| Time |
Nick |
Message |
| 04:00 |
|
MTDiscord joined #minetest-dev |
| 04:49 |
|
fluxionary joined #minetest-dev |
| 04:58 |
|
calcul0n joined #minetest-dev |
| 05:25 |
|
YuGiOhJCJ joined #minetest-dev |
| 06:00 |
|
calcul0n_ joined #minetest-dev |
| 06:31 |
|
YuGiOhJCJ2 joined #minetest-dev |
| 08:22 |
|
Warr10246 joined #minetest-dev |
| 08:29 |
|
vampirefrog joined #minetest-dev |
| 08:46 |
|
Warr1024 joined #minetest-dev |
| 09:41 |
|
vampirefrog joined #minetest-dev |
| 09:45 |
|
copygirl joined #minetest-dev |
| 09:45 |
copygirl |
I have a question that might be related to Minetest engine internals and might even be a "flaw" in the engine. |
| 09:46 |
copygirl |
This is sort of as a player, but if I can help out Warr1024 with this that would be nice. |
| 09:48 |
copygirl |
So in NodeCore I build machines that depend on entities and nodes updating. That is, item entities fall, and settle into nodes, and optics will tick and do things to the nodes. This can be pretty timing sensitive, so any out-of-syncness of the entity and node updates will often cause my machine to do something before the item has settled to become a node in-world, and thus break. |
| 09:50 |
copygirl |
Unfortunately I'm unsure how NodeCore processes its node updates. Perhaps it's an internal buffer that gets filled with future upates that then get processed at certain intervals. I'll have to dive through the code, or maybe Warr1024 can fill me in. |
| 09:52 |
copygirl |
So I'm wondering, what can be done to ensure that entity updates run more syncronously with node updates? Is there an event or hook that runs after all entities have been processed? Maybe that would be a good place to check for if enough time has passed to run the next batch of node updates. |
| 09:54 |
sfan5 |
that's quite abstract. which engine feature is nodecore using to implement those "node updates" |
| 09:54 |
sfan5 |
? |
| 09:57 |
copygirl |
Looks like it uses `register_globalstep` to process optics updates. |
| 09:59 |
copygirl |
Just for reference, I *believe* this is where the optics updates hook into: https://gitlab.com/sztest/nodecore/-/blob/master/mods/nc_optics/api.lua#L289 |
| 10:00 |
copygirl |
(Currently seeing if nodecore's functions does anything different from minetest's builtin.) |
| 10:02 |
copygirl |
From what I can see it's just calling the minetest function, you just also provide a label. |
| 10:31 |
sfan5 |
that looks like it's keeping track of the nodes on its own and then the updates run every server step |
| 10:32 |
sfan5 |
(entity updates also run every server step) |
| 10:34 |
[MTMatrix] |
<localhost> engine update -> mod update, eg. update legacy mods for new engine requirements, i think so đ¤ |
| 10:37 |
[MTMatrix] |
<localhost> But yes, as sfan5 said, sync every server step, and server step valid for all versions of minetest đ¤ |
| 10:39 |
copygirl |
In what order is entity physics, entity events and global step being processed? |
| 10:43 |
copygirl |
I guess by entity event I mean `on_step` in particular. |
| 10:52 |
copygirl |
Is the tick rate in Minetest constant? Can it be made constant? |
| 10:53 |
copygirl |
Or rather, can the delta time between each step/tick/update or whatever you call them be made constant? |
| 10:58 |
[MTMatrix] |
<localhost> for example, try make dedicated_server_step = 1.00 #, and players (player controlled entities) will be looks like glitch moves, also maybe valid for entities. I maybe wrong, wait for devs answer |
| 10:59 |
[MTMatrix] |
<localhost> for example, try make dedicated_server_step = 1.00 #, and players (player controlled entities) will be looks like glitch moves, also maybe valid for entities (not player controlled). I maybe wrong, wait for devs answer |
| 11:17 |
copygirl |
Oh sorry by "delta time between each tick" I rather meant the delta time passed to those systems to be a constant value. If the server stutters for one reason or another, it would call update functions with that constant delta time until it's caught up. |
| 11:17 |
copygirl |
I believe that's what some modern time-sensitive games do, such as FPS games. To code the game to be delta-time variant, but to keept the delta-time constant between each update for consistency. |
