move luacheck to github actions by nerzhul
10:57 nerzhul i will publish some PR step by step, i think moving to GH actions like we are using gitlab CI is better than the old travis with no OS maintained
11:08 nerzhul sfan5 #9635 is finished ? if yes i added my approval, we can merge
11:08 ShadowBot https://github.com/minetest/minetest/issues/9635 -- CSM fixes/features by sfan5
11:10 sfan5 it's finished yes
11:12 nerzhul okay i merge unsquashed
11:13 sfan5 thanks
11:13 nerzhul if you have time #9633 is ready for a review, and works fine :)
11:13 ShadowBot https://github.com/minetest/minetest/issues/9633 -- Move SAO to dedicated files & drop the old now useless factory by nerzhul
11:13 nerzhul that may help servers which does many SAO creations/removal a little bit too
11:13 sfan5 hmm I was sure I added [CSM] to the commit message
11:13 sfan5 wonder why that disappeared
11:14 nerzhul it was in the PR but not the commits it seems
11:15 sfan5 oh I know why
11:15 sfan5 I reimported the commits using .patch files yesterday and that strips any [...] prefixes
11:16 Fixer joined #minetest-dev
11:18 sfan5 <nerzhul> sfan5: i tried with 127.0.0.1 too and same problem
11:18 sfan5 would be nice if you could debug it and find the cause sometime
11:23 nerzhul yep, i will try in a few time, take a time to look my PR, i'm finishing the GH actions PR to see it we can use clang-tidy and clang-format here too for the first version included :)
11:23 nerzhul what do you think  about gh actions sfan5 ? good move or i waste time ?
11:24 sfan5 haven't used them in any projects yet but from what I've seen they work just like travis
11:24 nerzhul yes with inclusion per step directly in the github ui, faster to review, and also more flexible because you are like gitlab, very close to the repo workflow
11:38 Krock rubenwardy: "Authorize ContentDB" "Repository webhooks and services   Read and write access". Why does it need write access?
11:40 Krock rubenwardy: and wouldn't it suffice to fetch the atomic feed daily?
11:40 Krock -ic
11:42 sfan5 merging #9639
11:42 ShadowBot https://github.com/minetest/minetest/issues/9639 -- Use TILE_MATERIAL_ALPHA for use_texture_alpha entity flag by GreenXenith
12:00 nerzhul nice thanks
12:00 nerzhul hello Krock, glad to see you
12:01 nerzhul sfan5: travis is better now on #9633 you should be able to put an approval :D
12:01 ShadowBot https://github.com/minetest/minetest/issues/9633 -- Move SAO to dedicated files & drop the old now useless factory by nerzhul
12:01 Krock welcome back nerzhul :)
12:01 Krock it's been a while
12:01 sfan5 hm still no logs
12:01 sfan5 but the build passes
12:02 nerzhul yeah it runs but logs are totally fucked off :(
12:02 sfan5 btw I answered your comment on #9617
12:02 ShadowBot https://github.com/minetest/minetest/issues/9617 -- Improve protocol-level receiving code by sfan5
12:03 nerzhul fine will take a look
12:18 Darcidride joined #minetest-dev
12:27 proller joined #minetest-dev
12:41 behalebabo joined #minetest-dev
12:43 rubenwardy [12:38:50] <cddc39Krock> rubenwardy: "Authorize ContentDB" "Repository webhooks and services   Read and write access". Why does it need write access?
12:43 rubenwardy To create a webhook. It's much easier to use webhooks than to add more state to CDB. The Git use is the most fragile part of CDB, the library is poorly documented and painful to work with
12:44 rubenwardy You can set it up manually
12:46 rubenwardy (making sure there's at least one op whilst ChanServ disappears)
12:47 rubenwardy I'm sick of dynamic typing
12:48 sfan5 rewrite contentdb in c++?
