| Time |
Nick |
Message |
| 02:19 |
hmmmm |
ahh today's such an off day |
| 02:20 |
hmmmm |
well, crap, it looks like i won't be having pumpkins in the Nether by tomorrow |
| 02:21 |
hmmmm |
i want to explode Farmesh into nothingness; it makes me define these get_have_beach and tree_amount_2d things that make silly assumptions on how maps are generated and what kind of content they can have |
| 02:22 |
hmmmm |
and it's going to be hilariously incorrect with the new mapgen anyway |
| 02:25 |
hmmmm |
i think i'm going to want to add some fields to the map DB format, like generation parameters, mapgen version, etc. so a new executable will be able to continue generating old maps just as fine, and maps that are shared will still be able to maintain coherency |
| 02:26 |
hmmmm |
so what happened from 0.3 -> 0.4 doesn't happen again |
| 03:42 |
|
jyfl987 joined #minetest-dev |
| 04:46 |
|
hmmmm joined #minetest-dev |
| 05:11 |
OldCoder |
|
| 05:12 |
OldCoder |
VanessaE, If you see this, thexyz indicates that the db_leveldb= edit to the .conf file should now be omitted. It is apparently sufficient to edit the world.mt file. Request is to edit your LevelDB post accordingly. |
| 05:12 |
OldCoder |
|
| 05:13 |
OldCoder |
thexyz, VanessaE, celeron55: LevelDB seems to work well with latest git WorldEdit. I transplanted a building into a regenerated world and it loaded instantly and without apparent problems. //metasave and //metaload may be working now. |
| 05:13 |
OldCoder |
|
| 05:14 |
OldCoder |
VanessaE, implications are that RedCrab may wish to convert to LevelDB before attempting a WorldEdit copy of his map into a regenerated world. May speed things up. |
| 05:14 |
OldCoder |
|
| 05:15 |
OldCoder |
Note: These are core dev issues as they relate to decisions about LevelDB which is a fundamental issue. |
| 05:15 |
OldCoder |
|
| 05:59 |
jyfl987 |
OldCoder: good, i am very happy that you guys now considering leveldb |
| 05:59 |
OldCoder |
Hi |
| 05:59 |
OldCoder |
jyfl987, More than considering... I am switching all my worlds |
| 05:59 |
OldCoder |
Some may run into glitches but this is unavoidable |
| 05:59 |
jyfl987 |
OldCoder: i remember i have request this |
| 05:59 |
OldCoder |
Very well. It is proceeding. |
| 05:59 |
jyfl987 |
after i check the sql file |
| 05:59 |
OldCoder |
You may thank thexyz for his hard work |
| 06:00 |
jyfl987 |
oh, thexyz again, i found he always give nice technical solution |
| 06:00 |
OldCoder |
Yes |
| 06:00 |
OldCoder |
I ran into a serious memory problem and he was kind enough to address it |
| 06:00 |
jyfl987 |
OldCoder: so what about the performance improment? |
| 06:01 |
OldCoder |
It is significant |
| 06:01 |
OldCoder |
The advantages are as follows: |
| 06:02 |
OldCoder |
1. Faster runtime. This seems to mean fewer crashes. There are bugs that manifest when things are slow. |
| 06:02 |
jyfl987 |
OldCoder: so there wont be lag when you play on localserver? |
| 06:02 |
OldCoder |
2. WorldEdit seems to be much faster. Coupled with the recent fixes to the //meta commands WorldEdit may now become an everyday tool. |
| 06:02 |
OldCoder |
I have not tried local server |
| 06:02 |
OldCoder |
But there should be improvement |
| 06:02 |
OldCoder |
3. Disk usage is lower |
| 06:03 |
jyfl987 |
and i think the improment on write will perform better |
| 06:03 |
OldCoder |
Savings are 25% to 50% and up |
| 06:03 |
OldCoder |
Yes |
| 06:03 |
jyfl987 |
yep, that's cool , |
| 06:03 |
