Time  Nick        Message
00:31 gregorycu   Who can help me
00:32 acerspyro   Talk about being undescriptive!
00:32 gregorycu   I had a pull request, which got merged, and then unmerged due to a deadlock on linux
00:32 gregorycu   A fix has already been done, and there is an outstanding PR
00:33 gregorycu   What should happen in this scenario?
00:43 hmmmm       update the pull request with the fixed version
00:45 gregorycu   But the first pull request was already closed
00:45 gregorycu   I think I know what I need to do
00:45 hmmmm       make a new one
00:46 gregorycu   Rebase my fork, rebase my branch, merge the fixes (that already exist on the branch) and then resubmit a PR
00:46 gregorycu   I'm getting good at this git shit
00:47 hmmmm       git shit
00:47 hmmmm       hehehheh
00:47 gregorycu   Sounds like an insult
00:47 gregorycu   "You're a fucking git shit"
00:48 gregorycu   Oh... I think I need to do a pull of the branch too
02:31 VanessaE    can someone answer this guy?  https://forum.minetest.net/viewtopic.php?f=4&t=1583&start=250#p168725
02:31 VanessaE    I got nuthin' at all.
02:32 VanessaE    (normal usage for HDX at max resolution with normalmaps, should be around 6GB)
02:40 hmmmm       we need a stacktrace for that too
02:40 hmmmm       it could be a legitimate case of running out of memory, or it could be caused by heap corruption, or maybe the allocation function was passed a bogus size value on accident
02:41 hmmmm       whatever the reason is, that error originates from the lua interpreter
02:42 hmmmm       also note that he says it never exceeds 10GB of ram, but that's probably the working set size and not reserved
02:43 hmmmm       i bet if he checked the memory reserved size of the process he'd see his PC close to exhaustion, or maybe it went up against a per-process memory usage quota
02:46 VanessaE    a legit case of running out of memory?
02:46 VanessaE    um, the guy has 24 GB... HDX takes 6 tops.
02:47 VanessaE    anyways I only pass it to here because I can't do anything to help him :-/
02:50 VanessaE    his "for any meaningful amount of time" suggests he's being hit by another memory leak somewhere
02:52 Zeno`       dreambuilder test game massif output. With mesh cache: https://www.dropbox.com/s/oqsirhqi58468dq/massif.out.11177?dl=0
02:52 Zeno`       Without mesh cache: https://www.dropbox.com/s/2kubvrxo3hd3bd0/massif.out.11283?dl=0
02:52 VanessaE    ah, the perfect man to step in at this time :D
02:52 Zeno`       several days old but not much has changed since then
02:53 Zeno`       Should I hide?
02:53 VanessaE    nah
02:54 VanessaE    see the link I pasted
02:54 Zeno`       I have read it ;)
02:55 VanessaE    Minetest REALLY should print its version number/commit ID to the console at startup
02:56 puhfa       ^
03:04 VanessaE    so, he's using the standard 0.4.11 release build.
03:04 VanessaE    I pointed him to sfan5's and recommended he figure out how to use a debugger
03:16 hmmmm       Zeno`, I can't install massifg :(  could you make the graphs for me?
03:17 Zeno`       does massifvisualizer work?
03:20 Zeno`       with mesh cache: http://i.imgur.com/aNYKHZF.jpg
03:20 Zeno`       without: http://i.imgur.com/fLhhZNJ.jpg
03:21 hmmmm       wait, why are we even looking at this
03:21 Zeno`       dunno
03:21 hmmmm       what tool is that you're using btw
03:22 Zeno`       massif-visualizer
03:22 hmmmm       honestly wth
03:22 hmmmm       "qt qmake not found!"