12:48 sfan5 ;)
12:49 Calinou #TeamMypy
12:49 Calinou it also has Final type hints so you can have immutable variables
12:55 Krock hmm... manual setup? will look into that
13:02 rubenwardy here: https://content.minetest.net/help/release_webhooks/
13:02 rubenwardy it's not that hard
13:02 rubenwardy make an API token in contentdb, make a webhook using it on github
13:03 rubenwardy nerzhul!  https://rwdy.uk/20ZRR.png
13:13 rubenwardy ContentDB also has an API: https://content.minetest.net/help/api/
13:15 Krock \o/ thanks
13:16 rubenwardy given the low adoption of webhooks, I will probably implement polling at some point
13:16 rubenwardy but I'd make it infrequent
13:20 Krock the manual GitHub instructions seem to be outdated? I can only find the path through Settings -> Developer settings -> OAuth application (? maybe?)
13:21 rubenwardy Krock:  https://rwdy.uk/W5OlD.png
13:21 rubenwardy are you in the project settings?
13:21 Krock well, I need to generate a secret first?
13:21 rubenwardy no the API token is from contentdb
13:22 Krock oh, gotcha. visiting the ContentDB profile, not the GitHub one
13:25 Krock https://i.imgur.com/dq0dVjc.png  I'd assume the generated security token would show up after "save"..
13:26 Krock but "Reset" is needed additionally
13:27 rubenwardy huh, it should show after the intial save
13:27 rubenwardy but only once
13:28 Krock pressing "Reset" is always required after creating a new token
13:28 rubenwardy so it is
13:28 rubenwardy I'll file a bug
13:31 Krock will push fix for #9634 in 15 minutes: https://krock-works.uk.to/u/patches/0001-Formspec-No-spec-ID-for-static-text-labels.patch
13:31 ShadowBot https://github.com/minetest/minetest/issues/9634 -- Text cursor used when hovering non selectable StaticText
13:32 Krock oh. DS-minetest also found the responsible lines
13:33 Krock rubenwardy: HTTP status code 403 > You are not allowed to access this page.
13:34 rubenwardy what's the json body?
13:34 rubenwardy double check the settings?
13:34 Krock https://pastebin.com/raw/Gj8vn5AK
13:34 rubenwardy oh, response
13:35 Krock plus { on top
13:35 Krock response from ContentDB: https://pastebin.com/raw/Lg2HKYXc
13:36 rubenwardy what URL are you using?
13:36 rubenwardy ohhhhh
13:36 rubenwardy damn it
13:36 rubenwardy nginx is replacing the JSON response
13:36 Krock could it be that the shown access token is outdated? so that the reset happens after showing?
13:36 rubenwardy I never tested the error JSON responses on the production server
13:36 nerzhul rubenwardy: lol, is this due to the fact i'm cleaning up gitlab useless issues ? :D
13:38 nerzhul sfan5 thanks for the review, i will just cleanup the trailing whitespace you showed me
13:38 rubenwardy Krock: I've disabled nginx hijacking the error messages
13:39 Krock {"error":"Invalid authentication","success":false}
13:39 Krock already better now
13:39 rubenwardy that means that it can't find a matching API token
13:40 nerzhul rubenwardy, Krock as you are here, what do you think about https://github.com/minetest/minetest/pull/9641/checks?check_run_id=578960535 ? :)
13:40 Krock I reset it just now, and again the same error after updating the webhook & force re-deliver
13:41 Krock nerzhul: isn't this the same as the TravisCI overview page, but now integrated into GitHub?
