| Time |
Nick |
Message |
| 00:00 |
hmmmm |
i have no idea how to even start debugging this especially since it's only happening semi-consistently for you |
| 00:00 |
VanessaE |
it seems to take about 5-6 hours of uptime before it starts to happen |
| 00:00 |
VanessaE |
and during that 5-6 hours, it performs flawlessly. |
| 00:05 |
VanessaE |
it is conceivable that your mapgen work will solve that as a side effect :-) |
| 00:06 |
hmmmm |
probably, but that's not solving the real problem |
| 00:06 |
VanessaE |
yeah I know. |
| 00:06 |
hmmmm |
are you able to compile minetest for your server? |
| 00:06 |
VanessaE |
Oldcoder seems to think compiler flag tweaks plus leveldb may be a solution (but again wouldn't solve the root provlem) |
| 00:06 |
VanessaE |
oh sure, I always compile from git. |
| 00:06 |
hmmmm |
...compiler flag tweaks? |
| 00:06 |
VanessaE |
yeah, he didn't specify what tweaks he meant, I'll find out more tomorrow I expect. |
| 00:11 |
VanessaE |
OldCoder is going to drop in and explain. |
| 00:11 |
VanessaE |
(I gotta see this ;) ) |
| 00:14 |
OldCoder |
Hi |
| 00:14 |
OldCoder |
The "tweaks" are trivial and people may therefore disagree. I will comment regardless. I need a little while to prepare. |
| 00:15 |
OldCoder |
I will drop other things for a bit and focus on this |
| 00:20 |
hmmmm |
vanessa, add enable_mapgen_debug_info = true in your config if you haven't done so already... |
| 00:20 |
hmmmm |
if you already did add that, search for "EmergeThread: p=" |
| 00:20 |
VanessaE |
done. |
| 00:20 |
hmmmm |
your logs are gonna be really verbose |
| 00:21 |
VanessaE |
they already are anyway |
| 00:21 |
VanessaE |
and I've enabled it via /set also |
| 00:21 |
OldCoder |
working |
| 00:21 |
hmmmm |
ahh good |
| 00:21 |
VanessaE |
the setting was not turned on (but you knew that) |
| 00:22 |
hmmmm |
well |
| 00:22 |
hmmmm |
i was reading your description from yesterday |
| 00:22 |
hmmmm |
and i noticed you said ABMs were occasionally running |
| 00:22 |
VanessaE |
yes, intermittently during the lockdown |
| 00:23 |
hmmmm |
ABMs are run in two threads, the ServerThread on step and the EmergeThread after loading the block |
| 00:23 |
VanessaE |
lock-up, hang, whatever word you prefer :-) |
| 00:23 |
hmmmm |
now what are the symptoms of the "server hang" exactly? |
| 00:24 |
hmmmm |
I mean my intuition is telling me that it's ServerThread that's hanging |
| 00:24 |
OldCoder |
Which setting? *scroll* |
| 00:24 |
VanessaE |
ok: 100% CPU, packets stop flowing to the client (the lowest rate I've seen is 1 packet every 2-3 seconds). Commands sometimes execute, usually not. Players can move around from their own perspectives, but I don't see them moving. Chat sometimes works, usually not. |
| 00:24 |
hmmmm |
oh, tell OldCoder not to worry, he's not included in the conversation since he ignored me for whatever stupid reason |
| 00:24 |
VanessaE |
hmmmm: but here's the kicker: |
| 00:25 |
OldCoder |
How do I unignore somebody? |
| 00:25 |
hmmmm |
100% cpu |
| 00:25 |
hmmmm |
that's an important detail |
| 00:25 |
VanessaE |
sometimes, the only nasty thing is that the map loader stops executing - but players cna chat, commands still work, etc. |
| 00:26 |
OldCoder |
nvm |
| 00:26 |
VanessaE |
that is, players can walk around, not sure if they can build (I forgot already), but the map just doesn't load. |
| 00:27 |
VanessaE |
if for example, the spawn hasn't been visited in a long while and the server has already unloaded, and I teleport there from somewhere far away, I am dropped into a grey void and nothing happens after that. |
| 00:27 |
