Time Nick Message 05:24 while does anyone have a one-line command to install the necessary packages to build minetest on debian 10.6? 05:53 MTDiscord i think there was a command for it floating about in the forums? 05:54 MTDiscord while ^ 16:34 sfan5 how does it look with maybe dedicating some time to release this evening? 16:34 sfan5 if we can get #10752 merged and #10987 fixed (likely trivial) we should proceed with the release asap 16:34 ShadowBot https://github.com/minetest/minetest/issues/10752 -- Allow overwriting media files of dependencies by Desour 16:34 ShadowBot https://github.com/minetest/minetest/issues/10987 -- Minimap modes aren't translated 16:34 sfan5 it's taken way too long already 16:39 rubenwardy sure 16:45 * rubenwardy attempts to find apk signing script 16:47 sfan5 "${ANDROID_HOME}/build-tools/some.version.number/apksigner" sign --ks path/to/the.jks path/to/the.apk 16:49 rubenwardy jarsigner, but yeah 16:50 rubenwardy jarsigner and a zipalign 16:50 rubenwardy I wish it would just use Android's builtin tool 16:50 sfan5 i'm pretty sure jarsigner is for bundles 16:50 sfan5 and I haven't ever needed zipalign 16:52 rubenwardy the commands nerzhul gave me was: 16:52 rubenwardy jarsigner -verbose -digestalg SHA1 -sigalg MD5withRSA -keystore path/to/keystore-minetest.jks build/outputs/apk/Minetest-release-unsigned.apk Minetest 16:52 rubenwardy /opt/android-sdk/build-tools/25.0.1/zipalign -v 4 build/outputs/apk/Minetest-release-unsigned.apk Minetest-release.apk 16:52 rubenwardy Using Android Studio to sign builds doesn't work, it says "Keystore was tampered with or the password was incorrect" 16:54 rubenwardy perhaps the JKS needs to be converted 16:59 rubenwardy aha 16:59 rubenwardy using keytool, it tells me the format 16:59 rubenwardy I'll try converting it 17:03 rubenwardy (keytool -list even handly gives me the command to upgrade the format) 17:05 sfan5 the one I've used apksigner with is "Keystore type: JKS Keystore provider: SUN" too (I assume, since it also recommends converting it) 17:06 rubenwardy Android Studio is still building, but it's gotten further than last time I tried to sign a build within it 17:11 rubenwardy build succeeded, I'm shocked. Time to test 17:14 rubenwardy Ok, that seems to work 17:15 rubenwardy nice, I'm glad I spend the 15 minutes doign this 17:16 rubenwardy rc2 being a dumb tag, probably late for that https://github.com/rubenwardy/minetest/releases/tag/5.4.0-rc2 17:20 nerzhul cool great job rubenwardy, we can now build from android studio ? 17:20 rubenwardy yeah 17:21 rubenwardy the command is: keytool -importkeystore -srckeystore ./build/android/keystore-minetest.jks -destkeystore ./build/android/keystore-minetest.jks -deststoretype pkcs12 17:21 rubenwardy I'm not sure this should be commited or if the jks should be removed - does this give more info to cracking the keystore? 17:21 nerzhul okay, not so hard :) 17:21 nerzhul we should commit the keystore 17:21 nerzhul as there is creds, we don't care 17:21 nerzhul it permits to verison it 17:21 rubenwardy it's really bad security, it never should have been commited 17:21 nerzhul version* 17:21 nerzhul it's not a problem on my side 17:21 nerzhul as it's double password protected 17:27 rubenwardy nerzhul: https://monitor.rubenwardy.com/d/9TgIegyGk/ctf 17:27 rubenwardy been working on a Lua API for metrics and some more engine metrics, backported from buck's monitoring mod 17:27 nerzhul nice 17:27 nerzhul i don't remember, we have prometheus exporter directly on MT ? 