| 11:18 |
copygirl |
And then you can adjust the tick rate for your server depending on how much resources you have available. |
| 11:20 |
erle |
please explain how you can have a constant delta if the load changes |
| 11:22 |
erle |
calling update functions with a constant delta until it's caught up only works up to a point |
| 11:22 |
rubenwardy |
fixed time steps are done independently from the frame rate |
| 11:25 |
erle |
rubenwardy what are you planning to use the thing for that you added where items have custom textures? |
| 11:25 |
copygirl |
erle: When you get to the point where you can't keep up anymore you probably want to start skipping updates and print a big warning message. |
| 11:26 |
rubenwardy |
I'm not planning to use it currently |
| 11:27 |
erle |
oh, that explains a lot |
| 11:27 |
erle |
(i wondered why comments on the PR about use cases were marked offtopic) |
| 11:28 |
erle |
rubenwardy anyway, it will be useful for handheld maps i think |
| 11:28 |
[MTMatrix] |
<localhost> also, steps can be skipped if server got perfomance issues? |
| 11:28 |
rubenwardy |
yeah. And enchanted tools or very dynamic items |
| 11:31 |
erle |
i guess that now, if a handheld map item has a base64 encoded bitmap stored in meta you can use the ${foo} syntax and [png texmod to generate a nice framed item (instead of having to include it thrice for hud, handheld texture, inventory texture) |
| 11:31 |
erle |
is that correct? |
| 11:35 |
copygirl |
rubenwardy: I gotta ask, was that in response to erle or me? |
| 11:36 |
erle |
i hope it was in response to me |
| 11:36 |
erle |
bc that's how i interpreted it and responded myself |
| 11:38 |
[MTMatrix] |
<localhost> [micro offtop] messengers without replay/quote support be like :3 |
| 11:40 |
copygirl |
I mean, in IRC you just add the name of the person you wanna respond to. Though yes, it does lack the specific message you're responding to. |
| 11:41 |
copygirl |
One thing I was curious about is whether this was reported on the issue tracker yet? |
| 11:42 |
copygirl |
Mismatches in the way entities and nodes update (the latter done on globalstep) causes otherwise reliable machines to break seemingly randomly so it's certainly a bug of sorts. And I'm not sure if it's one NodeCore can work around to fix itself. |
| 11:43 |
rubenwardy |
copygirl: to erle |
| 11:44 |
[MTMatrix] |
<localhost> minetest can run it? https://piped.thedroth.rocks/watch?v=FDiapbD0Xfg đ |
| 11:59 |
|
YuGiOhJCJ2 joined #minetest-dev |
| 12:07 |
|
YuGiOhJCJ2 joined #minetest-dev |
| 12:23 |
|
Desour joined #minetest-dev |
| 12:42 |
pgimeno |
<copygirl> "I mean, in IRC you just add the name of the person you wanna respond to. Though yes, it does lack the specific message you're responding to." - Depends on how lazy you are ;) |
| 12:43 |
copygirl |
True true |
| 13:08 |
[MTMatrix] |
<localhost> ofc |
| 13:45 |
|
json joined #minetest-dev |
| 13:48 |
json |
What time is the core developers meeting today? |
| 13:51 |
Krock |
in 1 hour |
| 13:52 |
json |
at 15 o'clock? |
| 13:52 |
json |
UTC |
| 13:56 |
json |
https://github.com/minetest/minetest/pull/12926 |
| 13:58 |
json |
Can you discuss whether to close this PR or not? |
| 13:58 |
json |
a lot of time has passed and no activity |
| 14:04 |
|
Desour joined #minetest-dev |
| 14:05 |
Krock |
sure |
| 14:05 |
|
json joined #minetest-dev |
| 14:06 |
json |
thank you! |
| 14:11 |
Krock |
It's still a draft and on the roadmap, thus I wonder why there's a particular interest in closing it. v-rob might want to resume the work when he's got time for it again. Also: do you happen to be known under another username? I'm asking in case you're a minetest organization member or related |
| 14:11 |
Krock |
there's smarter things to do than closing draft PRs |
| 14:50 |
|
srifqi joined #minetest-dev |
| 14:59 |
Krock |
something came up, will be around in 15 minutes. |
| 15:00 |
srifqi |
okay |
| 15:08 |
Krock |
back |
| 15:09 |
Krock |
general meeting ping Desour srifqi rubenwardy sfan5 ? |
| 15:09 |
Krock |
https://dev.minetest.net/Meetings#2023-10-15 |
| 15:10 |
*Â Desour |
is here |
| 15:11 |
Krock |