OldCoder |
Main problem is: |
| 06:03 |
OldCoder |
Slower shutdown but it is not a serious issue |
| 06:03 |
OldCoder |
The only thing I'm concerned about is possible bugs that will manifest over time |
| 06:03 |
OldCoder |
Memory leaks etc. We will see. |
| 06:03 |
OldCoder |
Thank you for your interest jyfl987 |
| 06:04 |
jyfl987 |
OldCoder: i was guessing is it the thread model's problem? |
| 06:04 |
OldCoder |
It is not a problem; it is to be expected |
| 06:05 |
OldCoder |
And I am not seeing as significant a slowdown as I thought |
| 06:05 |
OldCoder |
It seems to vary |
| 06:05 |
OldCoder |
One issue is... |
| 06:05 |
OldCoder |
Building under Windows |
| 06:05 |
OldCoder |
We need to factor out Boost if possible |
| 06:06 |
OldCoder |
cornernote has made an interesting point jyfl987. For some types of tools MySQL will be better as only one process can use a given LevelDB database at a time. |
| 06:07 |
jyfl987 |
well ,yes its another problem |
| 06:08 |
jyfl987 |
i have reported the win32 built version 's sound doesnt work on win7 64bit |
| 06:08 |
OldCoder |
jyfl987, Is this true of the latest Windows client I have built? |
| 06:08 |
OldCoder |
Probably; but should be checked |
| 06:09 |
jyfl987 |
OldCoder: i built that myself on linux using sfan5's script which modified from celeron55's |
| 06:09 |
celeron55 |
is it the matter of wrap_oal.dll or whatever like that not being included? |
| 06:09 |
jyfl987 |
nope, i think it might be the mingw's problem |
| 06:09 |
jyfl987 |
because the same file could sound on my win xp in a vbox |
| 06:09 |
celeron55 |
very unlikely; instead what i said is very likely |
| 06:09 |
OldCoder |
jyfl987, I'd be curious as to whether my build manifests the issue. But it probably does. |
| 06:10 |
celeron55 |
jyfl987: you just happen to have openal explicitly installed on them |
| 06:10 |
jyfl987 |
OldCoder: i will check your version tonight |
| 06:10 |
jyfl987 |
OldCoder: can you give me your address of your built? |
| 06:10 |
OldCoder |
All right. Use the 30MB version. This will be much faster. |
| 06:10 |
OldCoder |
Yes |
| 06:10 |
jyfl987 |
OldCoder: btw, i want to try the leveldb version |
| 06:11 |
OldCoder |
Most of my worlds are LevelDB on the server side now. I have not built a LevelDB client package yet. |
| 06:11 |
OldCoder |
http://minetest.org/clients/minetest-121020.zip |
| 06:11 |
OldCoder |
That is a 30MB Windows build from a couple of weeks ago |
| 06:12 |
jyfl987 |
OldCoder: but isnt that the client and server share the same code on serving? |
| 06:12 |
OldCoder |
Yes |
| 06:12 |
OldCoder |
Technically my servers are clients |
| 06:12 |
OldCoder |
But I have not packaged this for download |
| 06:12 |
jyfl987 |
btw. i found that some sound mod makes the expierences much better |
| 06:13 |
OldCoder |
Once it is loaded yes |
| 06:13 |
OldCoder |
The slow load needs to be looked at |
| 06:13 |
OldCoder |
VanessaE has suggested streaming |
| 06:13 |
jyfl987 |
OldCoder: yes, this is another problem |
| 06:14 |
OldCoder |
Preloaded cache helps but that is simply a work around |
| 06:14 |
jyfl987 |
OldCoder: streaming is bad for bandwidth |
| 06:14 |
OldCoder |
Then another solution is needed |
| 06:14 |
OldCoder |
The packet size may be wrong etc. |
| 06:14 |
jyfl987 |
why is slow? |
| 06:14 |
OldCoder |
Don't know yet |
| 06:14 |
jyfl987 |
i found even on local server , the loading is also slow |
| 06:15 |
jyfl987 |
is it caused by the writting to disk IO? |