03:23 hmmmm       when it's clearly there
03:25 Zeno`       I think I had to edit that from memory... it's been a while
03:26 Zeno`       the build system kind of sucks
03:30 hmmmm       a monkey must have written FindQt4.cmake.  it's atrocious
04:57 gregorycu   #2193
04:57 ShadowBot   https://github.com/minetest/minetest/issues/2193 -- Settings fixes by gregorycu
04:59 gregorycu   #2173 and #2191 are small changes, if anyone feels inclined to merge
04:59 ShadowBot   https://github.com/minetest/minetest/issues/2173 -- Speed up Profiler::avg by a factor of 10 by gregorycu
04:59 ShadowBot   https://github.com/minetest/minetest/issues/2191 -- Fix chat scrolling bug (trying to negate unsigned) by gregorycu
05:13 Zeno`       2913 looks good. Testing in a minute
05:13 Zeno`       2193*
05:22 Zeno`       ok will (re)merge 2193
05:27 Zeno`       there is a bug in chat scrolling?
05:29 gregorycu   I had a compiler warning
05:29 gregorycu   Which is legit
05:29 gregorycu   Though I didn't test, to be honest
05:29 Zeno`       weird. The standard says it should be implicitly cast
05:31 Zeno`       I'll merge it because it's a trivial change but I don't really understand why MSVC is warning about that
05:32 Zeno`       Can you tell me the warning?
05:32 gregorycu   Sure, one moment
05:34 gregorycu   I just tested, it does not get promoted to signed
05:35 Zeno`       bizarre
05:36 Zeno`       will merge #2191 in 10 minutes
05:36 ShadowBot   https://github.com/minetest/minetest/issues/2191 -- Fix chat scrolling bug (trying to negate unsigned) by gregorycu
05:36 Zeno`       I'll remove the extra parenthesis though if that's ok
05:36 Zeno`       i.e.   m_console_buffer.scroll(-(s32)m_console_buffer.getRows());
05:36 gregorycu   ...
05:36 gregorycu   What the fuck
05:37 gregorycu   Oh
05:37 gregorycu   I've come across this before
05:37 Zeno`       Don't tell me MSVC is doing something else stupid
05:38 gregorycu   No, I'm testing on gcc
05:38 Zeno`       Just out of interest, does the warning go away if the line is: m_console_buffer.scroll(0 - m_console_buffer.getRows());
05:38 gregorycu   I'm tipping it would
05:38 gregorycu   http://www.tutorialspoint.com/compile_cpp_online.php
05:39 gregorycu   I'm going to give you some code, I want you to run it
05:39 Zeno`       just put it in codepad?
05:40 Zeno`       But I'll run it anyway if you want, heh
05:40 gregorycu   Oh yeah, that can run code too
05:40 gregorycu   http://codepad.org/8hbesiDo
05:41 gregorycu   What is strange there?
05:42 Zeno`       nothing
05:43 Zeno`       when it's passed to the function is cast to int
05:43 Zeno`       it's*
05:43 gregorycu   What's cast to an int?
05:43 Zeno`       -i
05:44 Zeno`       but in the first cout expression it's not
05:44 Zeno`       i.e. line 13
05:44 gregorycu   What value is cast to an int?
05:44 Zeno`       4294967276
05:44 gregorycu   Yeah, and then it overflows
05:45 Zeno`       hmm, I'm possibly using the word cast incorrectly. It's "cast" as if by assignment IIRC
05:46 Zeno`       http://codepad.org/8ygqJfMb
05:47 Zeno`       anyway... those snippets are working as I would have expected
05:48 Zeno`       what's the overflow, btw?
05:48 gregorycu   Well, we're trying to store 4294967276 in something that can't hold it
05:48 Zeno`       what can't hold it?
05:49 gregorycu   The signed number, which is the parameter to the function
05:49 Zeno`       (unsigned)4294967276 is reinterpreted (cast to) an int
05:50 Zeno`       the underlying bit pattern of 4294967276 and -20 is exactly the same
05:52 gregorycu   These conversion rules have always confused me
05:52 gregorycu   So, we know a cast happens then?
05:52 gregorycu   I mean, with any other type, it tries a conversion
05:53 gregorycu   It's an implicit cast, if it's a cast
05:53 Zeno`       yeah, it's implicit cast
05:53 Zeno`       the "as if by assignment" is for return values from functions
05:53 gregorycu   Which happens for related types
05:54 Zeno`       http://codepad.org/ABT5gNmm
05:54 Zeno`       these snippets are why I'm wondering WTF msvc is complaining about
05:55 gregorycu   It's complaining about the negation of an unsigned variable
05:55 Zeno`       because foo((int)(-i)); makes no difference
05:55 Zeno`       I should see if gcc warns that
05:55 Zeno`       sec
05:55 gregorycu   Is that what I did?
