Time Nick Message 00:36 rubenwardy Merging #13191 in 10 00:36 pgimeno https://github.com/minetest/minetest/pull/13191 -- Add jpeg and png libraries to the Fedora DNF install hint. by sofar 03:34 lhofhansl Can I get a quick review for #13225? So that we can close the blocker? Should be an easy review (review ignoring whitespace) 03:34 pgimeno https://github.com/minetest/minetest/pull/13225 -- Fix infinite viewing_range by lhofhansl 04:14 ashstar1248 Where is the stuff for parsing the mod.conf file? 04:16 ashstar1248 nvm found it 06:15 pgimeno has anyone pinged ShadowNinja about ShadowBot? 12:00 Zughy[m] Reminder that there is a meeting today 18:01 Krock Meeting time. who's available? 18:08 Zughy[m] Me? 18:20 Krock I'd say just two people are not a sufficient starting point ... 18:22 MTDiscord Well if youve got some spare time, #13189 is pretty simple and could use one more approval :] 18:22 pgimeno https://github.com/minetest/minetest/pull/13189 -- Add chat HUD flag by GreenXenith 18:23 Zughy[m] Krock yeah, I agree; I just want to say that this is the third meeting skipped in a row and that the problems listed in the meeting notes will only keep piling up. Also note that one point is still from Jude 18:24 Krock at least that one's postponed together with the release itself 18:24 Desour it's not the first time that most core devs are not available at meeting time. maybe it would make sense to announce the meeting more prominently already 1 or 2 days in advance 18:25 Desour (just a suggestion :=) 18:25 Krock github discussions will shut down in march so I suppose we'll fall back to IRC pinging 1 hour before the meeting 18:25 Desour :)* 18:25 Krock also I haven't seen much activity on those discussion posts anyway so it's not too big of a loss 18:26 MTDiscord Only team discussions are being deprecated, just to be clear between repo discussions 18:29 Krock yes that's what we used to announce engine-related meetings 18:30 MTDiscord Aha 18:32 Krock GreenXenith: shouldn't the flag be inverted? IIRC the hud_flags packet is set to 0 on older servers 18:33 Krock either that or you'll need a protocol version bump, I guess? 18:34 MTDiscord Why so, if the hotbar flag isn't inverted 18:34 MTDiscord (Which should usually be visible, even on old clients) 18:34 Krock actually nvm. it's implemented by delta-masking 18:35 Krock that's a neat trick 18:35 MTDiscord So it should work fine with a new server and old client? 18:36 Krock I meant the inverse 18:36 MTDiscord I suppose it should be tested (not home atm otherwise I could do that real fast) unless we're confident 18:36 Krock I'm currently at it 18:37 MTDiscord Ah, as in the client expecting information the server doesn't send 18:37 MTDiscord I get it 18:48 Krock it works correctly on older servers. what however does not work too well is the minimap that somehow make the game very laggy 18:48 Krock or rather: it results in stutter 18:50 MTDiscord That's not a result of the chat flag, right? 18:50 Krock entirely unrelated 18:50 MTDiscord Still wacky 20:06 Desour #13234 20:06 pgimeno https://github.com/minetest/minetest/pull/13234 -- Throw Hocroft-Karp onto shapeless recipes by Desour 20:06 Desour a single crafting check can take around 80 ms, just because of one single recipe 20:08 MTDiscord oh wow, this is awesome 20:09 Krock looks like black magic. and you're sure there's no helper for that in the std library? 20:09 MTDiscord Krock: standard libraries aren't all that large, this is pretty specific 20:10 MTDiscord correct me if I'm wrong, but I don't think the C++ standard library has any graph facilities? 20:10 Desour correct, afaik 20:11 Desour I agree that it looks like black magic. if you want to know how the algo works, look at wikipedia or some other source 20:13 Desour are there any existing unittests for crafting? 20:15 Krock not that I know of 20:15 MTDiscord Desour: this is still linear time, right? 20:15 MTDiscord (at least linear time in the worst case) 20:17 Desour linear in the number of recipes? yes. 20:17 Desour I didn't change the filtering yet 20:17 Desour (but have plans) 20:17 MTDiscord okay 20:17 MTDiscord good 20:18 MTDiscord I've found finding a proper index structure to be a hard problem because of groups 20:21 Desour my idea is to give each recipe an u32 index, and store all recipes in an array (this idx can later be used for sorting). then for PRIORITY_SHAPED_AND_GROUPS for example, store in a hashtable a sorted list of such recipes for each pair(itemname, shape). then at craft lookup these lists for each input item and intersect them with sorted-merge. (with shape I mean something like a bitmap that says at which slots there's and item and which are empty) 20:22 Desour and such lists are normally quite short, as can be seen in craft guides: if you look at what you can craft with an item, I've never seen anything more than 200 or something 20:23 Desour s/'s and item/'s an item/ 20:48 Desour see here for details: https://gist.github.com/Desour/e351fb7063ef10ed50549f6e1262d8b7