so today's plan is to enter the feature freeze |
| 15:11 |
|
proller joined #minetest-dev |
| 15:12 |
nrz_ |
oh i'm half around this time đ |
| 15:12 |
sfan5 |
copygirl: you can follow the broad sequence here https://github.com/minetest/minetest/blob/master/src/serverenvironment.cpp#L1328 |
| 15:12 |
sfan5 |
I'm online but not guaranteed I can take part entirely |
| 15:13 |
Desour |
is it realistic that all the remaining milestone points are completed within 2 weeks? |
| 15:15 |
Krock |
there's tough ones like #13849 |
| 15:15 |
ShadowBot |
https://github.com/minetest/minetest/issues/13849 -- Inventorylist: It's no longer possible to collect items from slots via drop and pickup doubleclick |
| 15:15 |
Krock |
I mean the double-click feature could be removed but that might undo an explicit feature of the PR |
| 15:16 |
Krock |
closed the set_list() bug. It works on master. At least one down. |
| 15:17 |
srifqi |
i wouldn't say that 13849 is a bug, though. it is _different_ than the prev version indeed, but not in a bad way. the descibed action can be done with click + drag for example |
| 15:17 |
Krock |
unfortunately grorp is not available today so that we could know what happened to #13885 to become a draft |
| 15:17 |
ShadowBot |
https://github.com/minetest/minetest/issues/13885 -- Clean up and improve mainmenu theme / game theme code by grorp |
| 15:17 |
srifqi |
i was going to ask about that, too |
| 15:17 |
Desour |
before I opened it, I've looked how hard is it to fix 13849. I haven't found an easy way, due to how doubleclicking is handled. but I could craft together a hacky fix |
| 15:20 |
Krock |
> Left-double-click to pickup up all of a selected item |
| 15:20 |
Krock |
well I suppose we have to undo this |
| 15:20 |
Krock |
fortunately there's shift click + listrings that make the movement easier |
| 15:22 |
Krock |
how about the MTG debundling? |
| 15:22 |
Krock |
#13850 improved on that a bit |
| 15:22 |
ShadowBot |
https://github.com/minetest/minetest/issues/13850 -- Notify users that they need to reinstall MTG if they still want it by grorp |
| 15:23 |
Desour |
I'm not sure if a revert is justified. I've added the bug label, but it's just changed input. my workflow of moving around items is not a documented feature. I mean, I would prefer a revert over no change, but idk how others feel about this |
| 15:24 |
Krock |
sorry I did not mean an entire revert but to undo that particular feature |
| 15:25 |
Krock |
theoretically the first click should not be counted if there's a selected item stack that will be or was dropped |
| 15:26 |
Krock |
(first click is where you are about to drop the dragged itemstack onto another stack) |
| 15:26 |
Desour |
(I didn't mean an entire revert either) |
| 15:27 |
Desour |
the doubleclick is not handled by the item move code, but by some other code in guiformspecmenu, that triggers special doubleclick events, iirc |
| 15:28 |
srifqi |
i tried to exclude that specific case, but hasn't succeed |
| 15:28 |
Desour |
anyway, I'll try to come up with a dirty fix. refactoring can happen later |
| 15:29 |
Krock |
also failed to find any relevant double-click code right now. anyway. I'd appreciate it if you could come up with a fix for that |
| 15:30 |
Krock |
otherwise maybe OrgelGames has an idea on how to fix it |
| 15:30 |
Krock |
-r |
| 15:31 |
Krock |
next important open point is #13799 as a requirement for #13818 |
| 15:31 |
ShadowBot |
https://github.com/minetest/minetest/issues/13799 -- Allow games (just MTG?) to enforce lock-step upgrades with versions |
| 15:31 |
ShadowBot |
https://github.com/minetest/minetest/issues/13818 -- Debundle Minetest Game by rollerozxa |
| 15:32 |
Krock |
we've discussed this in the last meeting and it turned out that preinstalled games cannot be detected by CDB |
| 15:32 |
Krock |
(or rather: their version cannot be detected) |
| 15:34 |
|
fluxionary joined #minetest-dev |
| 15:36 |
erle |
given that both issue #13794 and PR #13877 are more downvoted than upvoted and TGA is in use and the only way some things can be done right now from a mod perspective, is there any support for PR #13877 ? |
| 15:36 |
ShadowBot |
