| Time |
Nick |
Message |
| 00:00 |
MTDiscord |
<greenxenith> Actually all the issues are partially solved by the fact that the client will already have the code available and can choose to use it even if for some reason the names/releases/versions mismatch due to a CDB fault |
| 00:05 |
|
Eragon joined #minetest |
| 00:05 |
MTDiscord |
<greenxenith> Another advantage of signatures: If a CDB release disappears, you are SOL for obtaining it from CDB using what the server provided. Storing every possible version of a mod on CDB is technically what it is meant for, but practically not what happens. If CDB only had to store signatures for every release made, it could keep them around permanently even if the relevant releases disappears or packages got shuffled |
| 00:06 |
MTDiscord |
<greenxenith> Neither method addresses the fact that mods may want to provide client code without needing to depend on an entirely separate package |
| 00:09 |
ireallyhateirc |
I'm late to the party but essentially the above means that installed locally games would come with CSMs bundled? |
| 00:10 |
MTDiscord |
<greenxenith> CSMs being bundled with games is a tiny part of the conversation |
| 00:10 |
|
illwieckz joined #minetest |
| 00:12 |
rubenwardy |
I'd probably disallow deleting releases of CSMs |
| 00:12 |
MTDiscord |
<greenxenith> That is true, but an arbitrary necessity that could be avoided with the server sending CSMs itself |
| 00:14 |
MTDiscord |
<greenxenith> Honestly I don't think CSMs should be separate packages. They should be handled like media: A mod can provide client code as files which are stored in the client cache as a hash, the hash of which ideally corresponding to a hash on CDB with a signature that ideally matches the signature the server sent with that file |
| 00:15 |
ireallyhateirc |
My point is that for singleplayer games installed locally it would be optimal to treat server and client code equally. Even without the "server-sent" part server code could assume that the client has the game installed (or ask them to install the game) |
| 00:16 |
MTDiscord |
<greenxenith> Letting mods provide CSMs as media would support that without using special situational logic. The client connects to a SP server, the SP server sends those file hashes, sends the files if the client doesnt have them cached, and everyone is happy. |
| 00:20 |
|
simon816 joined #minetest |
| 00:22 |
ireallyhateirc |
SSCSMs sound like a good solution for games you play without installing (services/servers proprietary or not) but anything else? |
| 00:23 |
MTDiscord |
<greenxenith> The system should work for both multiplayer and singleplayer so we arent creating unique logic where we dont have to |
| 00:29 |
ireallyhateirc |
if the outcome would be that server code and CSMs would be downloaded as a bundle and just run out of the box then sounds good |
| 00:29 |
MTDiscord |
<greenxenith> Sure, something like that |
| 00:33 |
ireallyhateirc |
ideally the process should be straightforward even without CDB (a clientmods subdirectory in the game's tree along with ordinary mods) |
| 00:33 |
MTDiscord |
<greenxenith> That is what is being discussed |
| 00:33 |
repetitivestrain |
erle: that's already been fixed, the sheep i mean |
| 00:35 |
repetitivestrain |
but sheep that were already upgraded will retain their new colors, unfortunately |
| 00:38 |
|
garywhite joined #minetest |
| 00:41 |
|
garywhite joined #minetest |
| 00:42 |
erle |
shaft without checking. i bet it's gun2 |
| 00:43 |
erle |
repetitivestrain the new colors are, unfortunately, white lol |
| 00:49 |
|
jaca122 joined #minetest |
| 00:49 |
shaft |
rifles is better |
| 00:54 |
MTDiscord |
<greenxenith> Client code as media with signatures |
| 00:54 |
MTDiscord |