| 06:15 |
OldCoder |
I don't think so |
| 06:15 |
OldCoder |
I suspect something simple and fundamental |
| 06:15 |
OldCoder |
Such as packet size |
| 06:15 |
celeron55 |
the slow load time is due to too simple bandwidth throttling by the network stack |
| 06:15 |
jyfl987 |
then you cant explain why it slow on local server |
| 06:15 |
OldCoder |
Is it? |
| 06:15 |
celeron55 |
if somebody has any idea how to do it better, it's fairly easy to add |
| 06:15 |
|
Calinou joined #minetest-dev |
| 06:16 |
OldCoder |
celeron55, I've actually designed protocol stacks but it was a long time ago. Could look at it next year. |
| 06:16 |
celeron55 |
minetest has it's own reliability layer on top of UDP (which probably everybody knows) |
| 06:16 |
OldCoder |
In lieu of TCP I assume |
| 06:17 |
jyfl987 |
yep, but the protocol.txt still empty |
| 06:17 |
OldCoder |
Calinou, Hi |
| 06:17 |
celeron55 |
OldCoder: take a quick look at Peer::reportRTT() in connection.cpp |
| 06:17 |
OldCoder |
Proceeding now |
| 06:18 |
celeron55 |
it currently just switches between speeds based on raw average ping... which works sometimes, but usually not too well |
| 06:18 |
OldCoder |
Reading |
| 06:18 |
celeron55 |
or, well, varies the speed |
| 06:18 |
jyfl987 |
so its a soft limitation? |
| 06:18 |
OldCoder |
Not too long |
| 06:19 |
OldCoder |
Heuristic |
| 06:19 |
OldCoder |
Not limitation but heuristic |
| 06:19 |
OldCoder |
Could use more comments |
| 06:20 |
OldCoder |
Would be easy enough to fiddle with these hardcoded values |
| 06:20 |
celeron55 |
ideally it should measure the derivative of RTT to speed, and settle to the point where RTT starts to raise too much |
| 06:20 |
OldCoder |
I can't produce an intelligent solution myself until next year |
| 06:20 |
OldCoder |
Wait; I can probably do that |
| 06:20 |
* OldCoder |
sighs |
| 06:20 |
OldCoder |
celeron55, for reasons alluded to recently I can't do much until 2013 |
| 06:21 |
OldCoder |
But this looks like fun |
| 06:21 |
OldCoder |
So track RTT |
| 06:21 |
OldCoder |
Approximate derivative |
| 06:21 |
OldCoder |
Adjust and see how that affects the derivative? |
| 06:21 |
OldCoder |
Something like that? |
| 06:21 |
jyfl987 |
celeron55: can you add a `locale` item to configure so that in win32 we could also use our own locale? |
| 06:22 |
celeron55 |
jyfl987: i have done and won't do anything; ask someone else, who then should report to eg. thexyz |
| 06:22 |
jyfl987 |
i have opened debug, and found in win7 , the locale is not like zh_CN(my locale on linux), its about `simplyfied chinese`, that's not good |
| 06:22 |
celeron55 |
i just hang around as a random consultant currently 8) |
| 06:22 |
jyfl987 |
celeron55: you have done what? |
| 06:23 |
OldCoder |
celeron55, random consultants can be quite useful :-) |
| 06:23 |
celeron55 |
jyfl987: nothing |
| 06:23 |
jyfl987 |
give up for this project? |
| 06:23 |
OldCoder |
jyfl987, No |
| 06:23 |
OldCoder |
jyfl987, He is acting as an advisor |
| 06:23 |
celeron55 |
jyfl987: i use my coding efforts on other projects |
| 06:23 |
OldCoder |
jyfl987, things will move forward with time |
| 06:23 |
celeron55 |
because minetest isn't the only thing in the world (suprising, huh?) |
| 06:24 |
jyfl987 |
celeron55: so you wont maintain the project? |
| 06:24 |
* OldCoder |
chuckles |
| 06:24 |
OldCoder |
jyfl987, see above. It is fine. |
| 06:24 |
jyfl987 |
who is the new maintainer? |
| 06:24 |
OldCoder |
jyfl987, the group |