VanessaE |
unloaded that area I mean |
| 00:28 |
hmmmm |
it sounds to me like ServerThread is holding the envlock while using lots of cpu while EmergeThread is waiting on it |
| 00:28 |
hmmmm |
but that's pure speculation |
| 00:29 |
hmmmm |
you did disable all mods right? |
| 00:29 |
VanessaE |
right now the server is up and running normally, there are three of us online (two active), packets are streaming to the client at close to 10/sec, and the server is only occupying about 10-15% CPU. |
| 00:29 |
VanessaE |
this is my normal config with a full set of mods. |
| 00:30 |
hmmmm |
hrmmmmmm |
| 00:30 |
OldCoder |
Script completed. Testing before release. |
| 00:30 |
hmmmm |
alright so here's what i would have you do |
| 00:30 |
hmmmm |
keep letting the server run like normal |
| 00:30 |
hmmmm |
and you have the mapgen debugger enabled |
| 00:30 |
VanessaE |
yes |
| 00:30 |
hmmmm |
so then wait and see if it freezes up around the 5-6 hour mark again |
| 00:30 |
VanessaE |
ok. |
| 00:31 |
hmmmm |
if that happens then give me your latest server log |
| 00:31 |
hmmmm |
and |
| 00:31 |
hmmmm |
i'll tell you what to do from there i guess |
| 00:31 |
VanessaE |
okey dokey. I guess I better go kill the old log file and let the server start a fresh one then |
| 00:31 |
VanessaE |
whoa holy shit on a shingle |
| 00:31 |
VanessaE |
-rw-rw-r-- 1 vanessa vanessa 1369496045 Dec 19 19:32 .minetest/debug-server.txt |
| 00:31 |
hmmmm |
sfgsdfg |
| 00:32 |
VanessaE |
13 GIGA!? |
| 00:32 |
VanessaE |
GIGS* |
| 00:32 |
hmmmm |
lots of spamming |
| 00:32 |
hmmmm |
there was some message in there repeated ad nauseam |
| 00:32 |
OldCoder |
Heh |
| 00:32 |
VanessaE |
yeah, you saw one example earlier in the channel log |
| 00:32 |
hmmmm |
that would be a cause of the 100% cpu i reckon |
| 00:32 |
hmmmm |
well hold on |
| 00:32 |
hmmmm |
that much? |
| 00:32 |
VanessaE |
yup |
| 00:32 |
hmmmm |
all of that blah can't possibly be those messages |
| 00:33 |
hmmmm |
once a peer times out, that's it |
| 00:33 |
hmmmm |
it's not going to keep repeating that |
| 00:33 |
VanessaE |
I'm not sure when that log file was last deleted/created |
| 00:33 |
hmmmm |
hrm |
| 00:33 |
hmmmm |
anyway i think it's time to delete the log |
| 00:33 |
hmmmm |
hah |
| 00:33 |
hmmmm |
watch that solve the problem |
| 00:34 |
VanessaE |
lol |
| 00:34 |
hmmmm |
but anyway i'm curious as to what compiler flags OldCoder thinks can solve the issue |
| 00:34 |
hmmmm |
i just want to know |
| 00:35 |
VanessaE |
ok, server restarted with a fresh log file |
| 00:35 |
VanessaE |
and already 2.3 MB in the...whoa |
| 00:36 |
VanessaE |
map loader is DOA |
| 00:36 |
VanessaE |
19:36:55: INFO[EmergeThread]: EmergeThread: p=(-5,3,27) only_from_disk=1 |
| 00:36 |
VanessaE |
lots of that |
| 00:36 |
VanessaE |
19:36:07: INFO[EmergeThread]: EmergeThread: not in memory, attempting to load fr |
| 00:36 |
VanessaE |
om disk |
| 00:36 |
VanessaE |
19:36:07: INFO[EmergeThread]: EmergeThread: p=(-4,8,27) only_from_disk=1 |
| 00:36 |
VanessaE |
and a significant amount of that |
| 00:36 |
VanessaE |
and now the map just loaded. |
| 00:37 |
OldCoder |
Hello |
| 00:37 |
OldCoder |
Done |
| 00:37 |
VanessaE |
hey. |
| 00:37 |
OldCoder |
Shall I post? |
| 00:37 |
VanessaE |
wait |
| 00:37 |
OldCoder |
And I have not followed this discussion. I don't know what will solve what. I can only tell you what I have observed and share a script. |
| 00:38 |
VanessaE |
looks to be spewing about 20-25 of those events per second. |
| 00:41 |
VanessaE |
thankfully this runs from a SSD so speed of the log and map accesses isn't an issue :-) |