https://github.com/minetest/minetest/issues/13794 -- Remove TGA deprecation and put it in a âdo not deprecate againâ list |
| 15:36 |
ShadowBot |
https://github.com/minetest/minetest/issues/13877 -- Remove deprecation mark on TGA texture format in lua_api.md by corarona |
| 15:36 |
ShadowBot |
https://github.com/minetest/minetest/issues/13877 -- Remove deprecation mark on TGA texture format in lua_api.md by corarona |
| 15:37 |
srifqi |
Krock: i don't have any opinion on MTG debundle because i still can't get the idea or the effects it would be when it finally debundled, sorry |
| 15:38 |
erle |
srifqi that sounds like âthis should be deferred until the effects are clear, but mark it as deprecatedâ |
| 15:38 |
Krock |
I suppose the easiest solution for the version detection is to rely on package managers so that MTG is removed automatically on update. And in the case of Windows users that they use a new directory rather than replacing files |
| 15:38 |
erle |
by package managers you mean apt and stuff? |
| 15:38 |
Krock |
what else would I mean? |
| 15:39 |
erle |
i have no idea, which is why i am asking |
| 15:39 |
erle |
uh, who was the debian maintainer for minetest again? |
| 15:39 |
erle |
maybe other projects have experience with unbundling |
| 15:40 |
sfan5 |
"preinstalled games cannot be detected by CDB" can someone fill me in (or add a link) for the context? |
| 15:40 |
proller |
time to schedule merging https://github.com/minetest/minetest/pull/12142 ! |
| 15:42 |
erle |
proller how does âNOT FINISHED, HAVE BUGSâ and a locked PR that has commit messages like âfixâ, âfixesâ, âbetterâ and âcleanâ (LOL) have anything to do with anything rn? |
| 15:42 |
srifqi |
sfan5: i'm guessing it is the release field in game.conf. |
| 15:42 |
Krock |
sfan5: CDB adds specific fields to mod.conf or game.conf so that it can detect which version is installed |
| 15:43 |
erle |
Krock srifqi is anything keeping you from adding that field to the pre-installed version of minetest_game for exactly one minetest release? |
| 15:43 |
Krock |
so the concern was that people with RUN_IN_PLACE=1 builds would carry over their previous MTG version through various methods (replacing files or moving games/ and mods/) so that they're forever stuck with an old version |
| 15:43 |
Desour |
sfan5: that was discussed in last meeting, iirc. essentially, it's not possible to differentiate manual installations of mtg (e.g. with git instead of with contentdb) from shipped mtg installations (e.g. by windows users that decompress a zip to replace their old minetest installation) |
| 15:44 |
sfan5 |
ah, I see |
| 15:44 |
sfan5 |
can't really fix that can we? |
| 15:45 |
proller |
erle: 1. switching to 32 should be disabled by default for first time   2. it have no bugs in old 16bit mode  3. its should be merged with squash |
| 15:45 |
Krock |
we'd have to bundle "fake" CDB data for one or two Minetest versions including the bundled Minetest Game |
| 15:45 |
srifqi |
the alternative might be releasing MTG first, then bundling it (from CDB) with the ZIP download |
| 15:45 |
srifqi |
ah, wait, no |
| 15:45 |
erle |
Krock that sounds like a plan? |
| 15:46 |
sfan5 |
are we actually planning to remove MTG this release then? |
| 15:46 |
Krock |
relevant history https://irc.minetest.net/minetest-dev/2023-10-01#i_6118957 |
| 15:47 |
Krock |
sfan5: I've seen this request being brought up many times but I am not sure how many different people actually care about it |
| 15:47 |
erle |
proller query |
| 15:47 |
Krock |
tbh I do not care whether or not MTG is bundled forever |
| 15:48 |
Krock |
at least #9509 shows some interest |
| 15:48 |
ShadowBot |
https://github.com/minetest/minetest/issues/9509 -- Ship MTEngine with no games, choose and install from the Content Database (after main menu improvements) |
| 15:48 |
sfan5 |
I thought the consensus was to do it but only when |
| 15:48 |
sfan5 |
only a question of when* |
| 15:48 |
erle |
i don't see how it is a bad thing if the transition can be done properly. with the exception that minetest_game would have to be backwards compatible (but it is not a moving target much) |
| 15:49 |
sfan5 |
