Time  Nick       Message
05:43 nerzhul    i'm pushing one morning modernize patchset
05:44 sofar      playing with a 0.4.16dev client on a 0.4.15 server is terribly broken atm
05:45 sofar      eye offset is broken, can't walk up steps
05:46 sofar      this should not happen in this way - are we really breaking ABI that badly?
05:46 nerzhul    yes
05:46 nerzhul    it's normal
05:46 sofar      it's bad
05:46 nerzhul    no we are on 0.5 way
05:46 sofar      I can't even run 0.4.16 anymore
05:46 nerzhul    it's not 0.4.16
05:46 sofar      2017-08-17 22:45:02: ERROR[Main]: /usr/share/minetest/builtin/init.lua:40: attempt to call field 'setting_get' (a nil value)
05:46 nerzhul    it's 0.5.0-dev this should be clarified
05:46 sofar      no, vanilla 0.4.16
05:46 sofar      I can't run it anymore
05:46 nerzhul    for packaged 0.4.16 it works
05:47 sofar      I have to delete my -dev builds or git build an old release
05:47 nerzhul    very strange
05:47 sofar      well the lua files are incompat
05:47 nerzhul    for the eye offset we solved (with breakage no choice) the player collision box problem we have since ages
05:48 nerzhul    now it's properly settable and modifiable server side, and client side (client side was hardcoded)
05:48 nerzhul    if i remember we changed the setting api
05:48 nerzhul    use RUN_IN_PLACE
05:48 nerzhul    no the shared space sofar
05:48 sofar      meh
05:48 nerzhul    keep share for stable builds :p
05:48 sofar      I'm gonna stop tracking master, sorry
05:49 nerzhul    master is not stable for end user usage, it's in a huge development process and we are trying to fix all things which needs a break
05:49 nerzhul    also master should always being build with -DRUN_IN_PLACE
05:49 sofar      could have forced a protocol mismatch
05:49 sofar      but you didn't
05:49 nerzhul    protocol was bumped
05:49 nerzhul    but not the min version
05:50 sofar      exactly
05:50 nerzhul    i will open an issue to suggest first to change the version model to a proper model, use next-version-dev instead of previous-version-dev
05:51 nerzhul    it's bad for multi branch model and bad for end user to understand it's not the previous version
05:59 nerzhul    i like this: "// This is a temporary backwards compatibility fix" date: 03/02/2011
05:59 nerzhul    :p
06:03 nerzhul    i opened #6266 for a version scheme update vote
06:03 ShadowBot  https://github.com/minetest/minetest/issues/6266 -- Change version model
06:21 nerzhul    pushing the last code modernization this morning
06:21 nerzhul    i think i'm at 50% work on it
08:37 nerzhul    merging #6244 in ~10 minjs
08:37 ShadowBot  https://github.com/minetest/minetest/issues/6244 -- Change README to Markdown format, add LICENSE.txt by rubenwardy
09:51 nerzhul    hello Krock thanks for the review :)
09:51 nerzhul    merging #6265 in ~10 mins
09:51 ShadowBot  https://github.com/minetest/minetest/issues/6265 -- server.cpp: unroll setting when sending mapblocks by nerzhul
09:53 nerzhul    and #6269 too
09:53 ShadowBot  https://github.com/minetest/minetest/issues/6269 -- Add missing ctime include by adrido
11:01 paramat    devs please add your vote #6266
11:01 ShadowBot  https://github.com/minetest/minetest/issues/6266 -- Change version model
11:04 paramat    Krock game#1876 updated string recipe (2 vertical cotton returns 2 string), are you +1?
11:04 ShadowBot  https://github.com/minetest/minetest_game/issues/1876 -- Farming: Make cotton look like cotton, add crafted string item by paramat
11:35 nerzhul    paramat, thanks for promoting this important change :)
11:38 Krock      yes, paramat
11:43 nerzhul    paramat, see http://en.cppreference.com/w/cpp/language/range-for norm explanation
11:44 paramat    ok thanks and thanks
11:58 paramat    to clarify, devs are ok with enabling zoom using a new 'player object property' as in alongside collisionbox and stepheight?
12:02 nerzhul    it sounds more reasonable than mixing it with hud as it's player not hud
12:03 paramat    yeah i agree now
12:05 paramat    anyway what i'm considering to do now is to add a new HUD flag for minimap radar mode, as discussed, since it is a special ability very different to basic minimap. any objections?