| 00:41 |
OldCoder |
I need to move on. Shall I post? |
| 00:42 |
VanessaE |
post. |
| 00:42 |
OldCoder |
I am scheduled to teach somebody MySQL in a bit. And to upload new worlds. |
| 00:42 |
OldCoder |
O.K. |
| 00:42 |
OldCoder |
|
| 00:42 |
OldCoder |
I've had good results with servers recently. I attribute this to six things: (a) Fixes in core code. (b) Changing compiler flags. The default settings cause corruption and crashes. One of you doubts this but it is true. (c) Use of LevelDB. (d) Optimized .conf settings. (e) Enhancements added to my server framework. (f) Upgrading one of my VPSes. |
| 00:42 |
OldCoder |
|
| 00:42 |
OldCoder |
I hope to post my latest LevelDB notes soon. My server framework will be discussed over time. Regarding compiler flags and .conf settings, here is a shell script for you. It downloads vanilla C55 git code, compiles it, installs MT, and adds a good .conf file. Enjoy. Note: NDEBUG is required or things will break. |
| 00:42 |
OldCoder |
|
| 00:42 |
OldCoder |
http://minetest.org/mtbuildc55.txt |
| 00:42 |
OldCoder |
Done |
| 00:42 |
hmmmm |
oh he's *scheduled* to teach someone mysql |
| 00:43 |
hmmmm |
what a big important guy |
| 00:43 |
OldCoder |
hmmmm, fuck you |
| 00:43 |
hmmmm |
he's so busy |
| 00:43 |
OldCoder |
I was invited here today |
| 00:43 |
hmmmm |
he's taking time out of his valuable life to bestow his grace onto us |
| 00:43 |
hmmmm |
but |
| 00:43 |
OldCoder |
VanessaE, is the above sufficient to address what you asked of me? |
| 00:43 |
hmmmm |
vanessa, that stuff in the logs looks normal |
| 00:44 |
VanessaE |
hmmmm: I figured so. |
| 00:45 |
OldCoder |
hmmmm, I don't know what your problem is but we have enough to deal outside the ivory tower here. I came here in good faith to help. |
| 00:45 |
VanessaE |
OldCoder: I am looking at it now |
| 00:46 |
hmmmm |
i have no idea what your problem is in the first place. you ignored me and started acting like some sort of big shot |
| 00:46 |
OldCoder |
s/deal/deal with/ |
| 00:46 |
hmmmm |
now i don't like you |
| 00:46 |
OldCoder |
hmmmm, I /IGNOREd you because you referred to your "bunghole" moments after meeting me |
| 00:46 |
hmmmm |
lol |
| 00:46 |
OldCoder |
And the MySQL issue is that I keep my promises to people. I doubt you would understand that. |
| 00:47 |
OldCoder |
I am a bit tired. Where are *you* when the LM issue needs to be dealt with? |
| 00:47 |
OldCoder |
With your thumb up your "bunghole" I assume |
| 00:47 |
OldCoder |
You are the one who is more than a bit self-important |
| 00:47 |
OldCoder |
The ability to contribute does not in and of itself make you a contributor |
| 00:48 |
hmmmm |
w/e |
| 00:48 |
OldCoder |
The ability to go more than, let's say, three seconds without urinating on people is good too |
| 00:48 |
OldCoder |
w/e indeed |
| 00:48 |
hmmmm |
anyway it seems like your 'good config file' just lessens the load on the server with ultra-conservative settings |
| 00:48 |
OldCoder |
Try it. Go to five seconds maybe. You can probably do it. |
| 00:48 |
OldCoder |
Of course |
| 00:48 |
OldCoder |
I've said this before |
| 00:49 |
OldCoder |
I believe gcc bugs may contribute to some issues though |
| 00:49 |
hmmmm |
what is NDEBUG for |
| 00:49 |
hmmmm |
LevelDB? |
| 00:49 |
OldCoder |
And BTW regarding "don't like" me I am not especially fond of you either |
| 00:49 |
OldCoder |
NDEBUG is... |
| 00:49 |
OldCoder |
A preprocessor flag used in MT to enable or disable some debug code |
| 00:49 |
OldCoder |
If it is not asserted various things will break |
| 00:50 |
OldCoder |
This suggests that core cleanup will be a good idea |