<greenxenith> https://cdn.discordapp.com/attachments/749727888659447960/1318380653032640603/wErILeD4Z7PYAAAAABJRU5ErkJggg.png?ex=67621d2a&is=6760cbaa&hm=4432abbdfb88182c85fe0210a0bb1ede2e037c3fb915688581dbfb4dd155ace1& |
| 01:00 |
|
Bombo joined #minetest |
| 01:00 |
|
Bombo joined #minetest |
| 01:01 |
MTDiscord |
<greenxenith> This method keeps modules unfied (mods provide client code whether in a game or not), makes client code bundling consistent between games and mods, makes client code consistent between singleplayer and multiplayer, lets servers provide private client code if clients allow it, utilizes the existing connection to do the download, leverages existing media systems, can function entirely without CDB, doesnt rely on potentially missing or |
| 01:01 |
MTDiscord |
renamed releases, doesnt rely on the client knowing what version of code they have, and makes it easier to support third-party signature authorities. |
| 01:03 |
MTDiscord |
<vejou> make sure they cant be "trusted" mods though |
| 01:03 |
MTDiscord |
<greenxenith> Client code has its own sandbox |
| 01:04 |
MTDiscord |
<greenxenith> Whether or not the mod is trusted has no bearing on the client sandbox (for now) |
| 01:04 |
MTDiscord |
<vejou> dont want "Accept alternate version of FilesAPI" to give you a virus |
| 01:05 |
MTDiscord |
<vejou> so for now its not like that |
| 01:05 |
MTDiscord |
<greenxenith> Client mods wouldnt be able to do that in general |
| 01:05 |
MTDiscord |
<greenxenith> And if you are trusting a mod locally, obviously you are already letting it do whatever it wants |
| 01:05 |
MTDiscord |
<vejou> yes but the copy you have locally |
| 01:05 |
MTDiscord |
<vejou> not the copy the server requests you download |
| 01:06 |
MTDiscord |
<greenxenith> Once again, the client code is not trustable to leave the sandbox in any way, locally or not |
| 01:07 |
MTDiscord |
<vejou> as long as it stays that way then that isnt an issue |
| 01:22 |
MTDiscord |
<greenxenith> Missing flow component: Upon receiving the media file the client would need to rehash the file to verify that the hash the server sent was accurate |
| 01:30 |
MTDiscord |
<greenxenith> Also the part connecting to CDB is optional if the CDB is using a public key already on the client |
| 01:45 |
|
amfl2 joined #minetest |
| 01:46 |
MTDiscord |
<greenxenith> Please excuse my braindeadness; I should clarify: The signature should be sent with the hash, and the client should verify that the hashes match before even bothering to request the file (if not cached). If the signatures match, then it can request the file and compare the hashes to make sure everything is cool. |
| 01:54 |
MTDiscord |
<greenxenith> I guess the connection to cdb part is still valid for the OCSP-like check |
| 02:51 |
|
Blockhead256 joined #minetest |
| 02:51 |
Blockhead256 |
dear admin of the new dev wiki at dev.luanti.org |
| 02:51 |
Blockhead256 |
the site does not automatically choose HTTPS when I navigate to it in my browser by typing the domain name, can you please configure the web server to automatically upgrade to HTTPS? |
| 02:51 |
Blockhead256 |
thank you |
| 03:11 |
MTDiscord |
<warr1024> Green: I'd add OCSP stapling to the process so the client doesn't have to be able to hit CDB if the server can. |
| 03:13 |
erle |
i really would like to know what prompted this scramble, given that the simplest thing that could work – the jordan4ibanez CSM tab and letting CSMs be hosted on CDB – has not been done for years. |
| 03:14 |
erle |
and also that AFAIK none of the CSM API enhancements from lizzy or cora were ever upstreamed |
| 03:14 |
erle |
(i might be wrong on that last point, but then please tell me what *was* upstreamed) |
| 03:14 |
MTDiscord |
<greenxenith> Glad the CDB hosting didnt happen since that would have been terrible |