17:27 rubenwardy yeah 17:27 rubenwardy you did that 17:27 rubenwardy lol 17:28 nerzhul i didn't remember if we built in in core or not 17:28 nerzhul or if it was only in my fork 17:28 nerzhul if i miss it, tell me when 5.4 is out 17:29 nerzhul on the graphics part you wanted my advise, i prefer the stable version 17:29 nerzhul anyway if you want to use irrlicht 1.9 and backports the opengl feel free, but it can be hard 17:29 nerzhul i suggest that if we do that, and the first version works, we remove all useless irrlicht features in order to ensure we have our own fork, 17:30 nerzhul and move our own irrlicht code from MT to this new shiny library 17:30 nerzhul and i'm not sure about the SDL part as some talk about concerns on performance, and we need to ensure it works on android 17:30 nerzhul and also test the command lag 17:31 rubenwardy SDL does work on Android, and it would have to be really badly written to perform badly given we'd using it just for input and window management 17:33 ronoaldo Hey all! It looks like the form input does not handle special characters like "ã", which require you to type '~' followed by the 'a' character. Is this a know issue or something new? 17:34 ronoaldo I'm testing the portuguese translations I've made with a custom build from 5.4.0-dev but I'll also try from master/5.30 stable. 17:34 rubenwardy OS? 17:35 ronoaldo Debian Linux testing. 17:35 ronoaldo https://i.imgur.com/h3w6goz.png 17:35 ronoaldo This is what I got when testing a search from inventory. Now that I have translated, some results won't show up... 17:38 ronoaldo Same thing from 5.3.0 shipped from Debian repo: https://i.imgur.com/nXczSZv.png 17:40 nerzhul rubenwardy i agree but we need a strategy 17:41 ronoaldo looks like this issue: https://github.com/minetest/minetest/issues/302 17:42 nerzhul maybe irrlicht switch + sdl is intended to be a 6.0 release then, as it's a very major more 17:42 nerzhul move* 17:42 sfan5 ronoaldo: yeah I'd say that's a known issue 17:43 ronoaldo sfan5: thanks! i'll add a comment to it and watch it then. 17:43 sfan5 though specifically the creative inventory could handle that better by showing results for "algodão" when you enter "algodao" 17:43 sfan5 (that'd be a thing for MTG) 17:43 ronoaldo hmm, I see, so let me check it out 17:44 ronoaldo well, it shows no result either... O.o 17:44 sfan5 I said it could, not that it does 17:44 ronoaldo oh ok sorry, misinterpreted your statement. 17:45 ronoaldo from the issue i believe there is no solution because it is from upstream, correct? 17:46 sfan5 yes 17:46 sfan5 but this might be solved relatively soon since we're making our own modifications to Irrlicht for the next version 17:52 ronoaldo cool! thanks :) 19:10 Krock Q: how can I force a locale update? 19:11 sfan5 Q: what do you mean by "locale update"? 19:11 Krock re-generating po 19:11 Krock to generate compiled minetest.mo files 19:12 Krock it seems to happen incrementally when files are modified tho 19:12 sfan5 ./util/updatepo.sh 19:12 sfan5 if you're looking at #10987, shouldn't the strings already exist in the .po files? 19:12 ShadowBot https://github.com/minetest/minetest/issues/10987 -- Minimap modes aren't translated 19:12 Krock well I couldn't find them in the *.mo, that's why 19:13 Krock but they're in *.po 19:13 sfan5 cmake handles the po -> mo, I think it runs `msgfmt` for that 19:13 Krock sorry. I didn't explain it well. It's not about re-collecting the strings from the source, but to compile *.po files 19:14 Krock ashtrayoz: the problem there is that MTG doesn't strip whitespaces 19:15 Krock or whatever the issue is 19:16 Krock > make translations 19:18 Krock msgid "Minimap in radar mode, Zoom x%d" <-- in minetest.po but no such occurrence within minetest.mo ?! 19:21 sfan5 interesting 19:21 Krock # fuzzy <-- this is the cause for it 19:22 Krock apparently it's marked for removal or something? no idea 19:22 sfan5 fuzzy is used to mark translations that are not correct / need changes 19:22 sfan5 so those are not actually used, but persist in the file 19:22 Krock oh then it's fine 19:22 sfan5 if you remove the fuzzy it should apply the translation 19:23 Krock yes it does 19:25 Krock will push https://krock-works.uk.to/u/patches/0001-Minimap-Fix-default-label-translations.patch in 20 minutes if there are no objections 19:26 Krock another option would be to put gettext() at the