| 00:50 |
OldCoder |
It is not limited to LevelDB IIRC |
| 00:50 |
hmmmm |
by that you mean sqlite3, right? |
| 00:50 |
hmmmm |
because that's where it's mostly used |
| 00:50 |
OldCoder |
Explain, please |
| 00:50 |
OldCoder |
Probably; it's been a while since I looked |
| 00:50 |
OldCoder |
But I think it broke network code |
| 00:51 |
OldCoder |
Rather, its absence did |
| 00:51 |
OldCoder |
Are we done? Got parting mockery before you head back to ignoreland? And I never said this would be useful. I was asked to post this. |
| 00:52 |
OldCoder |
Regarding important I am not. But I got a life like you do. And it's sort of fucked but I am here to help anyway. |
| 00:52 |
OldCoder |
D: |
| 00:52 |
* OldCoder |
goes to pour artificial orange juice |
| 00:53 |
OldCoder |
Did I mention I have scurvy? Goddamn scurvy? |
| 00:53 |
OldCoder |
You have 60 seconds to mock me |
| 00:53 |
OldCoder |
Or ask questions |
| 00:53 |
OldCoder |
Then I am outta here |
| 00:53 |
VanessaE |
OldCoder: thank you for trying to help. |
| 00:55 |
OldCoder |
VanessaE, you are welcome. As he has not mocked me I take it he has joined Jordan Ibanez in ignoring me first. Fine. VanessaE I owe you some upgrades and responses to other issues. And tell RBA I want to talk to him about MineBest |
| 00:55 |
OldCoder |
So long Mr. hmmmmbug |
| 00:55 |
OldCoder |
*ignore* |
| 00:55 |
OldCoder |
^ sound effect |
| 00:56 |
* OldCoder |
nods |
| 00:56 |
OldCoder |
LevelDB scripts and shit will be posted for all of you to use |
| 00:56 |
OldCoder |
Plus my server framework. Oh, wait... My server framework may not be core dev. So it is fucking useful but I gather you don't need it. |
| 00:56 |
|
OldCoder left #minetest-dev |
| 01:11 |
VanessaE |
default is -O3 for a release build? I can see why he toned that down to -O2. I've had problems with -O3 in the past e.g. under gentoo |
| 01:11 |
VanessaE |
(programs tended to crash) |
| 01:11 |
VanessaE |
but that was years ago |
| 01:59 |
hmmmm |
it's highly unlikely that it's a compiler bug |
| 02:02 |
VanessaE |
I would tend to agree but who knows, right? |
| 02:02 |
hmmmm |
and I was away making/eating dinner |
| 02:02 |
hmmmm |
I don't get what his deal is |
| 02:02 |
hmmmm |
he's like |
| 02:02 |
hmmmm |
too serious or something |
| 02:02 |
hmmmm |
he takes IRC and minetest way too serious. |
| 02:04 |
VanessaE |
well, |
| 02:05 |
VanessaE |
hrm |
| 02:05 |
hmmmm |
but anyway about NDEBUG |
| 02:05 |
VanessaE |
strike that, better I don't say anything. |
| 02:05 |
hmmmm |
NDEBUG is the flag for "NO" debugging in SQLite3 |
| 02:05 |
hmmmm |
it optimizes things |
| 02:05 |
VanessaE |
I assumed that was the case actually |
| 02:05 |
hmmmm |
it's also used for a couple of things in minetest, nothing too important though |
| 02:06 |
hmmmm |
but it's ironic because celeron must have assumed it was the other way around |
| 02:06 |
VanessaE |
(I looked briefly at CMakeLists.txt earlier to try to understand what is beign changed) |
| 02:06 |
hmmmm |
with NDEBUG defined it'll enable catching acceptions and what not |
| 02:06 |
VanessaE |
being* |
| 02:07 |
hmmmm |
so i think it's okay if you turn off enable_mapgen_debug_info |
| 02:08 |
hmmmm |
my guess is that you'll be okay... if it acts up then at least we know what the problem isn't 8) |
| 02:08 |
VanessaE |
ok, turned it off (false) |
| 02:08 |
VanessaE |
haha |
| 02:08 |
VanessaE |
right |
| 02:08 |
hmmmm |
and for good measure delete debug.txt |
| 02:08 |
hmmmm |
because i bet it's up to like 100mb by now |
| 02:08 |
VanessaE |
while the server is running? |