12:05 paramat    Krock rubenwardy sfan5
12:06 sfan5      sounds useful
12:06 paramat    so that, in MTG basic minimap is enabled by crafting a simple paper map (as in my PR) then radar will need a much more special item, either magickal or technological
12:07 Krock      just get the priv away. I don't care whether it gets added to the HUD or player object properties
12:07 Krock      *magical
12:07 Krock      ImageMagick is something different, paramat :P
12:09 nerzhul    paramat, a hud flag to disable radar only ? i'm okay
12:11 paramat    ok thanks
12:15 paramat    my plan is to leave the minimap HUD flag as the master enabler, then the new flag becomes a suboption to enable the radar mode
12:15 nerzhul    paramat, yes it's good
12:41 jas_       if i have to hit ZOOMKEY twice to zoom once, i am going to be ... depressed
12:44 jas_       (in quake, it's cg_zoomtoogle, 0 for off, 1 for enabled.  right now, zooming feels amazing, i hope its behavior, besides requiring a map item, does not change.  yes, i'm thinking about lag.  eager to try a pr.)
12:56 jas_       well sorry for that!  it felt identical.
13:00 paramat    no change in behaviour with the zoom key, we're working only with the permission to zoom
15:53 paramat    #6273 tested and ready for review
15:53 ShadowBot  https://github.com/minetest/minetest/issues/6273 -- Minimap: Add new HUD flag for minimap radar mode by paramat
16:06 nerzhul    merging #6267 in ~10 mins
16:06 ShadowBot  https://github.com/minetest/minetest/issues/6267 -- Modernize various files (src/m*) by nerzhul
16:09 paramat    ok nice
16:17 paramat    #6257 +1 can be merged
16:17 ShadowBot  https://github.com/minetest/minetest/issues/6257 -- Make world creation menu automatically generate a random world name by lisacvuk
16:17 paramat    tested
16:53 nerzhul    i see we have a very old mapsector backend
16:54 nerzhul    on each server start we create a old sector it seems to remove it later lol
16:56 Krock      with some re-arranging we could load the mapblocks directly and deprecate the mapsectors or am I wrong there?
16:57 nerzhul    i don't think
16:57 nerzhul    i added and errorstream on m_sectors on servermap and i have an increase on maploading :)
16:58 nerzhul    but there is old code yes
16:59 nerzhul    i just need to understand why on my map loading i save 6200 mapblocks whereas i only dig 10 nodes
17:01 Krock      depending on the active block range, there may be ABMs active
17:01 nerzhul    i think yes, but it's very huge, it seems we save unconditionally
17:02 nerzhul    no grace timer ? i need to find it
17:02 Krock      nevertheless, backtracing what causes this ridiculous amount of saved data would be helpful
17:03 nerzhul    seem we have dead code, the sector meta on disk :p
17:04 nerzhul    differs_from_disk is always false on mapsector, and meta are into db
17:04 nerzhul    celeron55, when sqlite was added ?
17:04 nerzhul    it seems it's in 2011 or 2012 no ?
17:05 Krock      each set_node() call from Lua marks the mapblocks as changed
17:05 nerzhul    / DEPRECATED? Sectors have no metadata anymore :> 05/02/11
17:05 nerzhul    i think this part can be dropped :p
17:06 nerzhul    03/09/11
17:06 nerzhul    very very old map
17:06 nerzhul    :p
17:06 nerzhul    sqlite3 introduce at this date
17:07 nerzhul    i will propose a PR to remove that part
17:07 celeron55  you're mixing up a lot of things
17:07 celeron55  don't be so hasty
17:07 nerzhul    yes i know
17:07 nerzhul    i have multiple branches
17:07 nerzhul    i will just add a PR for the metadata part
17:07 celeron55  MapSector doesn't have anything to do with the map format called the "sectors format"
17:08 nerzhul    which is disabled since ages
17:08 nerzhul    (in the mapsectors on disk)
17:08 celeron55  it contained some data a long time ago but doesn't anymore
17:08 celeron55  i'm not sure if the old format is supposed to be supported still or not
17:09 celeron55  people have proposed removing it for a long time, maybe it has happened or maybe it hasn't
17:09 nerzhul    2011 ? there were no network, no server owner should be concerned
17:09 celeron55  ???