05:55 gregorycu   Oops
05:55 gregorycu   I meant to cast the i first
05:56 Zeno`       err
05:56 Zeno`       you did I think
05:56 Zeno`       typo ;)
05:57 Zeno`       compiling http://codepad.org/ABT5gNmm with g++ (and every single warning possible, even pedantic, I get no warning)
05:57 Zeno`       but I'll push your fix
05:58 Zeno`       maybe I'll change the title to Fix MSVC warning in ChatBackend
05:59 Zeno`       foo(-(int)i); is equivalent to foo(-i) (in the snippets) because of the implicit cast but MSVC is obviously being overly pedantic
06:00 gregorycu   It sees the -i in isolation
06:00 Zeno`       yeah
06:00 gregorycu   To be fair, it's basically saying "That number is going to do strange things, it's probably not what you expect"
06:01 gregorycu   Negating an unsigned number falls under bitfiddling
06:01 Zeno`       Can you make sure:   m_console_buffer.scroll(-(s32)m_console_buffer.getRows());   suppresses the warning as well?
06:02 Zeno`       But it's not negating an unsigned number because there is an explicit cast of the argument. If it was not the argument to a function the warning would be valid, of course.
06:02 Zeno`       implicit*
06:02 Zeno`       I don't mind adding stuff to shut compilers up though :D
06:02 gregorycu   Yeah, that fixes the warning too
06:03 gregorycu   The cast happens after the negation
06:03 gregorycu   It's just not smart enough to know
06:03 gregorycu   And I'm not smart enough to know, I don't understand the promotion rules, or cast rules or whatever
06:04 Zeno`       It's easier to understand them from the C Standard (which is the normative standard for C++) and then cross-reference the minor differences that apply to C++
06:04 Zeno`       C++ Standard is too big :3
06:05 * Zeno`     sees what clang says about this snippet
06:05 Zeno`       nope, no warnings either. *shrug*. Just one of those things
06:10 Zeno`       ok, #2191 closed
06:10 ShadowBot   https://github.com/minetest/minetest/issues/2191 -- Suppress MSVC warning in chat.cpp by gregorycu
06:11 Zeno`       with altered name ;) :-o
06:14 Zeno`       I'm guessing VC has many more warnings as well?
06:25 gregorycu   shitloads
06:29 Zeno`       there's shitloads if you compile with gcc and -Wextra as well
06:30 Zeno`       so we need a standard for one shitload
06:30 Zeno`       so we can compare
06:30 VanessaE    how many warnings does the current Linux kernel produce? :)
06:31 Zeno`       I haven't compiled the Linux kernel in about 6 years (maybe longer)
06:31 VanessaE    me neither.
06:31 Zeno`       those days are in the distant past heh
06:32 Zeno`       were you going to use that as the reference for "one shitload"?
06:32 Zeno`       in the new minetest metrics standard?
06:33 VanessaE    yep
06:33 VanessaE    provided of course it produces a reasonable amount anyway
06:33 Zeno`       hmm, I don't feel like compiling a kernel
06:33 VanessaE    well, how about an older minetest, say 0.4.8?
06:33 Zeno`       lol
06:34 VanessaE    I seem to recall things were pretty lax regarding warnings then
06:34 Zeno`       sorry, I laughed because in my head I wondered if we'd need a 16-, 32- or 64-bit unsigned to store this number
06:35 Zeno`       might even have to install a multi precision library like GMP
06:35 Zeno`       so we can store really big numbers
06:38 gregorycu   I have many many warnings disabled
06:38 gregorycu   Sell, just 6
06:38 gregorycu   4101;4005;4244;4996;4305;4800
06:44 Zeno`       4005 is needed?