| 02:08 |
hmmmm |
hrm |
| 02:08 |
hmmmm |
can you do that? |
| 02:09 |
hmmmm |
i guess it depends on what sort of lock the file was opened with and what filesystem you're using |
| 02:09 |
VanessaE |
(and actually it's only at 14.7 MB right now) |
| 02:09 |
hmmmm |
nevermind then |
| 02:09 |
VanessaE |
I'm not sure if I can :-) |
| 02:09 |
VanessaE |
it's ext4 |
| 02:10 |
VanessaE |
heh... |
| 02:10 |
VanessaE |
21:08:34: ACTION[ServerThread]: CHAT: <InfinityProject> Damn I just love these shaders!!! |
| 02:10 |
VanessaE |
21:10:17: ACTION[ServerThread]: CHAT: <VanessaEzekowitz> beautiful, aren't they? :-) |
| 02:10 |
VanessaE |
21:10:41: ACTION[ServerThread]: CHAT: <InfinityProject> compared to this; 0.4.3 was poop |
| 02:10 |
hmmmm |
they really make that much of a difference huh |
| 02:10 |
hmmmm |
i can't wait to upgrade then :) |
| 02:10 |
VanessaE |
(I have a separate window with just chat output running so I can avoid using F10 for the mostpart) |
| 02:11 |
VanessaE |
you do realize something right? |
| 02:11 |
hmmmm |
huh |
| 02:11 |
VanessaE |
you upgrade to 0.4.4, and try to rebase your mapgen changes against it, and all hell is going to break loose by way of merge conflicts :-) |
| 02:12 |
hmmmm |
from what changes i'm aware of it shouldn't be too horrible |
| 02:12 |
hmmmm |
but that's why i want to finish this up ASAP |
| 02:13 |
VanessaE |
probably - I know you're careful, but you know how it is coding sometimes. You think, oh this will be easy. 2 weeks later you're still only half-way through your project because for every breakthrough you make in your code, 2 other things just break :-) |
| 02:13 |
hmmmm |
so true |
| 02:13 |
hmmmm |
and my original plan for the mapgen isn't working out so well |
| 02:13 |
hmmmm |
I am still looking at minecraft for inspiration |
| 02:15 |
VanessaE |
yeah, I saw your latest "oh fuck that was a waste of time" collision. |
| 02:15 |
VanessaE |
but, you'll figure it out. |
| 05:20 |
|
kotolegokot joined #minetest-dev |
| 06:18 |
|
rsiska joined #minetest-dev |
| 06:46 |
hmmmm |
alright |
| 06:46 |
hmmmm |
i'm really at the point where i do not want to hard code NoiseParam structures |
| 06:47 |
hmmmm |
i want to load them in ServerMap::ServerMap |
| 06:47 |
hmmmm |
can't get them from the config file, too many parameters |
| 06:48 |
hmmmm |
defining it in Lua would be the "good" way but I just need the structures much earlier |
| 07:04 |
celeron55 |
hmmmm: minetest behaves with NDEBUG just as i want it to |
| 07:05 |
celeron55 |
hmmmm: when NDEBUG is off, it is a debug build, and then you want it to let all uncatched exceptions go so you can see where they come in a debugger |
| 07:05 |
celeron55 |
when NDEBUG is on, it is a release build, and you're not running it in a debugger most likely, and then you want it to handle them by itself |
| 07:05 |
|
jin_xi joined #minetest-dev |
| 07:06 |
hmmmm |
whoops |
| 07:06 |
hmmmm |
you're right, sorry. i totally missed it saying ifNdef |
| 07:08 |
celeron55 |
there could be a bit more control over that but what it does is exactly what one wants in 99% of cases, so it's fine |
| 07:11 |
celeron55 |
hmmmm: actually, looking at what you said, you didn't miss anything 8) |
| 07:11 |
celeron55 |
it's just unintuitive how it should work |
| 07:37 |
celeron55 |
somebody should probably add this commit upstream https://github.com/dannydark/minetest/commit/b6e2679ac7c5f270eb493b5d4e9ea06eaf28f40a |
| 07:37 |
celeron55 |
(i can't now) |
| 07:37 |
celeron55 |
(it's the first one here https://github.com/celeron55/minetest/pull/349) |
| 07:56 |
darkrose |
pushed |