17:09 nerzhul    sqlite was added 03 sept 2011
17:09 nerzhul    with the migration code from files to sqlite
17:09 celeron55  sqlite was added by a server owner
17:10 celeron55  are you silly
17:10 nerzhul    oh ? i thought network was mor erecent than that :)
17:11 celeron55  anyway, for 0.5, both MapSector and the format can surely be removed
17:11 nerzhul    MapSector c++ objects seems to be used by core, we should be careful
17:11 nerzhul    i don't know why they are there, but they are there: p
17:12 celeron55  hmm actually
17:13 nerzhul    interesting meta, they add just version = 1 :p
17:13 celeron55  the block lookup optimization in MapSector actually probably still makes block fetching faster
17:13 nerzhul    celeron55, i don't see a noticeable usage in callgrind on this part
17:13 celeron55  the map caches the sector pointer that was last fetched, and each sector caches the block that was last fetched
17:14 celeron55  back then it made Map::getNode and stuff faster, i don't see why it wouldn't now
17:14 nerzhul    mapblocks are in mapsectors right ?
17:14 celeron55  yes
17:14 nerzhul    okay
17:14 nerzhul    the only useless part is the disk meta write/read which is empty
17:15 celeron55  essentially the map is stored as a std::map<v2s16, MapSector*> and MapSector has std::map<s16, MapBlock*>
17:16 nerzhul    yeah :)
17:17 nerzhul    seems removing the useless meta serialization means ServerMapSector and ClientMapSector are useless because they are only MapSectors :p
17:22 celeron55  seems correct
17:26 nerzhul    okay tiny PR, easy to test for anyone who wants
17:26 nerzhul    #6274
17:26 ShadowBot  https://github.com/minetest/minetest/issues/6274 -- ServerMap saving: cleanups by nerzhul
17:27 paramat    a mapsector is a complete vertical column of mapblocks
17:27 nerzhul    yeah
17:32 Krock      paramat, changing the 0.5*BS part in the itemCAO function doesn't have any effect
17:34 paramat    hmm
17:36 nerzhul    merging #6257 on ~5mins
17:36 ShadowBot  https://github.com/minetest/minetest/issues/6257 -- Make world creation menu automatically generate a random world name by lisacvuk
17:36 paramat    time to merge some mtg stuff, i'll merge game#1823 game#1844 game#1873 game#1876 over the next few mins, will take time due to squashing
17:36 ShadowBot  https://github.com/minetest/minetest_game/issues/1823 -- TNT explosion custom sound by tenplus1
17:36 ShadowBot  https://github.com/minetest/minetest_game/issues/1844 -- TNT: add explode_center flag by tenplus1
17:36 ShadowBot  https://github.com/minetest/minetest_game/issues/1873 -- default.register_fence: Allow setting nodedefs to 'false' by HybridDog
17:36 ShadowBot  https://github.com/minetest/minetest_game/issues/1876 -- Farming: Make cotton look like cotton, add crafted string item by paramat
17:37 Krock      it's GenericCAO
17:38 Krock      same thing there
17:38 nerzhul    if i remember itemcao is never used
17:38 nerzhul    all is genericCAO
17:38 nerzhul    (shown in debug stuff when you drop an item and hover)
17:38 nerzhul    i should be sure it's never used but maybe we should remove this thing too
17:38 Krock      removing the 0.5 BS offset fixes the problem
17:39 nerzhul    what is the problem ? the dark textures ?
17:39 Krock      yes, wrong light positions
17:39 Krock      I guess it was introduced when the player offset was moved to its feet
17:42 Krock      so why do we keep itemcao when it's unused?
17:42 nerzhul    i'm not sure
17:43 nerzhul    and it's complicated to find how it's instancied due to the factory model :(
17:44 nerzhul    okay i find where it's called
17:44 nerzhul    ClientEnvironment::addActiveObject
17:44 nerzhul    it depend on what server send
17:45 nerzhul    TOCLIENT_ACTIVE_OBJECT_REMOVE_ADD packet
17:47 nerzhul    gob_cmd_update_infant
17:47 nerzhul    ACTIVEOBJECT_TYPE_ITEM
17:47 nerzhul    a SAO with ACTIVEOBJECT_TYPE_ITEM
17:47 Krock      defined in activeobjcet.h
17:48 nerzhul    no SAO has this type
17:48 nerzhul    and SAO refuse to create a such object
17:48 nerzhul    serverobject.cpp L45
17:48 nerzhul    / These are 0.3 entity types, return without error.
