Minetest logo

IRC log for #minetest-dev, 2017-06-10

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

All times shown according to UTC.

Time Nick Message
00:03 halt_ joined #minetest-dev
00:04 halt_ joined #minetest-dev
00:06 halt_ joined #minetest-dev
00:07 halt_ joined #minetest-dev
00:09 halt_ joined #minetest-dev
00:10 halt_ joined #minetest-dev
00:41 Puka joined #minetest-dev
00:44 Tmanyo joined #minetest-dev
01:54 CalebDavis joined #minetest-dev
01:57 Miner_48er joined #minetest-dev
03:44 jsgrant_ joined #minetest-dev
05:25 Raven262 joined #minetest-dev
06:09 QwertyDragon joined #minetest-dev
06:25 Hunterz joined #minetest-dev
06:42 VanessaE http://www.minetest.net/customize/  <-- dreambuilder should be removed from this page (it is not a subgame).  carbone, BFD, and voxelgarden are all dead
06:42 VanessaE (under "popular subgames"_
06:42 VanessaE )
07:15 nerzhul joined #minetest-dev
07:37 jin_xi joined #minetest-dev
07:48 DS-minetest joined #minetest-dev
07:57 Krock joined #minetest-dev
08:12 kilbith joined #minetest-dev
08:44 cx384 joined #minetest-dev
09:01 Krock joined #minetest-dev
09:21 Krock any plans of implementing this mapgen? http://imgur.com/gallery/0nFro :P
09:26 Krock sadly, a mapblock couldn't be regenerated the same way if it was deleted
09:44 red-001 could someone review #5941 or #5852?
09:44 ShadowBot https://github.com/minetest/minetest/issues/5941 -- [CSM] allow the client to access and modify settings in the `client.` namespace. by red-001
09:44 ShadowBot https://github.com/minetest/minetest/issues/5852 -- Improve the path select GUI by red-001
10:14 YuGiOhJCJ joined #minetest-dev
10:47 Krock red-001, there's an unanswered question on 5852
10:53 red-001 I'm not sure, I'm pretty sure someone requested that I update the files but I can't remeber who, I can revent that part of the commit if it's an issue
11:04 Fixer joined #minetest-dev
11:10 nerzhul joined #minetest-dev
11:11 red-001 Krock, revented the change and rebased + squashed the PR
11:11 kilbith joined #minetest-dev
11:12 Krock s/revent/revert/
11:13 nerzhul merging #5946 in #10 mins
11:13 ShadowBot https://github.com/minetest/minetest/issues/5946 -- Order es_DrawType exactly like enum NodeDrawType in nodedef.h by Thomas--S
11:13 ShadowBot https://github.com/minetest/minetest/issues/10 -- Fixed null pointer dereference errors found by cppcheck by joshbeck
11:13 nerzhul oops why i setted a # befor 10 :p
11:13 nerzhul before*
11:14 cx384 joined #minetest-dev
11:16 nerzhul #5948 sounds good now Krock no ?
11:16 ShadowBot https://github.com/minetest/minetest/issues/5948 -- Add a server-sided way to remove color codes from incoming chat messages by red-001
11:40 nerzhul merging #5948 #5604 & #5945 in ~10 mins
11:40 ShadowBot https://github.com/minetest/minetest/issues/5948 -- Add a server-sided way to remove color codes from incoming chat messages by red-001
11:40 ShadowBot https://github.com/minetest/minetest/issues/5604 -- lua_api: Correct craft recipe param `type` to `method` by DS-Minetest
11:40 ShadowBot https://github.com/minetest/minetest/issues/5945 -- C++11 patchset 6: forbid object copy using assigment/copy function delete by nerzhul
11:53 DS-minetest joined #minetest-dev
11:56 nerzhul if somes wants to comment #5953
11:56 ShadowBot https://github.com/minetest/minetest/issues/5953 -- Remove JThread license and add MT license header by nerzhul
12:10 Taoki joined #minetest-dev
13:49 kilbith_ joined #minetest-dev
14:32 srifqi joined #minetest-dev
14:55 Megaf joined #minetest-dev
14:55 calcul0n joined #minetest-dev
15:03 Player_2 joined #minetest-dev
15:23 Megaf joined #minetest-dev
15:39 KaadmY joined #minetest-dev
17:15 nerzhul joined #minetest-dev
17:27 ShadowNinja Krock, sfan5, rubenwardy, nore, celeron55, nerzhul: Meeting in about half an hour.