| 08:03 |
hmmmm |
anyway i think i should add a getStructure() thing in Settings |
| 08:04 |
hmmmm |
idk how it's going to work, i guess I just specify a format string and takes varargs |
| 08:04 |
hmmmm |
that's horrible |
| 08:04 |
hmmmm |
i'll just use a lot of individual settings |
| 08:06 |
yunfan |
do you solved the win7 sound problem? |
| 08:06 |
hmmmm |
try installing openal |
| 08:06 |
* hmmmm |
hides |
| 08:09 |
yunfan |
i tried , its not the missing libarray problem, in xp it works, but in win7 it failed |
| 08:21 |
|
kotolegokot joined #minetest-dev |
| 08:34 |
|
SpeedProg joined #minetest-dev |
| 08:35 |
|
SpeedProg1 joined #minetest-dev |
| 09:52 |
|
Calinou joined #minetest-dev |
| 11:19 |
|
SpeedProg joined #minetest-dev |
| 12:36 |
|
Calinou joined #minetest-dev |
| 14:26 |
|
RealBadAngel joined #minetest-dev |
| 14:26 |
RealBadAngel |
hi |
| 14:27 |
RealBadAngel |
celeron55, are you here? |
| 14:28 |
RealBadAngel |
maybe somebody will know |
| 14:28 |
RealBadAngel |
i try to define new matrix4 matrix |
| 14:29 |
RealBadAngel |
with core::matrix4 rotation; |
| 14:29 |
RealBadAngel |
compiler says that matrix4 is not the member of core |
| 14:30 |
RealBadAngel |
but i can find many examples of this in game source |
| 14:30 |
RealBadAngel |
like for example shader.cpp line #234 |
| 14:50 |
thexyz |
RealBadAngel: then show us your code |
| 14:52 |
jin_xi |
make sure right headers are included or try irr::core::matrix |
| 14:52 |
RealBadAngel |
this is the code |
| 14:53 |
RealBadAngel |
just declaration, im not sure what headers are needed for it to work |
| 14:53 |
RealBadAngel |
irr:: doesnt work neither, already tried |
| 14:53 |
RealBadAngel |
‘matrix4’ is not a member of ‘irr::core' |
| 14:54 |
RealBadAngel |
in other cpp's both ways are used |
| 14:54 |
RealBadAngel |
im tryin to use it in mapgen.cpp |
| 14:55 |
jin_xi |
<matrix4.h> |
| 14:55 |
jin_xi |
sry gotta run now |
| 14:56 |
RealBadAngel |
cant see such header file |
| 14:56 |
jin_xi |
probably only client has it, mapgen is server so... |
| 14:56 |
jin_xi |
its included in clientmap.cpp for example |
| 14:56 |
RealBadAngel |
shaders |
| 14:56 |
RealBadAngel |
ahhh |
| 14:58 |
RealBadAngel |
jin_xi, it worked, thx alot |
| 14:58 |
jin_xi |
cool |
| 14:59 |
RealBadAngel |
so, lets build the tree now out of axioms :) |
| 14:59 |
RealBadAngel |
i do have axiom part already done |
| 15:00 |
RealBadAngel |
for sake of testin im goin to replace default make_tree routine |
| 15:01 |
RealBadAngel |
and goin to see how mt will handle it |
| 16:16 |
|
PilzAdam joined #minetest-dev |
| 16:19 |
|
hmmmm joined #minetest-dev |
| 16:31 |
|
rsiska joined #minetest-dev |
| 16:58 |
|
rubenwardy joined #minetest-dev |
| 17:17 |
|
rubenwardy1 joined #minetest-dev |
| 17:22 |
hmmmm |
erm anywya |
| 17:22 |
thexyz |
celeron55: i've started to work on migrating to STL, can you review https://github.com/minetest/minetest/commit/052512ea9fd2c5c07a9d4e881e9dfcab4a3f8a49 ? |
| 17:23 |
hmmmm |
i realized that lua gets executed far before the emergethread gets created |
| 17:23 |
VanessaE |
ok, map data is an sqlite database, but the sqlite command line util can't open it? |
| 17:23 |
thexyz |
it can |
| 17:23 |
hmmmm |
so i'm going to stuff the NoiseParam definitions in there - not in a game, it'll be part of builtin |
| 17:23 |
VanessaE |
(trying to find a way to "repair" it in case it's corrupted) |
| 17:24 |
hmmmm |
vanessa, it can't? :/ |
| 17:24 |
VanessaE |
vanessa rainbird:~/.minetest/worlds/My_World_server$ echo .dump |sqlite map.sqlite |sqlite map.sqlite-repaired |
| 17:24 |
VanessaE |