17:48 nerzhul    if (ACTIVEOBJECT_TYPE_ITEM <= type && type <= ACTIVEOBJECT_TYPE_MOBV2) {
17:48 nerzhul    return NULL;
17:48 nerzhul    }
17:49 nerzhul    then ItemCAO will never trigger a SAO creation => no packet sent => no ItemCAO created
17:49 nerzhul    class can be safely removed
17:49 nerzhul    s/ItemSAO/ServerEnvironment/
17:50 nerzhul    do you want to do the PR Krock ?
17:51 nerzhul    (if you do it, if you can remove u16 id unused param in ServerActiveObject::create can be nice :D
17:52 Krock      as it belongs to the protocol too, I can bundle this cleanup with the other planned particlespawner fix
17:52 nerzhul    it's not protocol related :p
17:52 nerzhul    the only removal is remove ItemCAO class and references
17:52 Krock      s/protocol/network
17:52 nerzhul    i don't think anybody will connect on a MT 0.3 server ? :p
17:52 nerzhul    (and protocol doesn't permit it)
17:53 Krock      it's deprecated for a long time, so removing it as soon as possible
17:53 nerzhul    yeah, it makes clear that item CAO are not ItemCAO class for PR owners :p
17:57 paramat    merged 3 of those, one conflicted with another
17:58 paramat    genericCAO ok. the issue isn't player offset, it's because the light sample point is inside the node above
18:00 paramat    0.4f * BS above should be ok
18:02 nerzhul    hardcoding various entity size with a such offset is not good i think, you want to fix what ? the player ?
18:02 paramat    #6268
18:02 ShadowBot  https://github.com/minetest/minetest/issues/6268 -- Objects very dark
19:06 troller    btw mapsector was completely removed in freeminer
19:07 troller    and block saving optimized
19:07 nerzhul    what do you mean by block saving optimized ? the current problem with block saving (except it can be saved 10 times per second) it the comprssion
19:08 troller    just dont save not changed blocks
19:09 troller    + optional skip saving generated and not changed blocks
19:09 troller    = 10x less base size
19:09 nerzhul    it's already the case in minetest, the block is saved only if modified flag is set :)
19:10 nerzhul    base size is not very important, the compression do the trick :p but it can be an idea
19:10 troller    i remember some bugs fixed in this place
19:11 troller    in worlds with 5 years history compression does nothing
21:04 Fixer      nerzhul: you sure? o_O
21:04 Fixer      nerzhul: then why just simply walking I get world sizes up to 200 mb and more... i have not modified anything...
21:26 Wuzzy      hi. i want to ask some stuff about warnings in the log I get from stock MTG 0.4.16
21:26 Wuzzy      "WARNING[Main]: NodeDefManager: Ignoring CONTENT_IGNORE redefinition"
21:26 Wuzzy      "WARNING[Main]: BanManager: creating /home/wuzzy/.minetest/worlds/xxd/ipban.txt"
21:27 Krock      hello. that's your fault. more precise: you're trying to use set_node with an unknown node or ignore
21:27 Wuzzy      and "WARNING[Main]: /!\ You are using old player file backend. This backend is deprecated and will be removed in next release /!\"
21:27 Wuzzy      !!!stock Minetest Game!!!
21:27 Wuzzy      no additional mods
21:27 Krock      and the backend will be converted automatically, I think
21:27 Wuzzy      this warning does not make sense. i thought default is SQLite3, and I use default
21:28 Fixer      Wuzzy: old player file backend is just playerfiles in *.txt format, in future they will reside in database
21:28 Wuzzy      why is this warning appearing?
21:28 Krock      a warning is supposed to warn the user.. seems logical
21:28 Wuzzy      ok
21:28 Fixer      Wuzzy: you can already use database for playerfiles
21:28 Fixer      just not by default
21:28 Wuzzy      but the CONTENT_IGNORE warning seems more serious
21:29 Fixer      calling it deprecated is a little bit confusing since it is not yet deprecated, but will be on next version
21:29 Wuzzy      oka
21:29 Wuzzy      why is this CONTENT_IGNORE happening?