anyway the 'fake CDB data' approach could introduce new issues if we're not careful, e.g. MT trying to update data in a read-only place |
| 15:49 |
srifqi |
the additional 2 weeks was added just for having enough time to debundle MTG (from last meeting) |
| 15:49 |
sfan5 |
but really it doesn't seem like so much of an issue, package managers will remove old files and windows users are on their own anyway |
| 15:50 |
Krock |
assuming that package maintainers follow the project path by not bundling MTG |
| 15:50 |
Desour |
most package managers are probably reachable per e-mail |
| 15:50 |
sfan5 |
if we remove it from our cmake scripts it should disappear |
| 15:50 |
Krock |
but I'd support adding those fields by using build scripts, for example. |
| 15:51 |
sfan5 |
I'd put a big note into the releases in addition |
| 15:51 |
erle |
would special detection for .git folders for minetest game help anyone? |
| 15:52 |
Krock |
normies do not clone. you can ignore that case. |
| 15:52 |
srifqi |
bundled MTG does not have .git folder (at least on Windows) |
| 15:53 |
erle |
developers of games do have git cloned stuff in their source dirs and i assume at least some of them are here |
| 15:54 |
Krock |
sfan5: do you think debundling for 5.8.0 could work? |
| 15:55 |
Krock |
erle: those people (should) know what they're doing, thus not a concern. it's their responsibility to be up-to-date |
| 15:55 |
Krock |
I mean it also doesn't hurt to continue 5.8.0 the usual way |
| 15:56 |
sfan5 |
I think it could work |
| 15:56 |
sfan5 |
#13799 is not a critically important part |
| 15:56 |
ShadowBot |
https://github.com/minetest/minetest/issues/13799 -- Allow games (just MTG?) to enforce lock-step upgrades with versions |
| 15:56 |
Krock |
right |
| 15:57 |
Krock |
based on that, I don't see any further required dependencies for #13818 - or are there? |
| 15:57 |
ShadowBot |
https://github.com/minetest/minetest/issues/13818 -- Debundle Minetest Game by rollerozxa |
| 15:57 |
sfan5 |
correct |
| 15:58 |
sfan5 |
worst case some engine change does break MTG and users will have to manually check CDB and click the 'update' button |
| 15:58 |
sfan5 |
(are update notifications planned?) |
| 15:58 |
Desour |
actually, it should be enough to add a big note saying that mtg is no longer shipped with mt, and if users want an up-to-date mtg installation, they should remove their old installation to get it from contentdb. if someone manually upgrades minetest by downloading a file, the extra step of deleting the old mtg shouldn't be hard |
| 15:58 |
sfan5 |
you'd add that note to www.minetest.net/downloads/? |
| 15:58 |
|
fluxionary joined #minetest-dev |
| 15:59 |
Krock |
sfan5: it would be easiest to have those notifications done by CDB which has its own said problems |
| 15:59 |
Desour |
sfan5: yes, but only where it's relevant, i.e. at the windows section |
| 16:00 |
sfan5 |
Krock: a notification that a mod/game/txp installed from CDB has an update is what I meant, yes |
| 16:00 |
Krock |
however I think the chances for breaking MTG by engine changes is rather unrealistic |
| 16:01 |
Krock |
sfan5: don't those notifications already appear in the main menu? If they've installed MTG via CDB that's not a problem |
| 16:01 |
sfan5 |
no idea |
| 16:01 |
sfan5 |
anyway we can consider this point solved |
| 16:02 |
Krock |
okay they do yet not show up directly in the mein menu but rather in the CDB formspec |
| 16:02 |
Krock |
although I remember a PR that would add an asterisk and update indicator to the tab |
| 16:02 |
srifqi |
it hasn't been merged |
| 16:03 |
srifqi |
#13807 |
| 16:03 |
ShadowBot |
https://github.com/minetest/minetest/issues/13807 -- Add package update detection on Content tab by rubenwardy |
| 16:04 |
Krock |
okay that would be nice to have |
| 16:04 |
Krock |
not too important either, though. |
| 16:06 |
srifqi |
about the black screen on Android, i can't do anything because i don't have any device that has that problem |
| 16:07 |
Krock |
do we yet know which devices that are affected? or rather - which kind of hardware? |
| 16:07 |
erle |
which issue is it? |
| 16:07 |
Krock |
#13519 |
| 16:07 |
ShadowBot |
https://github.com/minetest/minetest/issues/13519 -- Android: Black Screen |