17:44 paramat joined #minetest-dev
18:02 ShadowNinja Alright, meeting starts now.
18:02 VanessaE soi
18:02 VanessaE so
18:02 VanessaE first order of business
18:02 VanessaE put the PB&J Pup back into minetest_game, without the animations and sound effects.
18:02 ShadowNinja VanessaE: _game meeting.
18:03 ShadowNinja Calinou: Any progress on the website pictures?
18:04 ShadowNinja First PR is to fix the shader label alignment #5820
18:04 ShadowBot https://github.com/minetest/minetest/issues/5820 -- Menu: Do not use textlist for shaders in settings tab by octacian
18:05 ShadowNinja Positioning might need a bit of tweaking, but the formspec scaling might make it impossible to get it lined up perfectly for everyone, so I wouldn't worry about it too much.
18:05 Calinou ShadowNinja: no :(
18:05 Krock but this PR already improves it a bit
18:05 Krock better than nothing, at least
18:05 ShadowNinja Is it possible to display a disabled checkbox?  That would always line up, but I'm net sure if formspecs support it.
18:06 Calinou don't shaders have that already? in the options
18:06 ShadowNinja Calinou: Alright, well... whenever you get the time.
18:06 Calinou ShadowNinja: yeah, I still have an internship going on right now, I should be on summer break on June 27th
18:06 Krock "checkbox[<X>,<Y>;<name>;<label>;<selected>]" - no "enabled" arg available yet
18:06 ShadowNinja Calinou: Nah, they turn into a text list.  No checkboxes at all and they don't line up properly (See pics in PR).
18:07 Calinou ah, I forgot
18:07 * ShadowNinja mumbles something about how formspecs have to go.
18:09 ShadowNinja Krock: Do you want to test and merge it?
18:09 Krock already tested it. (that's where my image comes from)
18:10 ShadowNinja Krock: Alright.  It lines up pretty well?
18:10 Krock yes
18:11 Krock I think getting it done better will require checkboxes that can be disabled
18:11 Krock will merge in a minute
18:11 ShadowNinja Krock: Great, go ahead and merge then.  Next is removeing superfluous null checks #5914.  Supposedly requires a rebase, but GitHub disagrees.
18:11 ShadowBot https://github.com/minetest/minetest/issues/5914 -- Removed redundant pointer null checks by QrchackOfficial
18:11 Krock nerzhul, are you there?
18:12 Krock 5914 needs improvement - and maybe a rebase
18:13 ShadowNinja Looks like 5914 is pretty simple.  I can merge it.
18:14 ShadowNinja Krock: What changes does it need?  Just looks like there's a spacing change request -- and that's pretty pedantic.
18:15 Krock there was something about the shared pointer being removed
18:15 Krock but I'm fine with the PR if it's mergeable
18:16 ShadowNinja Yeah, it was removed in the PR and trunk, but looks like the conflict has been resolved.
18:16 Krock interesting
18:16 ShadowNinja Another to consider is whether to remove minetestmapper.py #5901
18:16 ShadowBot https://github.com/minetest/minetest/issues/5901 -- Remove minetestmapper from this repository by nerzhul
18:17 Krock Either move to the mapper repository or keep as-is. The code shouldn't be thrown away IMO
18:18 ShadowNinja I don't think anyone uses it anymore, and it's still in the git history if anyone wants it.  So I'm fine with just removing it.
18:18 ShadowNinja Or it could get its own repo -- but it will probably never be updated.
18:18 Krock okay, nerzhul and paramat share the same opinion. Well then, remove it.
18:19 paramat since its preserved in history removal is ok with me
18:20 Krock but the PR is incomplete, the colors file should be removed aswell
18:20 ShadowNinja Yes.
18:21 paramat and sectors2sqlite.py?