| 06:24 |
OldCoder |
He has opened the master sources to three people |
| 06:24 |
OldCoder |
Who will coordinate contributions from the community |
| 06:24 |
jyfl987 |
good, i want to suggest again , is it possible to use c instead of cpp |
| 06:24 |
OldCoder |
That is too major a change ATM |
| 06:24 |
OldCoder |
Possibly in the future |
| 06:25 |
OldCoder |
I could handle that |
| 06:25 |
jyfl987 |
and using a vm to instead the lua scripting api |
| 06:25 |
Calinou |
OldCoder: huh? opened the master sources? did you mean three people got commit access? |
| 06:25 |
OldCoder |
Calinou, IIRC |
| 06:25 |
darkrose |
jyfl987: no |
| 06:25 |
OldCoder |
I think they are darkrose thexyz and PilzAdam |
| 06:25 |
celeron55 |
need to go and shut down this temporary connection though -> |
| 06:25 |
Calinou |
probably not, else there would've been commits recently :P |
| 06:25 |
OldCoder |
Calinou, this was very recent |
| 06:25 |
thexyz |
Calinou: of course he did |
| 06:26 |
jyfl987 |
ok , just using c, this will made me feel better |
| 06:26 |
OldCoder |
jyfl987, I don't know if the group will accept it but I'll consider doing a C version in the future |
| 06:27 |
jyfl987 |
OldCoder: then i need your github's addr, want to trace on it |
| 06:27 |
Calinou |
thexyz should probably merge his branch with the official one |
| 06:27 |
OldCoder |
jyfl987, If I return to full time development O.K. |
| 06:27 |
OldCoder |
Calinou, Key changes will be merged. For example LevelDB. |
| 06:27 |
Calinou |
why not all of them? |
| 06:27 |
thexyz |
Calinou: i should firstly merge all my branches into my master one |
| 06:27 |
Calinou |
ah |
| 06:27 |
OldCoder |
jyfl987, My GitHub is on my website: |
| 06:28 |
OldCoder |
http://oldcoder.org/ |
| 06:28 |
thexyz |
then wait for somebody more experienced in c++/minetest to review my changes |
| 06:28 |
OldCoder |
There is no MT work there ATM |
| 06:28 |
jyfl987 |
OldCoder: oop |
| 06:28 |
* jyfl987 |
i need to work, its afternoon |
| 06:28 |
OldCoder |
jyfl987, Have fun |
| 06:30 |
OldCoder |
jyfl987, If you wish to try out LevelDB visit Calinou world at port 30003 on my main server VPS |
| 06:31 |
Calinou |
k |
| 06:31 |
jyfl987 |
OldCoder: i hope he wont install many mods |
| 06:31 |
OldCoder |
Calinou, BTW I'd like to find out how people get out of the spawn trap. Something about a trampoline? |
| 06:31 |
OldCoder |
jyfl987, Calinou? It is a reasonable set |
| 06:31 |
jyfl987 |
i have tried redcrab's server on aws, which is too lag |
| 06:31 |
OldCoder |
I have many mods and not too much lag yet |
| 06:31 |
jyfl987 |
anyway its my problem , i live behind the GFW |
| 06:31 |
OldCoder |
I delete misbehaving mods |
| 06:32 |
OldCoder |
Great Firewall? |
| 06:32 |
jyfl987 |
yep |
| 06:32 |
darkrose |
thexyz: where are you on github, url? |
| 06:32 |
jyfl987 |
welcom to china |
| 06:32 |
OldCoder |
If I move to China I will set up servers there |
| 06:32 |
jyfl987 |
OldCoder: now where r u? |
| 06:32 |
OldCoder |
jyfl987, Silicon Valley but I must leave soon |
| 06:32 |
OldCoder |
My worlds will be down for a week |
| 06:32 |
OldCoder |
I have been invited to China; I may try there |
| 06:33 |
thexyz |
darkrose: https://github.com/minetest/minetest |
| 06:33 |
jyfl987 |
OldCoder: ok, if you came to china, be good luck :[ |
| 06:33 |
OldCoder |
jyfl987, why, thank you |
| 06:33 |
darkrose |
guys, chatter about servers should go elsewhere |
| 06:33 |
darkrose |
thexyz: ta |
| 06:33 |