| 16:07 |
sfan5 |
graphics related problems on android are often unfortunately a "lol no idea" situation since every vendor cooks their own soup |
| 16:08 |
srifqi |
indeed, it heavily depends on the vendor |
| 16:08 |
Krock |
I could try it on my smartphone but that's hardly relevant due to being a dinosaur at this point |
| 16:08 |
erle |
Krock it might lead to new knowledge though |
| 16:08 |
Krock |
is there a way to find out whether shaders failed to compile? |
| 16:09 |
Krock |
in that case it might be possible to disable those automatically |
| 16:09 |
erle |
wouldn't that show up in debug.txt? |
| 16:09 |
sfan5 |
the log would show it |
| 16:09 |
Krock |
checking |
| 16:09 |
Krock |
ah okay |
| 16:10 |
|
json joined #minetest-dev |
| 16:10 |
Krock |
there's however no relevant log entry here: https://github.com/minetest/minetest/issues/13519#issuecomment-1562600342 |
| 16:11 |
sfan5 |
it would be useful if there was a way to produce a verbose log on Android |
| 16:11 |
sfan5 |
or wasn't that logged to logcat anyway? I don't remember |
| 16:11 |
srifqi |
all log level are outputted to logcat |
| 16:12 |
Krock |
did the Android build get removed from GitHub Actions? Those were pretty helpful to not compile it on my own |
| 16:13 |
sfan5 |
https://github.com/minetest/minetest/actions/runs/6484455037 still there |
| 16:14 |
Krock |
so it was smart enough to skip the build in commit 5e0f14266d42b5, which is why I was confused |
| 16:14 |
srifqi |
isn't there a list of files to be changed to trigger a build? |
| 16:15 |
Krock |
I'd assume that build system changes (e.g. CMakeLists.txt) is important enough. Such list would explain it, though. |
| 16:16 |
srifqi |
the issue creator said that it happens in 5.7.0, so we can just download from Play Store/GitHub release to test it |
| 16:17 |
sfan5 |
Krock: was probably forgotten since android now also uses cmake |
| 16:17 |
srifqi |
fyi, 5.7.0 can run fine on BlueStacks: |
| 16:17 |
MTDiscord |
<srifqi> https://cdn.discordapp.com/attachments/747163566800633906/1163148713963823235/version-5-7-0-on-bluestacks.png?ex=653e85a7&is=652c10a7&hm=9bf50aa5662d6ced1ab93c159800473b63d0c04250cf08dea11dd2ae6ce506ed& |
| 16:19 |
srifqi |
since it is not a crash, there might be no reports about it in Play Store dashboard |
| 16:20 |
sfan5 |
I'm sure you could see it in user reviews however |
| 16:20 |
erle |
regarding android, consider the following: https://github.com/gonetz/GLideN64/pull/2633 |
| 16:20 |
erle |
> Newer PowerVR devices need depth to be cleared every frame, otherwise most geometry is rendered behind the background. |
| 16:20 |
erle |
> It also needs fragment based depth to be disabled, otherwise geometry has various depth glitches. |
| 16:20 |
erle |
does this explain anything? |
| 16:20 |
erle |
i mean it is not a black screen, but the results are like this: https://github.com/gonetz/GLideN64/issues/2108#issuecomment-581099066 |
| 16:22 |
Krock |
Works on Mali-T624 |
| 16:22 |
Krock |
that's all I can tell |
| 16:23 |
|
srifqi1 joined #minetest-dev |
| 16:23 |
erle |
well this particular PowerVR stuff seems to have been an issue for a bunch of projects. like godot engine also has bug reports about rendering fails. something is weird about it ig. |
| 16:23 |
erle |
i mean the PowerVR GE8320 |
| 16:23 |
Krock |
in either case, that issue might be better placed in minetest/irrlicht |
| 16:24 |
Krock |
unless it turns out to be a shader compiling error |
| 16:25 |
|
srifqi1 joined #minetest-dev |
| 16:25 |
Krock |
we don't have enough information or devices to actually reproduce this issue, hence I am not sure what to do with it |
| 16:25 |
srifqi1 |
any shader error would result in a massive amount of errors in the debug.txt |
| 16:25 |
sfan5 |
close as can't reproduce/can't fix? |
| 16:25 |
srifqi1 |
shader compilation error* |
| 16:26 |
Krock |
sfan5: keep it open. it's an issue that needs to be fixed. But I'm not sure whether to keep it in the 5.8 milestone |
| 16:27 |
sfan5 |
IMO no |
| 16:28 |
erle |
the issue has someone who posted just now |
| 16:28 |
erle |
and can reproduce |
| 16:28 |
erle |
or so it seems. i asked them to come to #minetest-dev |
| 16:29 |
erle |