end and keep N_, but I think this is a saner approach to not conflict with mod-specified strings 19:26 pgimeno ronoaldo: looks more like #6573 to me (been suffering it since I started with minetest) 19:26 Krock (fixes #10987) 19:26 ShadowBot https://github.com/minetest/minetest/issues/6573 -- Compose and dead key handling don't work on Linux 19:26 ShadowBot https://github.com/minetest/minetest/issues/10987 -- Minimap modes aren't translated 19:45 Krock pushing 19:47 kilbith bug? https://forum.minetest.net/viewtopic.php?p=390717#p390717 19:47 kilbith I wish people would get into the issue tracker directly 19:49 LoneWolfHT Is there an issue for no longer being able to hold down view range +/- buttons? 19:49 sfan5 yes 19:49 sfan5 numpad or regular ones? 19:49 sfan5 and please bisect it 19:50 Krock kilbith: I remember there were model animation changes but also reports that certain file formats do not work well so.. *shrug* 19:51 Krock if it were important, they shall file an issue on github 19:51 MTDiscord this artifact (lower left corner) is bugging me 19:51 MTDiscord https://cdn.discordapp.com/attachments/747163566800633906/813498260781924372/Bildschirmfoto_von_2021-02-22_20-51-13.png 19:51 MTDiscord https://cdn.discordapp.com/attachments/747163566800633906/813498260781924372/Bildschirmfoto_von_2021-02-22_20-51-13.png 20:03 sfan5 LoneWolfHT: #10969 20:03 ShadowBot https://github.com/minetest/minetest/issues/10969 -- plus/minus keys work inconsistently 20:03 kilbith Krock: meanwhile bug reports should not be ignored because it's not on github... 20:04 MTDiscord #10989 is a critical exploit 20:04 ShadowBot https://github.com/minetest/minetest/issues/10989 -- Wrong number of items in allow_metadata_inventory_put/take callbacks 20:04 MTDiscord most likely not introduced by my swap fix PR 20:05 MTDiscord it's just that the limiting to the actual movable stack size happens after the allow callbacks have been run 20:05 Krock ikr 20:05 Krock kilbith: yes sure but there's no information whatsoever to reproduce it. I'd answer the post if I could.... 20:06 MTDiscord /* Limit according to source item count */ if (count > list_from->getItem(from_i).count) count = list_from->getItem(from_i).count; 20:06 MTDiscord this should basically be moved up 20:06 Krock @a​ppguru please comment in the PR 20:06 MTDiscord indeed 20:07 kilbith Krock: you misread that post then 20:07 Krock oh 20:08 Krock -- Player animation blending 20:08 Krock -- Note: This is currently broken due to a bug in Irrlicht, leave at 0 20:08 Krock > sets it to 1 anyway 20:08 Krock > wonders why it's not working 20:09 kilbith then the engine should prevent it to be 1 20:09 MTDiscord Or just fix it in the fork. 20:10 Krock or document it 20:10 MTDiscord Fair 20:11 LoneWolfHT My issue is the fact that you could hold down the keys and it would repeatedly change, now you have to spam it to get it to change more than once 20:13 MTDiscord #10991 20:13 ShadowBot https://github.com/minetest/minetest/issues/10991 -- Fix late limiting of invalid inventory actions by appgurueu 20:15 Krock @a​ppguru optimization possible: recycle src_item 20:16 MTDiscord yeah I just saw savillis PR 20:16 MTDiscord it's better but it should remove the limiting code below I think? 20:17 Krock yes. makes sense. "count" is not used anywhere else, hence should produce the exact same behaviour as with the other PR 20:28 sfan5 LoneWolfHT: so not the one I linked? 20:29 sfan5 I can reproduce that too, but is it a regression? 20:29 LoneWolfHT Not the one you linked 20:34 kilbith there was someone on Discord having some troubles with model[] 20:34 kilbith please manifest yourself here so I can help you 20:36 LoneWolfHT I'd call it a regression because I could do it before in earlier versions 20:42 sfan5 mind bisecting the commit that caused it then? 21:48 LoneWolfHT https://github.com/minetest/minetest/commit/291a6b70d/ is the cause (Again ?) 21:55 sfan5 hmm makes sense, thanks