06:44 Zeno`       those warnings might be useful to look at
06:52 gregorycu   I'm only a single dude
08:32 sfan5       sofar: you don't need to open a new pull request, you can (force) push to your old branch
08:46 gregorycu   Just be careful with force push
08:57 sofar       yeah, I think that preserves some of the pull request history as well
08:58 sofar       I just don't like github's way of doing git, been used to posting patches on MailingLists too long
09:00 * sofar     goes to bed
11:29 gregorycu   Do we have any crash-on-startup bugs in 0.4.11?
11:29 kilbith     afaik no
11:33 nrzkt       no
11:53 Krock       gregorycu, yes. set video_driver = null
12:02 gregorycu   Is deleting the config safe?
12:04 Fadi        It shouldn't break anything as far as I know
12:04 Krock       wtf is going on in the forum style
12:05 Krock       http://i.imgur.com/d6VNy2S.png
12:51 gregorycu   Who wants to be my guineapig?
13:13 * Fadi      is away: Away exploring the Gamma Quadrant.
13:13 gregorycu   That's my favourite Quadrant
13:14 Fadi        ;)
13:21 gregorycu   Fuck this CD
13:23 Fadi        What's totally not wrong :p?
13:23 Zeno`       getNodeRefUnsafeCheckFlags()
13:23 Zeno`       what if hmm
13:23 Zeno`       nvm
13:24 Zeno`       gregorycu, you need to ask hmmmm if flags are actually used
13:24 gregorycu   Yes, they are
13:24 gregorycu   The signal if a block is loaded
13:25 gregorycu   If I don't check the flags... funny stuff happens
13:25 Zeno`       I guess they do now that I've fixed that bug, but... yeah
13:25 Zeno`       #2156 will have to be checked by hmmmm I think
13:25 ShadowBot   https://github.com/minetest/minetest/issues/2156 -- Optimise MapBlockMesh functions by gregorycu
13:25 gregorycu   The one in the lighting code didn't used to check falgs
13:25 gregorycu   flags
13:25 Zeno`       so hassle him when he's online
13:26 gregorycu   During emerge, the border nodes were 100% lit
13:26 gregorycu   Because I was reading unintialised memory
13:26 Zeno`       yeas
13:26 Zeno`       yes*
13:27 Zeno`       so, ok.. they are used now because of my commit a week ago
13:27 Zeno`       at least the static (non-const) issue is fixed
13:27 Zeno`       it looks good to me
13:28 Zeno`       line 584 of voxel.h should probably be const qualified also ( as well as line 624 of voxel.cpp)
13:30 Zeno`       but those are minor points. I'm ok with it
13:31 gregorycu   Oh yeah
13:31 gregorycu   Oops
13:48 rubenwardy  Awesome, feature freeze.
13:50 gregorycu   Sarcasm?
13:53 rubenwardy  No. Has the Mapgen bug been fixed?
14:05 rubenwardy  Who has access to forum.minetest.net?
14:17 gregorycu   Curses
14:17 gregorycu   http://tinypic.com/view.php?pic=8xkjf6&s=8#.VMT6xS7AZ7Y
14:38 rubenwardy  Lol
14:39 gregorycu   I'm playing with CD, trying to fix a 4 year old bug
14:39 gregorycu   And not break EVERYTHING ELSE in the meantime
14:39 rubenwardy  Sometimes you have to break things to fix things
14:40 gregorycu   Well, I fixed the bug
14:41 gregorycu   And some people reported that you bounce and then sink through the air
14:41 gregorycu   Fucked if I know what that means
14:41 gregorycu   Other than... you know... not good
15:32 Zeno`       No need to be so aggressive, gcu
15:32 Zeno`       people report things that are fairly subjective (in most instances)
15:33 Zeno`       You asked for feedback so you need to accept that some of that feedback is going to be subjective
15:34 gregorycu   No need to get so defensive, I wasn't having a go at you
15:34 gregorycu   Geez
15:34 gregorycu   I wasn't having a go at anyone, i was having a go at the code
15:34 Zeno`       I didn't think you were having a go at me
15:35 gregorycu   Sometimes I think you go looking to get offended
15:36 Zeno`       It takes a *lot* to offend me ;)
15:37 Zeno`       But...