13:41 nerzhul yep it's directly in GH, no more external support needed, it's a little bit like gitlab ci but inside github
13:41 Krock oh! it's no longer TravisCI
13:41 nerzhul Travis is external to GH :)
13:41 nerzhul see the status on the build, you see the 3 linters status directly
13:41 nerzhul (in the PR view)
13:42 nerzhul i will merge #9633 after build pass in 1 hour i think due to travis slowness :D (fixed the whitespace sfan5)
13:42 ShadowBot https://github.com/minetest/minetest/issues/9633 -- Move SAO to dedicated files & drop the old now useless factory by nerzhul
13:42 Krock well, that was possible before as well.. except that it needs more mouse precision for clicking the status icon
13:44 rubenwardy ahhh I know the issue
13:44 rubenwardy the code requires you to limit tokens to a package, currently
13:45 rubenwardy code: https://github.com/minetest/contentdb/blob/master/app/blueprints/github/__init__.py#L88
13:45 Krock oh
13:45 rubenwardy GitHub doesn't give you the secret, it signs the payload
13:46 Krock and the payload must correspond 1:1 to a ContentDB entry
13:46 Krock probably
13:47 rubenwardy yeah
13:47 rubenwardy the payload should do that if the VCS repo field is set correctly
13:47 rubenwardy I need to make these errors more descriptive
13:47 Krock pushing patch
13:48 Krock done
13:48 Krock and the webhook works now
13:49 rubenwardy "Invalid authentication, couldn't validate API token. Make sure to limit it to a package"   is much more descriptive
13:49 rubenwardy although, "it"
13:50 rubenwardy so, to make it possible for you to have a token which supports multiple packages, I could do:       SELECT * FROM token WHERE package_id=package.id OR owner_id=package.owner.id;
13:51 Krock that would be very handy. saves effort.
13:51 Krock well, given that someone wants to grant access manually
13:52 rubenwardy I kinda want a fullback though - like select tokens with matching package ids, unless there are no results in which case select tokens with matching author ids and no package
13:54 Krock nerzhul: tbh I don't see much profit using the GitHub actions right now. Travis shows up all checks nicely, even though it's more complicated to get to the overview page
13:55 nerzhul oh no it's simpler :D
13:59 nerzhul https://github.com/minetest/minetest/actions you have that overview on the project side
13:59 nerzhul https://github.com/minetest/minetest/pull/9641/checks is the check on the PR side pretty simple (im learning how to use them properly, but it's not so hard
13:59 nerzhul and you have the test rendering done properly here: https://github.com/minetest/minetest/pull/9641
14:01 nerzhul regarding travis i think github actions will just overtake it, and travis will die slowly. I follow very very huge amount on GH project, and many dropped travis either for actions or for circle ci
14:02 Krock to add further dependencies on GitHub
14:02 nerzhul travis is already a GH dep
14:02 nerzhul :)
14:03 nerzhul you cannot run travis outside of gh, it's just a remove a deprecated tool :D
14:03 Krock so what are you going to do about the master builds?
14:03 kilbith joined #minetest-dev
14:04 nerzhul the master branch you mean ?
14:04 Krock oh. there's also an Actions tab for it
14:04 rubenwardy SELECT * FROM token WHERE package_id=package.id OR (package_is IS NULL AND owner_id=package.owner.id);     is what I want
14:04 nerzhul yeah it's the same than before, you have a build list, except now actions tab render everything
14:04 TC01_ joined #minetest-dev
14:06 rubenwardy will it show format suggestions prettily?
14:07 nerzhul maybe we can configure it, but currently it's the same rendering than travis with just the GH UI as far i see
14:07 nerzhul merging #9633
14:07 ShadowBot https://github.com/minetest/minetest/issues/9633 -- Move SAO to dedicated files & drop the old now useless factory by nerzhul
14:18 DS-minetest joined #minetest-dev
14:23 nerzhul honestly their integration is very fine, and unlike travis which has broken logs since this morning, it should be well supported as it's github and microsoft (and no i'm not microsoft fan, but GH customers are more than travis customers :D)
14:23 nerzhul See that nice compilation run :D https://github.com/minetest/minetest/pull/9641/checks?check_run_id=579124224
14:23 nerzhul i open a discussion on our private space on GH, feel free to comment
14:27 sfan5 travis is doing maintenance today afaik
14:30 nerzhul oh okay :D
15:08 calcul0n_ joined #minetest-dev
16:09 mizux joined #minetest-dev
16:10 proller joined #minetest-dev
16:15 kilbith joined #minetest-dev
16:21 nerzhul i just published another tiny MR to refacto #9642. I'm trying to restart a old project i got which split the responsibilities serverside on the ServerEnvironment and consorts to tinier objects
16:21 ShadowBot https://github.com/minetest/minetest/issues/9642 -- Reduce ServerEnvironment propagation by nerzhul
16:24 sfan5 nerzhul: if you're editing the pathfinder anyway, have you checked if it can work with Map instead of ServerMap?