18:21 kilbith I'm going to close #5457
18:21 ShadowBot https://github.com/minetest/minetest/issues/5457 -- GUI: Allow texture packs to customize the mouse pointer by kilbith
18:22 ShadowNinja paramat: Yeah, that could probably be removed too.  Doubt anyone's used sectors in a long time.
18:22 ShadowNinja Next PR: Minetest apparently has a built-in file browser, but it can't select folders: #5852
18:22 ShadowBot https://github.com/minetest/minetest/issues/5852 -- Improve the path select GUI by red-001
18:24 rubenwardy Irrlicht has an inbuilt file browser :P
18:24 Krock yes, that's what we use. But the directory selecting feature wasn't working right
18:24 Krock or still isn't
18:25 rubenwardy I suggest finding a wrapper library for native file browsers
18:25 ShadowNinja rubenwardy: Yeah, but it resets the locale every time you open it.  That's why the "local mod" option was removed from the mod store tab.
18:25 rubenwardy when doing the Node Box Editor, I used tinyfiledialog - it's probably unmaintained by now
18:27 ShadowNinja Qt works really well (native file pickers on most/all platforms), but it's a lot to bring in for just a file browser.
18:27 ShadowNinja But fixing the GUI is a but too big of a topic for now I think.
18:27 rubenwardy woah
18:27 kaeza joined #minetest-dev
18:27 rubenwardy tinyfiledialogs was last modified 13 hours ago
18:27 rubenwardy https://sourceforge.net/projects/tinyfiledialogs/files/
18:30 ShadowNinja So does the path select PR look good?
18:31 paramat no opinion
18:34 ShadowNinja Looks OK to me.
18:35 Puka joined #minetest-dev
18:35 sfan5 does that pr introduce a new dependency?
18:35 ShadowNinja sfan5: Nope, just some tweaks to existing code.
18:36 sfan5 that's fine then
18:37 sfan5 haven't looked at the code though yet
18:37 ShadowNinja We also have #5819 to look at.
18:37 ShadowBot https://github.com/minetest/minetest/issues/5819 -- Fix default item callbacks to work with nil users by raymoo
18:40 nerzhul Krock, i'm eating with my family, late today, then i just read
18:40 ShadowNinja The item_drop change looks pretty unrelated.
18:40 nerzhul tell me what to remove on 5901 and i will update it
18:40 Krock oh sorry for interrupting, nerzhul
18:41 nerzhul no probme, as you see there is latency :p
18:41 nerzhul problem*
18:42 nerzhul ShadowNinja, for 5953 we removed all code for mutexes, and the threading code only has the class and some function declaration but definition has changed
18:42 nerzhul look at 0.4.8
18:42 Krock nerzhul, you'll see the comments on github ;) colors.txt and sectors2sqlite.py can be removed
18:43 nerzhul okay i will update this evening or tomorrow then
18:43 Krock great. thanks :)
18:45 ShadowNinja nerzhul: Yes, but I believe it's still legally under the JThread license if it was modified from JThread, even if there's lottle or nothing left of the original code.
18:48 ShadowNinja Can we get a decision on timed move #1489  sfan5, rubenwardy, paramat?
18:48 ShadowBot https://github.com/minetest/minetest/issues/1489 -- Timed move by sapier
18:49 paramat i'm neutral
18:50 Krock modders will always find an usage for this feature
18:51 ShadowNinja I'd rather see it implemented in a more general way -- such as a movement queue or just CSM.
18:51 sfan5 not everything needs to be csm
18:51 sfan5 but a movement queue might be nice
18:51 Krock indeed
18:52 Krock "setvelocity({x=num, y=num, z=num}, time)" if "time" is set, add to the queue
18:54 ShadowNinja I was thinking something more like q = obj.movement_queue(); q.add({type="move_to", pos={...}, time=1}); q.add({type="acceleration", acceleration={...}})
18:57 VanessaE ShadowNinja: we need to settle the issue about 0.4.x backports and point releases, now that the 0.5.x cycle has started
18:57 Krock is that on the list? (where's the paste btw?)
18:57 ShadowNinja VanessaE: Do you want to do a bugfix release?