maybe one of you can help them to get the debug info out of the hardware |
| 16:29 |
erle |
(i can assist them with adb logcat, but i can not interpret the logs) |
| 16:30 |
srifqi |
ah, the device is not used anymore |
| 16:32 |
erle |
so with https://github.com/minetest/minetest/issues/13519#issuecomment-1763440866 being written, what's the next topic? |
| 16:34 |
sfan5 |
going for a walk now, afk |
| 16:34 |
Krock |
the other issues look rather minor |
| 16:34 |
erle |
what about https://github.com/minetest/minetest/pull/13877 |
| 16:35 |
Krock |
erle: no objections. can be merged within the feature freeze if it gets the approvals |
| 16:36 |
erle |
well, i was asking for approvals actually. since i don't want to do effort for nothing i'll basically stop work on everything that would be affected (i.e. every mod i made that does non-trivial things). |
| 16:38 |
Krock |
ok |
| 16:38 |
|
unbiquadium joined #minetest-dev |
| 16:39 |
Krock |
this meeting is again dragging on for quite a while. I think the topic update for feature freeze will have to wait on someone who's got the MinetestBot permissions for that |
| 16:39 |
Krock |
there's not too much important waiting from the milestone so I suppose it's about time for that |
| 16:40 |
Krock |
I also have to leave now; feel free to discuss the remaining topics if needed but from what I can see the most important ones are done now |
| 16:40 |
Krock |
thanks for participating |
| 16:41 |
srifqi |
thank you for organizing |
| 16:41 |
erle |
for reference, these are still open for the milestone: https://github.com/minetest/minetest/milestones/5.8.0 |
| 16:42 |
|
unbiquadium joined #minetest-dev |
| 16:43 |
|
unbiquadium joined #minetest-dev |
| 16:45 |
srifqi |
i'll leave for now, too |
| 16:45 |
unbiquadium |
if my question is ontopic, will the two finger stack splitting solved? |
| 16:46 |
unbiquadium |
in 5.8 i mean |
| 16:47 |
unbiquadium |
every client i found is based on 5.8 dev and its getting frustrating |
| 16:48 |
erle |
what issue? |
| 16:49 |
Desour |
unbiquadium: if you're referring to #13822, that's on the milestone and should be fixed till next release. you could try out the PR that already exists |
| 16:49 |
ShadowBot |
https://github.com/minetest/minetest/issues/13822 -- Regression in splitting inventory stacks on Android |
| 16:49 |
erle |
for future reference, these are the milestone things right now https://web.archive.org/web/20231015164259if_/https://github.com/minetest/minetest/milestones/5.8.0 |
| 16:49 |
MTDiscord |
<rollerozxa> it should be fixed with #13872, though I haven't tested it myself |
| 16:49 |
ShadowBot |
https://github.com/minetest/minetest/issues/13872 -- Formspec: Allow multiple mouse button click for touch-screen by srifqi |
| 16:49 |
erle |
13822 is on the list |
| 16:49 |
unbiquadium |
yes that |
| 16:49 |
erle |
for 5.8.0 |
| 16:49 |
erle |
so ig you are good unless that changes |
| 16:50 |
unbiquadium |
also, can i hope that anything similar to @grorp's fork add will be |
| 16:50 |
unbiquadium |
i mean virtual conteols android |
| 16:51 |
unbiquadium |
i played a LOT with it and i really like that  i can literally pvp with half limitations |
| 16:53 |
|
unbiquadium joined #minetest-dev |
| 16:57 |
unbiquadium |
forgive my english,just look at the screenshot, the leftclick and rightclick buttons are well requested https://github.com/minetest/minetest/issues/13229 |
| 17:04 |
sfan5 |
are we in feature freeze now? |
| 17:16 |
|
vampirefrog joined #minetest-dev |
| 17:25 |
Krock |
sfan5: last week's plan was to enter it today |
| 17:26 |
Krock |
however #13818 yet needs to be merged |
| 17:26 |
ShadowBot |
https://github.com/minetest/minetest/issues/13818 -- Debundle Minetest Game by rollerozxa |
| 17:33 |
|
Desour joined #minetest-dev |
| 17:37 |
Krock |
rubenwardy: how come Android does not remove MTG on update? |
| 17:37 |
sfan5 |
hopefully it actually deletes old files instead of overwriting, would be broken otherwise |
| 17:38 |
sfan5 |
might just be a special casing for 'games' tho |
| 17:38 |
rubenwardy |
There's no code to do so |
| 17:38 |
rubenwardy |
Minetest game won't be in the archive any more, so it won't touch that directory |