15:37 Zeno`       <gregorycu> Well, I fixed the bug
15:37 Zeno`       <gregorycu> And some people reported that you bounce and then sink through the air
15:37 Zeno`       <gregorycu> Fucked if I know what that means
15:38 Zeno`       ^^ that is not constructive
15:38 Zeno`       *shrug*. Just my opinion
15:39 gregorycu   I'll be more verbose: "Fucked if I know what bouncing and then sinking means in terms of how the effect could possibly be produced"
15:39 Zeno`       Maybe it's because I am 90 years old :P
15:39 gregorycu   Don't tell me how or how not be constructive
15:39 crazyR      has anyone considered creating a minetest.register_on_prenewplayer() method. in which we could stop a player account being created and do other checks if needed before the account is created
15:39 Zeno`       I didn't tell you anything other than my opinion
15:40 gregorycu   When I want your guidance, I'll ask
15:41 gregorycu   Don't project other devs onto me, I'm not here to be a dick, I'm here to be honest and get the job done. I'm not here to ensure I don't offend anybodies sensibilities or be seen as amazingly constructive
15:42 nrzkt       hi
15:44 Zeno`       I didn't project anything
15:45 Zeno`       I stated an opinion
15:45 Zeno`       An interpretation, if that's a better word
15:50 crazyR      anyone?
15:50 gregorycu   Out of interest, what would you check in the callback?
15:51 rubenwardy  Wat. Are you actually 90 years old?
15:52 rubenwardy  crazyR, could be good.
15:53 crazyR      cold check if the name exists in other formats. could help with stopping banned users from even creating an account. probably alot more too
15:54 crazyR      the xban mod would benefit from it.
15:58 crazyR      so the idea behind it is that its identical to the minetest.register_on_prejoinplayer() method, except if it returns then the account is not created and the message int he return is posted to screen same way as prejoin does
15:59 gregorycu   Sounds good
15:59 gregorycu   You should create an issue
16:02 gregorycu   I got latest, and I have trouble starting a game now
16:02 gregorycu   m_emerge = new EmergeManager(this);
16:02 gregorycu   m_emerge is NULL
16:02 gregorycu   wtf
16:03 rubenwardy  under what conditions does the EmergeManager constructor do delete this?
16:03 rubenwardy  That's what's happened. Or you're out of RAM.
16:03 rubenwardy  Or some C++ thing]
16:04 gregorycu   delete this will call the destructor
16:04 gregorycu   It can't possibly set my pointer to null
16:22 gregorycu   I'm off to bed, I can't figure this shit out
16:58 rubenwardy  #1869 would be a very useful addition. What's the current status?
16:58 ShadowBot   https://github.com/minetest/minetest/issues/1869 -- Add scriptapi interface to HTTPFetchRequest by Jeija
16:59 rubenwardy  I guess it relies on the mod security.
17:00 rubenwardy  #1606
17:00 ShadowBot   https://github.com/minetest/minetest/issues/1606 -- Add mod security by ShadowNinja
17:33 crazyR      ^ +1 that should definatly be merged :D
17:46 kaeza       crazyR, check `if minetest.auth_table[name] then ...` in `on_prejoinplayer`. that does what you want
17:47 kaeza       `minetest.auth_table[name]` is `nil` if the player is new (i.e. never logged in before)
17:50 crazyR      kaeza i know that. thats what i so. but... by the time it gets `to on_prejoinplayer` the account has already been created along with the files its creates for the user and the files that the mods create for the user. my suggest for a `on_prenewplayer` would allow us to stop the accounts being created before they have.  meaning less mes to clean up
17:50 kaeza       eh? no
17:50 crazyR      no what?
17:51 kaeza       one of the points of prejoinplayer is to avoid creating the account/initializing the connection/player/whatever
17:51 kaeza       so if prejoinplayer returns something, it is as if the player didn't connect at all to the server
17:52 crazyR      ahhh i undertsand.. i didnt realise thats how it works
18:08 sapier      ShadowNinja: time between those buttonbars is different on purpose. For my first try the left button bar was shown for too long. I assume now it's to short?