21:29 sfan5      the banmanager thing should not be a warning at all
21:30 Krock      oh, the ignore part could also be builtin, trying to re-register ignore somehow
21:30 Wuzzy      damn, IGNORE warning also appears in minimal
21:30 Wuzzy      yep
21:31 Wuzzy      its definitely builtin. it also happens in the empty subgame
21:38 nerzhul    !tell Krock the content_ignore redefinition was always here
21:38 ShadowBot  nerzhul: O.K.
21:38 nerzhul    for the player backend you should migrate to sqlite, see wiki about database to known how to migrate to sqlite or postgres
21:39 nerzhul    sfan5, i agree it should be in infostram
21:39 nerzhul    stream*
21:41 Wuzzy      nerzhul: weird. i get this warning even for new worlds
21:41 Wuzzy      did I manage to screw up my settings? I think i use pretty much default
21:44 nerzhul    for the depreciation yes, it's always the default, we don't migrate on 0.4.16
21:52 celeron55  it shouldn't really be a warning because migration will happen automatically in the future and the user doesn't need to do anything at any point if they don't feel like it
21:52 celeron55  but... whatever i guess
21:55 Wuzzy      Is this warning good or bad?:
21:55 celeron55  i do agree with wuzzy that there shouldn't be warnings by default when everything is working correctly
21:55 Wuzzy      WARNING[Server]: StaticObjectList::remove(): id=1545 not found
21:55 Wuzzy      i think some warnings are really cryptic. like, they dont include the information to hunt problems down
21:56 celeron55  it could be bad or not... the static object system isn't always behaving nicely
21:57 Wuzzy      could it be a screw-up in Lua scripting?
21:57 celeron55  no, lua isn't responsible for managing those
21:57 celeron55  but lua can do things that trigger bugs or undefined behavior, of course
21:58 Wuzzy      ok
22:24 paramat    now fixing the banmanager message
22:26 nerzhul    nice
22:27 bigfoot547 Hi nerzhul
22:30 paramat    because that's been bothering me too
22:30 paramat    #6277
22:30 ShadowBot  https://github.com/minetest/minetest/issues/6277 -- Banmanager: Move 'creating ..' message to infostream by paramat
22:31 paramat    there's a PR for the ignore thing
22:31 nerzhul    can you fix the spaces too please ?
22:32 paramat    yeah
22:32 nerzhul    when done, merge, don't wait the CI it's trivial
22:32 paramat    yes
22:33 paramat    erm what spaces?
22:34 nerzhul    between << operator in the log
22:34 nerzhul    infostream << "blabh" << "blah"
22:34 paramat    ok
22:34 paramat    yes good idea
22:44 paramat    i think i would prefer to just do the same for the ignore message, a warning appearing is alarming and not a good impression
22:46 paramat    but then maybe the warning is important to see, not sure
22:46 celeron55  what exactly is the content_ignore warning about?
22:46 nerzhul    the warning is important
22:47 nerzhul    not for this node but for other nodes when it happen
22:47 celeron55  it seems reasonable and can happen due to many reasons
22:47 nerzhul    it's a node redef
22:47 nerzhul    content ignore is defined by core and builtin
22:48 paramat    ok
22:48 celeron55  besides, we have no obligation to make good impressions 8)
22:49 celeron55  maybe i should make a PR that adds a "Your computer is about to explode" warning at MT startup
22:50 celeron55  "Sending personal information to Putin..."
22:50 celeron55  now i know what i'll do in my next project
22:51 paramat    hehe
22:51 paramat    i might make that player backend message less alarming
22:53 celeron55  the player backend message should say something like "You are using the file-based player backend. You can migrate it to sqlite manually, or wait for it to be automatically migrated in a future version"
22:53 paramat    yeah
22:54 celeron55  the "/!\ deprecated" stuff is quite useless
22:54 celeron55  how to tell if a message is stupid: it's stupid if it contains an exclamation mark
22:54 paramat    will do
22:55 bigfoot547 hah!
22:55 bigfoot547 That makes a lotta things stupid!
22:55 bigfoot547 hehe!
22:55 celeron55  ^ proof
22:56 bigfoot547 yeah!