| 17:39 |
sfan5 |
so it has just been overwriting the old files all the time? |
| 17:39 |
rubenwardy |
I don't think so |
| 17:39 |
ROllerozxa |
...yep |
| 17:39 |
ROllerozxa |
it just extracts the assets zip into the data folder, it doesn't remove anything first |
| 17:39 |
Krock |
that's stupid. I thought Android also had some sort of package manager to keep track of the installed files? |
| 17:39 |
rubenwardy |
The android app copies the archive contents to the user data dir, there's no share directory |
| 17:39 |
rubenwardy |
We're not using any package manager for this |
| 17:40 |
rubenwardy |
This is just the android app extracting a .zip file |
| 17:41 |
rubenwardy |
It would be better to somehow mount the archive so the data inside could be used like a share location, without having to copy it |
| 17:41 |
Krock |
regardless on whether we keep MTG bundled or not - that approach of overwriting needs to be fixed |
| 17:42 |
rubenwardy |
Even if it did delete the folders first instead of overwriting, that wouldn't have helped us here |
| 17:43 |
Krock |
mount-zip (via fuse) on android 8) |
| 17:43 |
rubenwardy |
Unless it somehow remembered what files were extracted |
| 17:43 |
sfan5 |
sounds like we need an one-time special code that deletes MTG |
| 17:44 |
rubenwardy |
Yeah, based on version code of the app? |
| 17:44 |
Krock |
could we check the directory creation timestamp? |
| 17:44 |
sfan5 |
maybe also switch to separating share/user at this opportunity |
| 18:05 |
erle |
mods can't write to the game directory, right? only to world dir? |
| 18:08 |
MTDiscord |
<luatic> I think mods can write to their mod folder, at least during load time? |
| 18:08 |
erle |
i just want to figure out if anything of value could get lost |
| 18:09 |
MTDiscord |
<luatic> "The following functions, which are not available to CSM / SSM-only, allow read-only access to all mod directories and write access to the current loading modâs directory only while itâs loading (a handle with write access to a file within the mod directory can however be stored and used at a later time) and read & write access to the world directory excepting the worldmods and game subfolders" - |
| 18:09 |
MTDiscord |
https://github.com/minetest/minetest_docs/blob/master/doc/environment.adoc |
| 18:09 |
MTDiscord |
<luatic> this was how i documented this a while ago, hope it's still accurate |
| 18:09 |
erle |
yeah okay but is any mod in minetest_game doing it |
| 18:09 |
MTDiscord |
<luatic> in minetest_game? not that i'm aware of |
| 18:10 |
erle |
should i take this as âi have not looked and am making it upâ or âi would remember if that was the caseâ? |
| 18:43 |
MTDiscord |
<jordan4ibanez> Sfan5 is right, we should delete mtg |
| 18:48 |
|
appguru joined #minetest-dev |
| 18:49 |
ROllerozxa |
jordan4ibanez: working on it |
| 18:50 |
erle |
delete mtg, the delete system32 of minetest |
| 18:51 |
erle |
ROllerozxa so what's your opinion on the version hack thing |
| 18:51 |
ROllerozxa |
~~comparing it to system32 implies MTG is a critical component of minetest~~ |
| 18:51 |
ShadowBot |
ROllerozxa: Error: You must be registered to use this command. If you are already registered, you must either identify (using the identify command) or add a hostmask matching your current hostmask (using the "hostmask add" command). |
| 18:51 |
ROllerozxa |
registered to who |
| 18:51 |
erle |
~~what~~ |
| 18:51 |
ROllerozxa |
anyways erle: what version hack? |
| 18:52 |
erle |
ROllerozxa add a CDB version field to minetest_game for one or two releases |
| 18:52 |
erle |
i mean to the included version |
| 18:56 |
ROllerozxa |
manually adding a release field to the game.conf? |
| 18:57 |
ROllerozxa |
I admit I haven't fully caught up on the current discussion as I was away during most of the meeting |
| 19:01 |
erle |
âcontext is what scrolls away at the top of the windowâ ROllerozxa |
| 19:17 |
sfan5 |
does someone maybe have a mac device and can test irr#252 |
| 19:17 |
ShadowBot |
https://github.com/minetest/irrlicht/issues/252 -- Backport upstream commits by sfan5 |
| 21:13 |
|
proller joined #minetest-dev |
| 22:34 |
|
panwolfram joined #minetest-dev |