18:58 ShadowNinja Krock: No, but Iconsidered discussing it anyways.  See -staff.
18:58 VanessaE ShadowNinja: not necessarily.  more a matter of "will we, won't we".  i.e. let's make a decision about it now.
18:58 paramat there will be a point release to sort out csm
18:59 paramat other fixes could be in that
18:59 VanessaE paramat: CSM is one issue that needs addressed, but I'm sure there will be other things as well
18:59 Krock I believe nerzhul came up with this idea first. So we can release 0.4.17 and 0.5.0 parallel. the first just for bugfixes to play on older servers and the other is the actual thing
18:59 VanessaE Krock: yes, but no one here has actually *agreed* on doing it
18:59 DS-minetest joined #minetest-dev
18:59 Krock that's why we're discussing it here now
18:59 paramat Krock that's a separate proposal
19:00 paramat red suggests a point release to restrict csm in around 2 months, maybe less
19:01 ShadowNinja At some point we should go through the CSM API and make sure that everything in there is what we want, and make some sort of proper seperation between mods (I believe it does the same thing as the server ATM -- just running all mods in one lua state.  I think it also uses the server security module -- which is more lenient than it probably should be since it was focused on backwards compat for the server.
19:02 VanessaE paramat: I'd say 4-6 weeks; will we do a protocol bump to knock "bad csm" clients out?
19:02 paramat maybe
19:02 VanessaE ok.
19:02 paramat actually yes
19:03 paramat because then the restrictions can be enforced
19:03 VanessaE make "strict checking" the default then
19:03 VanessaE (it's off by default, now, I believe)
19:03 paramat of course :]
19:04 VanessaE just making sure :)
19:04 paramat you know how ifeel about the issue :]
19:04 VanessaE perhaps ;)
19:05 red-001 VanessaE, then you lose a lot of players
19:05 VanessaE oh well.
19:05 red-001 look at the server list stats to see what I mean
19:05 VanessaE there'll be an official release when that happens.  users can upgrade.
19:06 VanessaE fwiw, right now daconcepts.com 30001, my dreambuilder survival server, is at the top of the list.
19:06 red-001 https://kitsunemimi.pw/tmp/serverlist_stats_2017-03-17.txt
19:06 VanessaE that one doesn't work well on mobiles, meaning PC users mostly, and those folks can upgrade easily.
19:06 red-001 only 05.89% where latest
19:06 red-001 were*
19:06 VanessaE right
19:06 paramat red, it's optional to prevent old clients, but necessary to completely enforce csm restrictions, as you explained
19:06 paramat only some servers will
19:07 VanessaE red-001: but then again, look at the stats closer:  the majority are on Multicraft.
19:07 VanessaE a mobile client
19:07 red-001 true
19:07 VanessaE so f--- them. :P
19:07 paramat hehe
19:07 red-001 for minetest clients it was even worse
19:07 red-001 most were 0.4.13
19:08 VanessaE red-001: think of it this way:  how would 0.5.x affect them, when servers start moving to that?
19:08 Krock it's not like they can't update minetest. strict protocol checking will just speed up the process
19:08 VanessaE answer of course is that everyone will break for a while
19:08 red-001 I do have to ask is 0.5.0 going to fully break backwards compatibility?
19:08 paramat yes this will give clients a reason to upgrade, so more will
19:09 VanessaE it will break backward compat enough to make it a problem for luddites, yes
19:09 VanessaE it already breaks backward compat with older OS's
19:09 paramat sorry red, that 'yes' was not for your question
19:09 Krock [limited to protocol backwards compability]
19:09 VanessaE (because of C++11 requirements)
19:13 paramat nore ShadowNinja what are your opinions on game#1543 ?
19:13 ShadowBot https://github.com/minetest/minetest_game/issues/1543 -- Add 'creative' privilege for survival servers by tenplus1
19:14 ShadowNinja paramat: Will conflict with unified_inventory.
19:15 paramat ok thanks, please add a comment
19:15 ShadowNinja (or at least, they both have the same priv, I don't know if Minetest allows redefinitions).
19:16 paramat ah so a new priv name is needed?