18:10 ShadowNinja sapier: Yes, far too short.  They should be the same and should auto-close if you click elsewhere.
18:11 sapier      hmm I don't think they should be same, for those long buttonbars you need more time to select the correct one. for the left one this time is ugly.
18:11 sapier      hide once clicking elsewhere ... that's gonna be difficult
18:12 sapier      e.g. how to behave on finger resting somewhere else on screen?
18:13 sapier      but I'm gonna give it a try, could work
18:14 ShadowNinja sapier: Close it.  Look at how Adnroid handles popup windows.
18:14 sapier      btw the hiding time can be changed quite easy it's a parameter passed on initialization of bar
18:15 sapier      but usual android interface doesn't use two hands
18:16 sapier      and the normal phone usage doesn't depend on any finger beeing there permanent but is more like short "precise" located clicks
18:16 sapier      still let's give it a try
19:03 kilbith     newly generated V6 : https://lut.im/Nm9EHtND/i55wpjAT
19:37 kilbith     ShadowNinja, default.gui_bg_img makes an obvious difference, you should add it.
19:41 crazyR      whats the deal with this guys, https://github.com/minetest/minetest/issues/2150
19:43 nore        sfan5, thoughts on game#414?
19:43 ShadowBot   https://github.com/minetest/minetest_game/issues/414 -- Sam III by kilbith
19:43 ShadowNinja kilbith: I daon't see that difference.
19:43 kilbith     ShadowNinja: MT-Modder screenshots.
19:45 PilzAdam    nore, does the new model have the same hat part?
19:45 nore        hm, I did not check that
19:46 Krock       kilbith, is it possible to make the arms not look like cut-off parts?
19:46 Krock       same for the legs
19:46 kilbith     *missed heart attack*
19:46 PilzAdam    Krock, wouldn't that add more unwanted detail?
19:47 Krock       PilzAdam, it would make the gameplay nicer
19:47 PilzAdam    yes, because visuals = gameplay
19:47 PilzAdam    ?
19:47 ShadowNinja kilbith: I can't reproduce that.
19:48 ShadowNinja And that shouldn't be an image.
19:48 kilbith     no, compensate the slight "cut-off part" would be heavier
19:48 kilbith     and anyways, not really wanted imo
19:48 Krock       slight..lol
19:49 kilbith     your pickyness is not slight for sure
20:59 kilbith     sfan5: https://github.com/minetest/minetest_game/pull/414#issuecomment-71391276
22:06 kilbith     PilzAdam: how i can check for the hat ? dunno what it is...
22:07 PilzAdam    kilbith, see the hair parts in the top-right of the current texture?
22:07 kilbith     yes
22:07 PilzAdam    these are drawn slightly above the cube hat, so it looks more 3D
22:07 PilzAdam    color them green or so and check if the new model uses them
22:08 kilbith     ok
22:08 PilzAdam    s/hat/head/
22:11 PilzAdam    kilbith, game#184 has nice comparison screenshots
22:11 ShadowBot   https://github.com/minetest/minetest_game/issues/184 -- Sam II: by Jordach
22:12 PilzAdam    kilbith, there also seems to be an cape layer (game#201)
22:13 PilzAdam    make sure that the new model supports that, too
22:13 kilbith     hmmm, i colored the right part in green, and i don't see any green both in Sam II and III
22:14 kilbith     idk where exactly i'm supposed see that on the head
22:14 kilbith     VanessaE, you know better this model, you can help me ?
22:18 kilbith     oops, saved the .png in the wrong directory
22:19 kilbith     PilzAdam, Sam III support the cap : https://lut.im/Ablk2c3Z/pTU1SeGA
22:24 VanessaE    The "cape" layer needs to be drawn from a second material imho
22:24 VanessaE    the old model had one but it pulls from a part of the texture file that, roughly 75% of the time in MC skins, has either garbage or a solid color not indicative of a cape.
22:26 kilbith     ^ sfan5, consider that
22:27 VanessaE    as for the hat layer, that's just a matter of re-using the same parts of the head as contain hair, which I think the standard Sam II texture already does.