16:25 nerzhul i didn't checked it yet, you think about CSM path finding ?
16:25 rubenwardy I dread pathfinding on the client
16:26 nerzhul CSM pathfinding can be fine for servers but risky in terms of pos sync
16:26 nerzhul i look if i can use the parent class
16:26 sfan5 perhaps but using Map instead of ServerMap is good either way
16:26 rubenwardy the plans for client-side scripting were to have a thread on the client
16:26 nerzhul yep you are right, looking
16:26 nerzhul running lua in a thread ? very big change, as same as a lua thread on the server
16:26 nerzhul many locking will occur i think no ? :D
16:27 rubenwardy csm is a big change
16:28 rubenwardy I think that the pathfinder should only be given a model for the map, which doesn't include any environment stuff. It's currently given access to the entire server environment, iirc
16:28 nerzhul oh yeah i know, but the threading on the lua part is very huge challenge, not sure it's a good idea, because you will need to make all core engine component safe, and the env is already threaded
16:28 nerzhul client is doing nothing currently in terms of load compared to server
16:28 rubenwardy in my non-mt game, the pathfinder is unit tested because its only requirement is a model
16:28 nerzhul i dropped the ServerEnvironment pointer in my PR, i have just a map and node def manager
16:28 rubenwardy if that were true, the client would be getting 60FPS
16:29 nerzhul sfan5, just changing the type is fine :)
16:29 Wuzzy joined #minetest-dev
16:31 nerzhul i pushed the change, feel free to use it on client now :D
16:31 nerzhul it was very easy change, it's why we need to refactor things sometime :D
16:32 rubenwardy thanks
16:39 rubenwardy Krock: I have literally no idea why contentdb isn't showing the tokens the first time, printf debugging shows that everything is set as it should be
16:40 rubenwardy to show the API token, ContentDB saves it in a session cookie that is then cleared
16:40 rubenwardy the session cookie is never set
16:40 rubenwardy (this is secure because it's only set at time of reset/creation and isn't used anywhere except for information purposes)
16:43 rubenwardy aha, I've found the problem
16:51 nerzhul #9622 requires another approval and #9642 is ready for people who wants pathfinder to run on client :D
16:51 ShadowBot https://github.com/minetest/minetest/issues/9622 -- Update API site build by pauloue
16:51 ShadowBot https://github.com/minetest/minetest/issues/9642 -- Reduce ServerEnvironment propagation by nerzhul
17:34 nerzhul hmm i just found some issues in our docker images if we really run it as non priv, discovered on my kube cluster
17:34 nerzhul we missed some chown on the data folders
17:38 Krock nerzhul: why is #9642 necessary for 5.3.0?
17:38 ShadowBot https://github.com/minetest/minetest/issues/9642 -- Reduce ServerEnvironment propagation (helps pathfinder to become client side friendly) by nerzhul
17:39 Krock the milestone should only be used for important features and bugfixes that ARE milestones.
17:39 Krock refactoring is nice to have and all, but not crucial.
17:39 nerzhul ah we use milestones like this, i thought it was used like any real project milestone to define what we push in the release :D
17:40 Krock the overview would get even worse if we did that IMO
17:40 sfan5 well everything goes into the release, so that would be redundant
17:40 nerzhul it's right we don't handle multiple branches like craziness, fine
17:40 nerzhul Krock, answered you on 9642
17:41 Krock huh? collisions
17:41 Krock I meant the pathfinder. objects could obstruct the path
17:41 nerzhul yep, but where are objects ?