| 03:17 |
|
garywhite joined #minetest |
| 03:32 |
MTDiscord |
<greenxenith> Blockhead256: I've enabled HTTPS enforcement |
| 03:49 |
|
Blockhead256 joined #minetest |
| 05:00 |
|
MTDiscord joined #minetest |
| 05:50 |
|
est31 joined #minetest |
| 07:55 |
|
MTDiscord joined #minetest |
| 08:00 |
MTDiscord |
<greenxenith> This flow chart is starting to look more like a circuit diagram |
| 08:44 |
|
sid0 joined #minetest |
| 08:49 |
|
Blockhead256 joined #minetest |
| 08:50 |
|
MacroFaxSax joined #minetest |
| 09:55 |
|
tarsovbak joined #minetest |
| 09:58 |
|
gregon joined #minetest |
| 10:52 |
|
TheCoffeMaker joined #minetest |
| 11:50 |
|
Glaedr joined #minetest |
| 12:08 |
|
SFENCE joined #minetest |
| 13:24 |
erle |
<greenxenith> Glad the CDB hosting didnt happen since that would have been terrible |
| 13:24 |
erle |
wdym? |
| 13:35 |
|
Kimapr joined #minetest |
| 13:45 |
|
jadedctrl joined #minetest |
| 13:57 |
|
ireallyhateirc joined #minetest |
| 14:02 |
|
MacroFaxSax joined #minetest |
| 14:10 |
|
SFENCE joined #minetest |
| 14:41 |
sfan5 |
rubenwardy: did we ever fix this https://www.reddit.com/r/Minetest/comments/1hgb45v/ive_put_100_hours_in_my_world_and_there_doesnt/ |
| 14:42 |
rubenwardy |
I believe you may need a computer now Google has locked out totalcommander |
| 14:42 |
rubenwardy |
Is there an issue to support import/export of worlds? |
| 14:43 |
sfan5 |
probably |
| 14:43 |
rubenwardy |
And GitHub unicorn |
| 14:43 |
|
jordan4ibanez joined #minetest |
| 14:44 |
sfan5 |
but as far as I know you can't read app private data from a pc either |
| 14:44 |
sfan5 |
unless it's in external storage maybe? |
| 14:45 |
rubenwardy |
https://dev.luanti.org/accessing-android-data-directory/#through-adb-needs-computer |
| 14:46 |
rubenwardy |
https://github.com/minetest/minetest/issues/10600 |
| 14:46 |
sfan5 |
ah yea that would be external storage |
| 14:46 |
sfan5 |
because adb can't read /data/data/net.minetest.minetest/... |
| 14:47 |
sfan5 |
might even work by attaching your phone to a PC, without adb |
| 14:51 |
|
mrkubax10 joined #minetest |
| 14:58 |
ROllerozxa |
fwiw I heard that there was a new workaround on some vendor ROMs where if you simply put a zero-width joiner in the "data" folder name for a path to the scoped storage it will just bypass it |
| 14:59 |
ROllerozxa |
(I'm not kidding) |
| 14:59 |
MTDiscord |
<rollerozxa> https://cdn.discordapp.com/attachments/749727888659447960/1318593482457219155/image.png?ex=6762e360&is=676191e0&hm=e2e52ddfb897e01a9860c232d41db194110a9a4bc6b84f9c7c9b3367c227398c& |
| 15:00 |
ROllerozxa |
I have not been able to reproduce it on my nokia phone though |
| 15:01 |
ROllerozxa |
but, I do still believe that if you connect your phone through regular MTP to your computer it should be accessible there? I just happened to use ADB instead when writing the instructions which may not be as friendly to non-technical users |
| 15:01 |
|
chilledfrogs joined #minetest |
| 15:02 |
|
Desour joined #minetest |
| 15:05 |
sfan5 |
someone should adjust that article (not me) |
| 15:11 |
celeron55 |
at least on my android 14 phone i can view and copy stuff under Android/data/net.minetest.minetest by just connecting the phone via USB as an MTP device (i.e. set "Use USB for" to "File transfer" in settings. i have to set this every time i plug it in) |
| 15:33 |
|
ireallyhateirc joined #minetest |
| 16:11 |
|
Glaedr joined #minetest |
| 16:21 |
lmat |
I grew some trees underground. Feeling like a goat. |
| 16:44 |
jordan4ibanez |
How does cmake work oh my goodness |
| 16:45 |
ROllerozxa |
it works very great, thank you for asking |
| 16:46 |
Desour |
lmat: ah yes, underground goats planting trees. I like that conspiracy! |
| 16:46 |
jordan4ibanez |