jyfl987 |
OldCoder: many foreighner came to china especially beijing, and i now live here, but the air is too bad |
| 06:34 |
OldCoder |
darkrose, you are not presently ignored. Therefore a response: Server was mentioned in conjunction with testing LevelDB. That is a core dev issue. |
| 06:34 |
OldCoder |
I do think LevelDB needs testing |
| 06:34 |
jyfl987 |
you could check some reporting from US embassy |
| 06:34 |
OldCoder |
jyfl987, that should be in the other channel; she would be right about that |
| 06:34 |
OldCoder |
Embassies are not core dev :-) |
| 06:35 |
jyfl987 |
ok , back to leveldb, i think a more clear protocol is important |
| 06:35 |
OldCoder |
How so? |
| 06:35 |
jyfl987 |
need documents |
| 06:35 |
OldCoder |
jyfl987, do come back in the next few weeks to speak with me elsewhere BTW. I need advice regarding China. PM me when you can. |
| 06:36 |
jyfl987 |
OldCoder: i think there're 2 or 3 websites which give advice for foreigor living in china, but anyway, i will give mine if you need |
| 06:36 |
OldCoder |
PM later |
| 06:36 |
jyfl987 |
ok |
| 06:59 |
OldCoder |
jyfl987, I have PMd you |
| 07:33 |
|
celeron55 joined #minetest-dev |
| 07:50 |
|
jyfl987 joined #minetest-dev |
| 08:07 |
|
celeron55 joined #minetest-dev |
| 08:13 |
|
Calinou joined #minetest-dev |
| 09:34 |
|
celeron55_ joined #minetest-dev |
| 09:59 |
|
PilzAdam joined #minetest-dev |
| 10:33 |
|
SpeedProg joined #minetest-dev |
| 11:52 |
|
Jeija joined #minetest-dev |
| 12:38 |
|
nyuszika7h joined #minetest-dev |
| 13:06 |
|
nyuszika7h joined #minetest-dev |
| 13:08 |
|
hmmmm joined #minetest-dev |
| 13:09 |
|
jyfl9871 joined #minetest-dev |
| 13:11 |
jyfl9871 |
OldCoder: hi where is your server? |
| 13:11 |
hmmmm |
.... |
| 13:11 |
hmmmm |
WTF does that have to do with engine development? |
| 13:12 |
OldCoder |
Hi |
| 13:13 |
OldCoder |
jyfl987, that is not core dev I assume. Therefore I will answer in the other channel. |
| 13:13 |
jyfl9871 |
ok i will be attention of the rule here |
| 13:13 |
OldCoder |
jyfl987, I have posted an answer in #minetest |
| 14:22 |
|
Jeija left #minetest-dev |
| 14:51 |
|
nyuszika7h joined #minetest-dev |
| 15:19 |
|
PilzAdam joined #minetest-dev |
| 16:31 |
OldCoder |
PilzAdam, hello |
| 16:57 |
|
darkrose joined #minetest-dev |
| 16:57 |
|
darkrose joined #minetest-dev |
| 17:28 |
|
Calinou joined #minetest-dev |
| 17:42 |
PilzAdam |
Suggestion: move the code of falling items to builtin, so not every game has to code it |
| 17:43 |
PilzAdam |
also you dont have to depend on default if you call nodeupdate(pos) |
| 17:46 |
darkrose |
and if a game doesn't want falling things? |
| 17:46 |
PilzAdam |
they dont have to define the falling group |
| 17:48 |
PilzAdam |
s/define/use |
| 17:48 |
darkrose |
sounds fine in theory then |
| 17:50 |
celeron55_ |
hmm... i think i should document somewhere how minetest releases are generally made |
| 17:50 |
celeron55_ |
because it kind of requires a bunch of non-trivial things |
| 18:14 |
PilzAdam |
https://github.com/celeron55/minetest/pull/270 |
| 18:15 |
PilzAdam |
https://github.com/celeron55/minetest_game/pull/17 |
| 19:12 |
VanessaE |
so seeing as how the dev survey is already closed, what were the questions on it? |
| 19:12 |
* VanessaE |
is curious. |
| 19:13 |
thexyz |
"what would you like to include to minetest" |
| 19:14 |
VanessaE |
oh. |
| 19:14 |
darkrose |
and "what have you already done" |
| 20:43 |
OldCoder |
VanessaE, Hi |
