| Time |
Nick |
Message |
| 01:03 |
hmmmm |
home from school |
| 01:03 |
hmmmm |
now... back to work |
| 01:04 |
hmmmm |
so what i ended up doing was having emergeBlock optionally generate a blank block (filled with CONTENT_IGNORE) and the emerge thread already takes care of the case where isGenerated() == false |
| 01:05 |
hmmmm |
i believe this is the optimal solution because it fits in the nicest with the rest of the code |
| 01:14 |
* VanessaE |
peeks in |
| 01:43 |
celeron55 |
https://github.com/celeron55/minetest/tree/tcp_connection |
| 01:43 |
celeron55 |
buggy as hell |
| 01:43 |
celeron55 |
it does work sometimes though, i haven't yet undersood which times |
| 01:44 |
celeron55 |
+t |
| 01:44 |
celeron55 |
somebody could set up a server with the ambience mod and see how it loads |
| 01:45 |
celeron55 |
it seems to work for me if i remove something from the cache prior to running; which doesn't make any sense, but anyway |
| 01:45 |
celeron55 |
otherwise no mapblocks will transfer at all |
| 01:45 |
celeron55 |
... |
| 01:45 |
RealBadAngel |
i will try |
| 03:26 |
RealBadAngel |
we have compiled it, made a server and wasnt even able to connect |
| 03:27 |
RealBadAngel |
server is runnin on minetest.org 30030 |
| 04:15 |
|
doserj joined #minetest-dev |
| 04:42 |
hmmmm |
man, GetNextBlocks still calculates the client's view area |
| 04:43 |
hmmmm |
whatever happened to that guy who was supposed to move that calculation out into the client? wasn't he basically finished with that task? why isn't it merged? |
| 05:47 |
|
doserj joined #minetest-dev |
| 06:06 |
celeron55 |
hmmmm: basically he didn't even remotely finish it |
| 06:06 |
celeron55 |
hmmmm: i have worked on it (and my work can be found from my repo), but it still has problems |
| 06:16 |
celeron55 |
but indeed combinin that with this could be full of win |
| 06:16 |
celeron55 |
+g |
| 06:32 |
hmmmm |
err... he looked like he did a lot with it. and from what i understand all that had to be done was to move some code around and add a field to the request blocks packet |
| 06:33 |
hmmmm |
what are the problems exactly? like bugs, or design flaws? |
| 07:10 |
celeron55 |
well, i basically made 5 commits worth of fixes, tweaking and further development, and then ran a test on a server with thexyz and somebody |
| 07:10 |
celeron55 |
it kind of worked, but had some odd pauses in transferring and didn't overally perform better than what there currnently is |
| 07:11 |
celeron55 |
i don't remember if i was able to figure out at all what the problems were |
| 07:14 |
celeron55 |
(as of how he left it, it was completely unusable) |
| 07:15 |
celeron55 |
(a good start though) |
| 08:00 |
celeron55 |
i tried to merge these; i've now got a funny problem of the new draw code now trying to access blocks that haven't been already loaded and the client_requests_blocks_2 code requiring it for deciding what to request |
| 08:00 |
celeron55 |
8D |
| 08:00 |
celeron55 |
not trying to access* |
| 08:03 |
celeron55 |
well, this is fixable |
| 08:07 |
|
Calinou joined #minetest-dev |
| 08:20 |
celeron55 |
https://github.com/celeron55/minetest/commits/tcp_blocks |
| 08:20 |
celeron55 |
works quite well |
| 08:24 |
|
saschaheylik joined #minetest-dev |
| 08:33 |
celeron55 |
i'll put this on my server with the ambience mod for testing |
| 08:40 |
celeron55 |
apparently the server screws something up |
| 08:59 |
celeron55 |
i guess it's because i left proper handling of send() out 8) |
| 09:01 |
celeron55 |
hmm, actually no, it's not |
| 09:20 |
celeron55 |
oh cool, now it works to some extent |
| 09:21 |
celeron55 |
it just oddly times out while sending if i put in the ambience mod... |
| 10:00 |
celeron55 |
great success \:D/ |
| 10:00 |
celeron55 |
it's a hack but works almost perfectly now |
| 10:02 |
celeron55 |
...except when it doesn't |
| 10:04 |
celeron55 |
c55.me:30001 |
| 10:07 |
celeron55 |
ehm... well yeah, it's still useless |
| 10:08 |
celeron55 |
it jams randomly |
| 10:54 |
celeron55 |
it probably needs a more thorough redesign if somebody is to make it work without exploding their brain |
| 10:55 |
celeron55 |
something along the lines of making TCP the primary thing and UDP a secondary thing; currently it just pops up the TCP connection on the side after it knows the peer id and routes stuff through it that was asked to be transferred reliably |
| 13:09 |
|
SpeedProg joined #minetest-dev |
| 13:17 |
celeron55 |
one thing i'd be eager to do is add an interface to the built-in TCP socket in Lua 8) it could be used for lots of weird things |
| 13:17 |
celeron55 |
not sure if for anything useful though |
| 13:52 |
|
rsiska joined #minetest-dev |
| 14:08 |
|
hmmmm joined #minetest-dev |
| 14:08 |
|
ecube joined #minetest-dev |
| 14:44 |
|
doserj joined #minetest-dev |
| 15:08 |
|
doserj joined #minetest-dev |
| 15:22 |
|
PilzAdam joined #minetest-dev |
| 15:45 |
|
Calinou joined #minetest-dev |
| 15:56 |
celeron55 |
urgh... yeah, i can't understand at all why and how this gets stuck |
| 15:57 |
* celeron55 |
appends network rework on an imaginary infinite to-do list |
| 16:38 |
Calinou |
echo "network rework" >> /home/celeron55/TODO |
| 16:38 |
Calinou |
du -h /home/celeron55/TODO |
| 16:39 |
Calinou |
2,9M TODO |
| 16:39 |
celeron55 |
over 9000 |
| 16:57 |
|
sstrandberg joined #minetest-dev |
| 18:57 |
|
rubenwardy joined #minetest-dev |
| 19:18 |
hmmmm |
did you try looking at it in a packet logger yet...? |
| 20:36 |
|
rubenwardy left #minetest-dev |
| 21:41 |
|
rsiska joined #minetest-dev |
| 22:03 |
hmmmm |
oh wow |
| 22:03 |
hmmmm |
i want to be able to put images on signs |
| 22:03 |
hmmmm |
that would be AWESOME |
| 22:05 |
hmmmm |
i'll have to do that after i'm finished with this |
| 22:05 |
hmmmm |
the armor bar thing can wait |