local man can't get cmake to build one file |
| 16:50 |
jordan4ibanez |
I forgot to capitalize the M in CMakeLists.txt lol |
| 16:52 |
SwissalpS |
UNIX isn't dOs |
| 16:54 |
sfan5 |
CakeLists.txt |
| 16:58 |
Desour |
the CakeLists.txt are a lie |
| 17:09 |
|
Verticen joined #minetest |
| 17:09 |
|
SFENCE joined #minetest |
| 17:14 |
|
jaca122 joined #minetest |
| 17:17 |
MTDiscord |
<luatic> i mean, it is a recipe. just not for a cake. |
| 17:19 |
SwissalpS |
CBakeList.txt |
| 17:41 |
|
Desour joined #minetest |
| 17:52 |
|
fluxionary joined #minetest |
| 18:01 |
jordan4ibanez |
cmake is blowing up with clang and modules nooo |
| 18:05 |
lmat |
I'm running luantiserver; and I get, "To run a Luanti server, you need to select a game using the '--gameid' argument." fair enough. Where do I get a valid gameid? I tried "1" and it says, "ServerError: Supplied invalid gamespec...Game "1" not found". It gives a link: "Check out https://content.luanti.org for a selection of games to pick from and download." I go there, but there are no "gameid" values I |
| 18:05 |
lmat |
think... |
| 18:06 |
ireallyhateirc |
isn't gameid just the name of the game? |
| 18:06 |
ROllerozxa |
the gameid is the ID of the game, usually the name of the folder |
| 18:07 |
ROllerozxa |
luanti is just an engine, so you will need to download a game alongside to run a server |
| 18:17 |
MTDiscord |
<bastrabun> Use luantiserver --gameid list to see the available game IDs |
| 18:22 |
ireallyhateirc |
no way! |
| 18:23 |
|
SFENCE joined #minetest |
| 18:43 |
|
SFENCE joined #minetest |
| 18:50 |
|
kamdard joined #minetest |
| 18:53 |
|
SFENCE joined #minetest |
| 19:10 |
erle |
seems hopper clocks are broken in mineclonia ._. |
| 19:35 |
|
SFENCE joined #minetest |
| 19:35 |
|
YuGiOhJCJ joined #minetest |
| 19:39 |
lmat |
Thanks for the tips. I did cd games && git clone https://github.com/minetest/minetest_game; then used luantiserver --gameid minetest_game; and it appears to be working! |
| 19:40 |
|
sometalgoo joined #minetest |
| 19:40 |
|
garywhite joined #minetest |
| 19:48 |
lmat |
I'm using Dmobs on my server. In the logs, I see, " - dmobs is missing: mobs" How do I tell which "mobs" it's missing? |
| 19:48 |
lmat |
I dont't see anything called "Mobs" in content: https://content.luanti.org/packages/?q=mobs. (and if I did, I would be skeptical...) |
| 19:48 |
ireallyhateirc |
what kind of mod soup did you install? |
| 19:49 |
MTDiscord |
<rollerozxa> mobs redo provides mobs |
| 19:50 |
lmat |
rollerozxa: I have mobs_redo. ...oh I'm looking at world.mt and some six more mods were added " = false". I'm changing them to " = true" and hoping... |
| 19:52 |
lmat |
rollerozxa: Looks like it worked, thank you! |
| 19:52 |
lmat |
ireallyhateirc: I love IRC <3 I installed areas, dmobs, mobs_monster, mobs_redo, mobs_water, and tpad. |
| 19:52 |
lmat |
ireallyhateirc: Another six or so were added automatically (vide supra) |
| 19:54 |
lmat |
I tried to set an administrator by setting "name=lmat" in minetest.conf. It looks like it's not working. I didn't see a "name" property in minetest.conf.example...is there a new way to set an administrator? |
| 19:55 |
ireallyhateirc |
did you try doing "/grandme all" after logging in under that name? |
| 19:55 |
ireallyhateirc |
sry, "/grantme all"* |
| 20:03 |
lmat |
ireallyhateirc: Thank you <3 |
| 20:13 |
MTDiscord |
<csperson> grand. dad. fleentstones. |
| 20:25 |
|
imi joined #minetest |
| 20:43 |
|
SFENCE joined #minetest |
| 20:46 |
|
SFENCE joined #minetest |
| 20:51 |
|
SFENCE joined #minetest |
| 21:13 |
|
fluxionary joined #minetest |
| 22:06 |
|
Lunatrius joined #minetest |
| 22:12 |
|
Verticen joined #minetest |
| 22:45 |
|
Trifton joined #minetest |
| 23:34 |
|
panwolfram joined #minetest |
| 23:55 |
|
calculon joined #minetest |