Unable to open database "map.sqlite": file is encrypted or is not a database |
| 17:24 |
thexyz |
VanessaE: can you send it to me? |
| 17:24 |
hmmmm |
how big is the map file |
| 17:25 |
VanessaE |
thexyz: sure. hmmmm: about 94MB. |
| 17:25 |
hmmmm |
hmmm.. |
| 17:26 |
VanessaE |
uploading it to my web space... |
| 17:28 |
VanessaE |
thexyz, hmmmm: http://digitalaudioconcepts.com/vanessa/hobbies/minetest/My_World_server.zip |
| 17:28 |
VanessaE |
(the whole world folder just in case you want to try to run it) |
| 17:29 |
thexyz |
got it |
| 17:29 |
thexyz |
remove it NOW |
| 17:29 |
VanessaE |
eh?> |
| 17:29 |
thexyz |
as it contains players' passwords |
| 17:29 |
VanessaE |
ok |
| 17:29 |
VanessaE |
didn't think about that |
| 17:30 |
hmmmm |
way to tell everyone |
| 17:30 |
VanessaE |
deleted. |
| 17:31 |
VanessaE |
guess I better upload a version that's just the map, in the future :-) |
| 17:31 |
VanessaE |
or at least, that lacks auth.txt |
| 17:31 |
thexyz |
sqlite> PRAGMA integrity_check; |
| 17:31 |
thexyz |
ok |
| 17:31 |
thexyz |
it's fine |
| 17:33 |
celeron55 |
why does VanessaE have sqlite3 called sqlite on the command line |
| 17:33 |
celeron55 |
on all of my computers it's sqlite3 |
| 17:33 |
hmmmm |
aha |
| 17:33 |
hmmmm |
wow |
| 17:33 |
VanessaE |
celeron55: dunno, that's just what ubuntu did with it on my box. |
| 17:33 |
hmmmm |
that went right over my head |
| 17:33 |
hmmmm |
well that's the problem :) |
| 17:34 |
VanessaE |
well shit |
| 17:34 |
celeron55 |
it's probably not the problem, but that is what i noticed |
| 17:34 |
VanessaE |
this is sqlite2 that I installed |
| 17:34 |
VanessaE |
vanessarainbird:~$ apt-cache show sqlite |
| 17:34 |
VanessaE |
[...] |
| 17:34 |
celeron55 |
it is? lol 8) |
| 17:34 |
VanessaE |
Version: 2.8.17-7fakesync1build1 |
| 17:34 |
hmmmm |
damn |
| 17:34 |
celeron55 |
DAMN I'M GOOD |
| 17:34 |
celeron55 |
-> |
| 17:35 |
VanessaE |
heh |
| 17:35 |
hmmmm |
you'd think that you'd see the problem right away when using 'sqlite' |
| 17:35 |
hmmmm |
ryantitan:/home/ryan/Downloads/My_World_server % sqlite |
| 17:35 |
hmmmm |
SQLite version 2.8.17 |
| 17:35 |
hmmmm |
"version 2.8.17" |
| 17:35 |
VanessaE |
didn't realize it had to be sqlite3 :-) |
| 17:35 |
hmmmm |
i'm more surprised that you didn't pull in sqlite3 as a dependency by now |
| 17:36 |
VanessaE |
eh? I better re-check that. I always pull that stuff in anyway |
| 17:37 |
hmmmm |
but anyway |
| 17:37 |
hmmmm |
is this what you ment by thinking your map file might've been corrupted |
| 17:37 |
VanessaE |
hm, lost 4MB of map data by piping it through sqlite3 |
| 17:38 |
VanessaE |
hmmmm: no, that was just a command line error - I was thinking of ot from the stantpoint that the emerge thread looks like it's trying to pull blocks from disk but is silently failing to do so. |
| 17:38 |
VanessaE |
a quick google had some people having this same sorta problem with sqlite databases in general |
| 17:38 |
hmmmm |
same here |
| 17:38 |
hmmmm |
well regardless |
| 17:39 |
hmmmm |
i think we should probably put some sort of sanity checks for block metadata upon deserializing it |
| 17:39 |
VanessaE |
sounds like a good idea |
| 17:39 |
VanessaE |
ok, let's see if this 4-MB-smaller version works. |
| 17:39 |
VanessaE |
back up the old one..mv the new to the old...start server... |
| 17:40 |
VanessaE |
nope. that had no effect. |
| 17:41 |
VanessaE |
except to make the problem slightly worse |
| 17:41 |
VanessaE |
hah! client crashed. |
| 17:41 |
VanessaE |
let's see...what was the "some exception" this time.. :-) |
| 17:42 |
hmmmm |