17:41 Krock but that'll probably never be implemented
17:42 nerzhul in a future maybe, and objects are in ActiveobjectMgr, not in ServerEnvironment :)
17:42 Krock just thought to mention the possible ideas behind giving it access to SEnv
17:42 Krock right, so these would be passed separately to Pathfinder
17:43 nerzhul yeah, currently i'm looking for current code, if there is justification in the future, the people who will implement will be happy to do it, but currently i'm trying to reduce the highly reponsible objects transmission on the core to use only the needed areas
17:43 rubenwardy objects blocking the path doesn't really make sense
17:43 rubenwardy static ones maybe
17:43 sfan5 well how do you know if one is static?
17:43 Krock A) you don't.
17:44 rubenwardy which is why I'd prefer not to do that
17:44 rubenwardy you could add some group system to objects and say which ones should block the route
17:44 rubenwardy the pathfinder should not have access to the serverenvironment
17:44 sfan5 you could misuse armor groups for that
17:45 sfan5 but either way those issues should be discussing when someone actually implements it
17:45 rubenwardy it should be given access to some map class, and an ObjectManager class - I forget if that still exists
17:45 rubenwardy yeah
17:45 Krock fair enough. best to limit the access to only what's currently needed in Pathfinder
17:45 rubenwardy also, I'm very thankful for Wuzzy's pathfinder fixes. It's made Conquer a lot easier to implement
17:46 Wuzzy what is Conquer
17:46 Krock a mod probably
17:46 rubenwardy https://content.minetest.net/packages/rubenwardy/conquer/
17:46 nerzhul anyway it's outside of my refactor PR :D i think it can be merged as implement crazy pathfinding using other objects collision is not for tomorrow :D
17:46 rubenwardy !title https://www.youtube.com/watch?v=jB_sFZGRUgw
17:46 ShadowBot rubenwardy: Error: That URL appears to have no HTML title within the first 4KB.
17:46 rubenwardy rude
17:46 nerzhul rubenwardy gg for the concept it was nice
17:47 sfan5 review reminder: #9621 #9617 #9586
17:47 ShadowBot https://github.com/minetest/minetest/issues/9621 -- Update wireshark dissector by sfan5
17:47 ShadowBot https://github.com/minetest/minetest/issues/9617 -- Improve protocol-level receiving code by sfan5
17:47 ShadowBot https://github.com/minetest/minetest/issues/9586 -- Implement DPI scaling for Windows by sfan5
17:47 sfan5 (it must be terrible for other people who have their PRs left lying around, luckily I have the luxury to require one less review)
17:48 nerzhul haha
17:50 nerzhul approved the DPI PR and answered you on my refacto pr
17:59 nerzhul merging #9586 & #9642
17:59 ShadowBot https://github.com/minetest/minetest/issues/9586 -- Implement DPI scaling for Windows by sfan5
17:59 ShadowBot https://github.com/minetest/minetest/issues/9642 -- Reduce ServerEnvironment propagation (helps pathfinder to become client side friendly) by nerzhul
18:01 nerzhul the wireshark pr is quite heavy to read, let me some time :D
18:03 sfan5 sure
18:03 sfan5 I don't think such helper scripts need much code review though, the code is there and it works
18:05 nerzhul yep just to be sure we don't miss a packet
18:05 nerzhul #9622 requires a second review and sounds fine to me too
18:05 ShadowBot https://github.com/minetest/minetest/issues/9622 -- Update API site build by pauloue
18:06 sfan5 can gitlab show me if that pr works? or does it not run the docker builds?