19:16 nerzhul red-001, we did this for breaking compat, and i hope it will be strictly done, we have some protocols problems, in packets for example we have some bad integer truncates or sign mismatch
19:17 nerzhul C++11 is just a compiler break, no functional change atm
19:17 ShadowNinja paramat: Looks like there are no conflict checks, so UI will just "win" and get to set the priv def.
19:17 paramat yes it's a good time to break compat, there's no better time
19:17 Krock builtin/game/privileges.lua: core.registered_privileges[name] = def
19:17 Krock redefinitions are no problem
19:18 paramat but a new priv name is better surely?
19:19 paramat i'll merge #5908 later
19:19 ShadowBot https://github.com/minetest/minetest/issues/5908 -- (Re)spawn players within 'mapgen_limit' by paramat
19:19 Krock I don't see any problem there. This change shouldn't conflict with u_i
19:20 ShadowNinja paramat: No, it's fine.
19:21 paramat ok you +1?
19:21 paramat concept approval
19:21 ShadowNinja paramat: Yep.
19:21 paramat ok
19:22 AntumDeluge joined #minetest-dev
19:24 paramat Krock +1?
19:28 Krock concept approval from my side too
19:28 paramat ok
19:29 paramat it can certainly be merged then
19:32 paramat game#1758 ?
19:32 ShadowBot https://github.com/minetest/minetest_game/issues/1758 -- Stairs: Use one recipe matching inventory appearence by paramat
19:33 paramat game#1764 is potentially good now
19:33 ShadowBot https://github.com/minetest/minetest_game/issues/1764 -- Adding backface culling to stairs by MrRar
19:35 DS-minetest the extra variable in game1764 is unneeded
19:36 paramat which variable?
19:36 DS-minetest stair_images
19:37 DS-minetest `images` can be overridden directly in the loop
19:37 paramat ok
19:38 paramat i see
19:39 DS-minetest and what is if backface_culling (or something else) is already given? imo it shouldn't override then (eg. for glass sairs)
19:41 paramat yeah
19:45 paramat thanks
19:47 ShadowNinja nerzhul: Can you take a look at this?: http://sprunge.us/BiQh?diff  It removes threads.h in favor of the C++11 types and methods, and it fixes the threadProc signature Since C++11 supports arbitrary thread function signatures.
19:50 sfan5 you should make a pr for that as it changes quite many lines
19:51 Agent07 joined #minetest-dev
19:52 ShadowNinja sfan5: My PRs usually just sit in the queue for years...  It's a good number of lines, but it's mostly just search and replace, so you can go over it quickly.
19:53 ShadowNinja It doesn't actually change any functionality too, so as long as it compiles you can be pretty sure it's correct.
20:28 YuGiOhJCJ joined #minetest-dev
20:36 nerzhul ShadowNinja, if you can do a PR i can review
20:36 ShadowNinja nerzhul: Can you just review the patch?
20:38 nerzhul too huge i prefer GH
20:38 nerzhul i notice you don't have the macro for COPY removal, you have only copy constructor removal not copy assigment removal
20:40 ShadowNinja nerzhul: Ehm, I think the C++11 one does both -- but I didn't change that I just moved it.  That change was from your previous C++11 threading PR.
20:41 nerzhul no it doesn't do both, in all docs i read about this usage both are required
20:42 nerzhul just use DISABLE_CLASS_COPY(class) after constructor it will generate it for you
20:42 nerzhul and provide a GH pr i will review & test, maybe someone else and we will merge it soon
20:42 nerzhul ShadowNinja, oh okay :) then i missed it, but it should be re-used
20:43 nerzhul i just want to finish those PR on C++11 base for our code
20:44 nerzhul we didn't discussed about #pragma once usage, too late i think, i and zeno was in favour (i should go too), don't hesitate to debate and add a rule if accepted :p
20:44 nerzhul i hope we can flush the pr queue and be under 160 PR :(
20:51 ShadowNinja I've also created a new .clang-format, which more closely matches the current style.
20:54 ShadowNinja !tell nerzhul https://github.com/minetest/minetest/pull/5957
20:54 ShadowBot ShadowNinja: O.K.