mmm yes, nodemetadata.cpp, line 60 |
| 17:42 |
hmmmm |
it's probably a good idea to set an upper bound for number of variables |
| 17:45 |
celeron55 |
VanessaE: it's 4MB smaller because running the database through a dump like that "de-fragments" it, in a way |
| 17:45 |
VanessaE |
gotchya. |
| 17:45 |
VanessaE |
fat lotta good it did though ;-) |
| 17:45 |
celeron55 |
unused space gets created in the file as stuff gets removed from the middle of it and sqlite doesn't have suitably sized stuff to put in the place |
| 17:46 |
celeron55 |
and sqlite doesn't bother squeezing it in while running because it's quite an expensive operation |
| 17:47 |
celeron55 |
http://www.sqlite.org/lang_vacuum.html |
| 17:47 |
celeron55 |
it can do it by using that command |
| 17:48 |
thexyz |
celeron55: have you read my message? |
| 17:56 |
PilzAdam |
hmmmm, your mapgen: http://www.zimg.eu/i/3981767363 (GitHub version) |
| 17:56 |
hmmmm |
why does it have those white spots everywhere |
| 17:56 |
PilzAdam |
not generated chunks |
| 17:57 |
celeron55 |
probably not |
| 17:57 |
PilzAdam |
s/chunk/mapblock |
| 17:57 |
hmmmm |
those are generated chunks |
| 17:57 |
celeron55 |
it's some kind of a bug in an optimization |
| 17:57 |
hmmmm |
they;'re all generated |
| 17:57 |
PilzAdam |
no, they arent |
| 17:57 |
PilzAdam |
ive seen this ingame too |
| 17:57 |
hmmmm |
nonsense, they're generated in 5x5x5 block chunks |
| 17:58 |
PilzAdam |
I was using this mod to generate the map: http://minetest.net/forum/viewtopic.php?id=2950 |
| 17:58 |
PilzAdam |
and made the interval to 1/4 second |
| 17:58 |
hmmmm |
besides, there's no way a non-square piece of the map can be not generated |
| 17:58 |
hmmmm |
map blocks are square |
| 17:59 |
hmmmm |
unless that's on the underside |
| 17:59 |
PilzAdam |
the block under the visible hill is not loaded |
| 17:59 |
hmmmm |
as in |
| 17:59 |
hmmmm |
so you did this with a mod |
| 17:59 |
hmmmm |
and the player wasn't looking downward i bet |
| 17:59 |
hmmmm |
also those parameters suck, try this one instead |
| 18:00 |
hmmmm |
http://pastebin.com/xwmCD3ae |
| 18:00 |
celeron55 |
thexyz: i don't see problems in it |
| 18:00 |
thexyz |
fine, i'll continue then |
| 18:01 |
hmmmm |
well |
| 18:02 |
hmmmm |
I would tune the biome parameters to be better, but basically you need to reduce the frequency of biome group 1, right now it's at .30 |
| 18:02 |
hmmmm |
make that more like .10 |
| 18:02 |
hmmmm |
and there's still way too much gravel |
| 18:02 |
hmmmm |
so change gravel's heat_max to something like 30 |
| 18:02 |
hmmmm |
and change regular beach's heat_min to 30 |
| 18:03 |
hmmmm |
you still need to add in more biomes |
| 18:04 |
hmmmm |
anyway i have a good idea on how to make the biome boundaries more interesting and less blocky. |
| 18:07 |
VanessaE |
as long as they aren't dithered :-) |
| 18:07 |
PilzAdam |
I dont like smooth biome transition |
| 18:07 |
hmmmm |
that's not smooth |
| 18:08 |
PilzAdam |
just saying |
| 18:08 |
hmmmm |
in order to work on the new mapgen i have to finish up patching up the old mapgen |
| 18:08 |
hmmmm |
like |
| 18:08 |
hmmmm |
i only work with one branch of minetest at any given time |
| 18:09 |
hmmmm |
it encourages me to finish what i started before moving onto something else |
| 18:09 |
PilzAdam |
gtg; bye |
| 19:23 |
|
doserj joined #minetest-dev |
| 19:47 |
|
doserj joined #minetest-dev |
| 20:12 |
|
rubenwardy1 joined #minetest-dev |
| 20:16 |
|
rubenwardy1 left #minetest-dev |
| 20:18 |
|
SpeedProg joined #minetest-dev |
| 20:36 |
|
mrmeow joined #minetest-dev |
| 23:48 |
VanessaE |
celeron55: around? |