18:06 kilbith joined #minetest-dev
18:09 sfan5 I guess I can just run it locally
18:12 sfan5 yep worked
18:14 nerzhul it builds the minetest server image but that's all
18:38 nerzhul i just published #9643 to fix our docker image build to work as non root too, not just as root
18:38 ShadowBot https://github.com/minetest/minetest/issues/9643 -- Fix docker image priv when running container as non root by nerzhul
18:38 nerzhul (works on any k8s cluster with strong PodSecurityPolicies enabled)
18:38 nerzhul and merging #9622
18:38 ShadowBot https://github.com/minetest/minetest/issues/9622 -- Update API site build by pauloue
18:40 kilbith joined #minetest-dev
18:44 sfan5 merging game#2568 in 5m
18:44 ShadowBot https://github.com/minetest/minetest_game/issues/2568 -- Add Basic Fire Description by IFRFSX
18:58 behalebabo joined #minetest-dev
19:05 DS-minetest joined #minetest-dev
19:31 sfan5 pushing http://sprunge.us/O7MQud?diff (trivial) in 5m
19:36 Miner_48er joined #minetest-dev
19:44 Krock What's the status of 5.2.0 for F-Droid? Are there issues?
19:46 sfan5 there's definitely this https://github.com/minetest/minetest/pull/9410
19:47 Krock that's only Android 10(+)
19:47 sfan5 sure but it's an issue
19:48 Krock ikr. also broken in 5.1.0, but there's a build on F-Droid
20:29 rubenwardy merging #9378 in 10
20:29 ShadowBot https://github.com/minetest/minetest/issues/9378 -- Formspecs: Add state-selection to style elements by Df458
20:40 rubenwardy done
20:58 rubenwardy and #9480 in 10
20:58 ShadowBot https://github.com/minetest/minetest/issues/9480 -- Image waypoints & waypoint improvements by appgurueu
21:00 appguru joined #minetest-dev
21:10 nerzhul joined #minetest-dev
21:22 appguru When is MT 5.3 planned?
21:23 appguru I. e. how much time do I have left if I want to get something merged?
21:23 appguru (which I still have to work on)
21:24 rubenwardy sooner than usual. We want to fix Android pretty soon
21:24 rubenwardy so more like 2-4 months rather than 6
21:24 rubenwardy but that's a guess
21:25 appguru Ah fine. That's enough time :)
21:29 DS-minetest how can I increase the chances that one of my PRs is merged before the next release?
21:29 rubenwardy mention it here
21:31 rubenwardy scroll_container is next on my list
21:31 appguru is it a scrollable container for formspec elements?
21:31 rubenwardy yes
21:31 appguru because that'd be high priority
21:32 rubenwardy it also adds the ability for elements to be set a parent, which is important to adding structure
21:32 DS-minetest 😀 yey
21:33 DS-minetest 🤔 normal containers should also set parent in new fs versions
21:34 nerzhul by fixing android you mean having the android PR merged as soon it's ready ?
21:34 rubenwardy yeah, that PR needs to be merged
21:34 rubenwardy and then we need time to fix any issues that arise
21:35 nerzhul i will try to make effort to review it as soon as we have the next PR from monte48 with the two core changes he ported
21:37 rubenwardy I have two qualms with the Android mega PR: the dependency infrastructure relies too much on Monte's presense (I want it to be an automated pipeline), and the asset extracting is no longer hash based (which is good) but there should be a way to "factory reset" it.  Like, say the native app crashes before you reach the menu, it should offer to clean and re-extract the share data. Including wiping settings
21:38 rubenwardy it's so much easier to build though
21:38 rubenwardy all it takes it opening it in Android Studio
21:39 appguru Hmm
21:39 appguru I'm fine with the dependency infrastructre
21:39 appguru But the builds don't work
21:40 rubenwardy you're fine with binaries in a git repo with no instructions how to build new versions ofthem?
21:40 appguru No, I built myself
21:40 appguru It uses Gradle.
21:40 rubenwardy _dependencies_
21:40 appguru Gradle is a proper build system.