20:57 Megaf joined #minetest-dev
21:03 diegom joined #minetest-dev
21:06 Gundul joined #minetest-dev
21:16 CalebDavis joined #minetest-dev
21:31 red-001 so #5852 can be merged?
21:31 ShadowBot https://github.com/minetest/minetest/issues/5852 -- Improve the path select GUI by red-001
21:38 ShadowNinja Hmmm, it seems that if you set clang-format to not add a newline between class Foo and the starting brace, it will also trim all the empty newlines before the class.
21:38 ShadowNinja Same for struct, enum, etc.
21:40 ShadowNinja Maybe it's a bug or there's some config option for that (although it doesn't look like it).  If not, we might have to switch to AStyle or something to get this to work.
21:43 ShadowNinja clang-format 4.0 does tha same thing, so probably not a bug (or at least not one likely to be fixed).
21:46 ShadowNinja Doesn't look like clang-format supports "class Foo {" without removing all empty lines befpre the befinition.
21:46 ShadowNinja AStyle is probably a good choice to replace it.
21:53 jin_xi joined #minetest-dev
22:13 DI3HARD139 joined #minetest-dev
22:19 paramat joined #minetest-dev
22:33 halt_ joined #minetest-dev
22:35 red-001 halt_ seems to have a bad connection
23:09 johnnyjoy joined #minetest-dev
23:34 red-001 does CSM need access to the jit table?
23:34 red-001 the fucntions in it don't seem like something a mod needs
23:34 red-001 and removing it will get rid of one more attack vector
23:36 paramat halt/grandolf should probably be banned here, they are in minetest channel
23:37 red-001 well they seems to quickly part whenever they join
23:37 paramat or at least grandolf is, but apparently they share a computer
23:37 red-001 CSM needs a serious security review
23:37 paramat they can read logs, there's no need for them to participate here :]
23:40 paramat rubenwardy ShadowNinja please could you ban halt_ ?
23:41 ShadowNinja paramat: Er, from where?
23:41 red-001 ShadowNinja, since you orginaly implemented mod security could you tell me if #5941 correctly implements the isClient function?
23:41 ShadowBot https://github.com/minetest/minetest/issues/5941 -- [CSM] allow the client to access and modify settings in the `client.` namespace. by red-001
23:42 red-001 I don't want mods to be able to modify the vaule of isClient since I use it to check if we need to enable more strict sandboxing for settings
23:42 ShadowNinja red-001: The server security code isn't supposed to be used on the client in the first place...
23:44 ShadowNinja red-001: You don't initialize the RIDX for the server.
23:44 ShadowNinja red-001: Also, adding a second seperate "secure" settings namespace seems like a bad idea.
23:46 red-001 huh so it's not certain that it will be nil/false if I don't initialize it?
23:46 ShadowNinja red-001: Your new security check is flawed -- it will work in all settings files instead of just for the main settings.
23:47 ShadowNinja Well -- yes, it would be nil.
23:47 red-001 well I don't think CSM is likely going to be allowed to load other setting files
23:48 red-001 since it's not meant to have filesystem access
23:48 ShadowNinja red-001: You erase the first 7 chars of every setting name in to_table?
23:48 red-001 yes
23:48 red-001 but only if they are "client."
23:49 red-001 it's so that mods don't have to deal with the abstraction of the "client." namespace
23:49 ShadowNinja Er, s/to_table/get_names/
23:49 red-001 both?
23:50 sfan5 what happens if i name a setting client.secure.something
23:51 ShadowNinja Oh, I see what you're doing.  I don't like it at all.  The security check macro is actually mutating local variables.
23:51 red-001 you have a bad naming scheme then
23:52 red-001 so should I move it into a function then?
23:52 sfan5 it's not about bad naming scheme, its about whether this could confuse mods or result in a security issue
23:52 ShadowNinja red-001: You're also not doing the conversion back in get.
23:52 ShadowNinja red-001: I think setting groups are just what you should be using here.
23:53 red-001 setting groups?
23:53 ShadowNinja red-001: https://github.com/minetest/minetest/blob/master/src/settings.h#L137
23:54 ShadowNinja Then do something along the lines of core.settings = LuaSettings(g_settings->getGroup("client"))

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