21:40 rubenwardy I know, I like gradle
21:40 rubenwardy my problem is with the dependencies
21:40 appguru But yeah, haven't investigated the depdencies much
21:40 appguru will have a second look
21:40 rubenwardy github.com/minetest/minetest_android_deps_binaries
21:41 rubenwardy https://github.com/minetest/minetest_android_deps_binaries
21:41 kilbith joined #minetest-dev
21:41 appguru also, current infrastructure is still relying on HTTP and therefore vulnerable
21:41 appguru which I consider rather scary
21:41 appguru (fetches some files from sfan5's website over HTTP)
21:41 sfan5 >the dependency infrastructure relies too much on Monte's presense (I want it to be an automated pipeline)
21:42 sfan5 this can (and should be) fixed afterwards
21:42 rubenwardy exactly
21:42 rubenwardy both of those qualms can be fixed afterwards, which is why some time is needed after merging that PR
21:43 rubenwardy DS-minetest: did some testing of scroll_container, couldn't find any issues. I'll go over the code in the next few days, I've looked at enough code for today :)
21:44 DS-minetest thanks!
21:44 appguru I'll probably go over the Android PR
21:45 appguru And retest
21:45 * DS-minetest just rebased the scroll container PR again btw.
21:45 rubenwardy tip: if the extraction fails, it won't try again. The extraction can fail if the build is incorrect (this happened to me because of a bug with the archive script)
21:47 appguru hmm
21:47 appguru something's failing :P
21:47 appguru it used to work
22:01 p_gimeno <appguru> also, current infrastructure is still relying on HTTP and therefore vulnerable
22:01 p_gimeno vulnerable to what?
22:01 appguru MITM attacks
22:02 nerzhul MITM from github to sfan5 server :p
22:02 nerzhul honestly we should just automate lib builds using gitlab ci or github actions, that's all
22:03 appguru acessing some weird http resources is not good practice
22:03 appguru and is prone to be insecure
22:03 appguru i agree, it's probably not very likely
22:03 appguru but still... afaik sfan5's server even serves https, so why not add that tiny "s" ?
22:04 p_gimeno I the people who are able to set up a MITM attack between those, can probably also obtain a legit looking certificate for https for the same domain
22:05 sfan5 is this about the travis script or ..?
22:05 p_gimeno I think* the people
22:11 calcul0n joined #minetest-dev
23:13 rubenwardy trivial bug fix:
23:13 rubenwardy !title https://github.com/rubenwardy/minetest/commit/3833396cfad7289b8c38ab64b69885b409d993ce
23:13 ShadowBot rubenwardy: Fix delayed error message in start game tab · rubenwardy/minetest@3833396 · GitHub
23:13 rubenwardy there was a missing `return true` that meant the error dialog wasn't shown
23:13 rubenwardy there was also duplicated code that made that possible
23:14 sfan5 hm it appears to be impossible to switch commit views to whitespace-ignoring diffs
23:15 rubenwardy works for me? https://rwdy.uk/8Utmq.png
23:15 rubenwardy this also moves the "update last game" out of the enable_server check
23:15 rubenwardy I don't know why that was there
23:15 sfan5 ther is no such button
23:16 rubenwardy oh, I'm using github enhanced
23:16 rubenwardy for firefox
23:16 rubenwardy so it's probably that extension
23:16 sfan5 well diff looks good to me
23:16 rubenwardy cool, I'll merge it in 10
23:16 rubenwardy that bug was annoying me so much
23:17 rubenwardy also, gamedata needs to go - global state sucks, it should be an argument to core.start
23:21 erlehmann joined #minetest-dev
23:36 rubenwardy merging #9373 in 10
23:36 ShadowBot https://github.com/minetest/minetest/issues/9373 -- Play 'player_jump' when player jumps by Wuzzy2
23:37 sfan5 [Mario sound]
23:37 rubenwardy ikr
23:51 sfan5 merging #9566
23:51 ShadowBot https://github.com/minetest/minetest/issues/9566 -- Update copyright in README to 2020 by testman42

| Channels | #minetest-dev index | Today | | Google Search | Plaintext