DDraceNetwork - developer

what if you haven’t played teworlds for a long time, then I went in and didn’t show the server ddrays

@Rafael Fontenelle @Ryozuki the AUR ddnet/ddnet-git packages should depend on sqlite

done

def-

[ddnet/ddnet] Issue opened: #2558 Can't load old saves after SQLite update

@Zwelf Could you take a look? *** Unable to load savegame: data corrupted Happens on every old save.

Zwelf: https://github.com/ddnet/ddnet/issues/2558

def-

[ddnet/ddnet] Issue opened: #2559 Teamranks don't get saved after SQLite update

@Zwelf Can you take a look? Seems like ranks got saved, but team ranks not.

Zwelf

[ddnet/ddnet] Pull request opened: #2560 Fix team rank not updating after SQLite update

Fixes #2559

09:02

def-

[ddnet/ddnet] Issue closed: #2559 Teamranks don't get saved after SQLite update

09:02

def-

[ddnet/ddnet] Pull request closed: #2560 Fix team rank not updating after SQLite update

09:02

def-

[ddnet:master] 2 new commits

a866c3e Fix team rank not updating after SQLite update - Zwelf c41826d Merge pull request #2560 from Zwelf/pr-fix-teamrank-not-updating - def-

Zwelf: Hm, I guess I should look for all team ranks and create them manually now, right?

That would be nice. If you still have the logs, you can grep for the string "found team rank from same team" and update the times.

bors[bot]

[ddnet/ddnet] Pull request closed: #2557 Add SQLite as dependency in README

09:09

bors[bot]

[ddnet:master] 2 new commits

3e0e4bd Add SQLite as dependency in README - def- 4928d52 Merge #2557 - bors[bot]

or alternatively go through the record_race and find things that look like team ranks but has no equivalent in record_teamrace?

not trivial to write that though is it?

09:10

hm, I guess you can group by timestamp

grepping the logs isn't trivial either, they get overwritten, might miss something

yes going through ranks would be good enough as well

oh well, not so many ranks actually, can do it by hand. 11 for now

09:12

but servers will only restart slowly, so I'll have to monitor

09:13

we should have some test scenarios with sqlite to test the sql code

09:13

before with a real db it was annoying, with sqlite would be nicer to build

(CI should also at least build mysql)

yeah, but then you need to start a real mysql server etc

If we can abstract out a "RaceManager" we could add a lot of tests

yes, but without building with mysql we get build errors like the one you reported earlier

sure, but build errors are not so bad

09:15

failing to save ranks is bad

yes

Zwelf: You're also not printing out the sql queries anymore, right?

09:19

Makes it hard to find out what the map even was, I liked seeing the SQL in there, would also have been easier to see the mistake then

Our logs need some work, I think it's very noisy to be useful

i prefer slightly too noisy

09:20

can still grep

With the prepared statements, there are only ? in the sql statement, so it isn't that useful anymore.

@deen that's true

Zwelf: Isn't there a way to print the prepared statement with the bound values?

Searched for it a bit, but didn't find a way to do this.

is the game id printed to log somewhere?

teehistorian name is printed

ah, good enough

(which contains the game id)

but this is still quite some manual work collecting all this info instead of copy-pasting the sql from log

the game id should also be in the (non-team) ranks btw

def-

[ddnet/ddnet] Issue opened: #2561 ID in record_teamrace is broken

@Zwelf Another one for you I guess: ID in record_teamrace contains weird characters when a team save is inserted for the first time: «[7SLÊ\nk+ vs previously: 912b616a-d1ae-11. Probably the ID shouldn't be a blob?

It's mind boggling how shitty the documentation for the cpp connector is

yeah, might as well skip using it and use straight C

would drop a dependency \o/

would also reduce windows problems with the cpp connector

might also be easier with the nice abstraction now

I wonder if there is a nice wrapper library that would let us use any sql implementation rather than just mysql

Zwelf found one, but it was GPL

09:30

ed

I can't reproduce the /load problem locally and would like to compare Savestrings. Could you send me the output of SELECT Savegame FROM record_saves WHERE Map = "Just2Easy" AND Code = "gear verse shock";

The C library does indeed look much better documented

Zwelf: https://ddnet.tw/sqlout.txt

oh ye sqlite rly good for testing

09:54

it even has a memory db option

09:55

im also for removing cpp connector

def-

[ddnet/ddnet] Pull request opened: #2562 Update language files ("Client message")

As noticed by TortiLeq

https://dev.mysql.com/doc/c-api/8.0/en/c-api-function-overview.html

Mh, copying mysql output doesn't seem to work well for tabs at the end of the line. Could you repeat the command and base64 encode the output: SELECT to_base64(Savegame) FROM record_saves WHERE Map = "Just2Easy" AND Code = "gear verse shock";

done, same link

thanks

10:12

Hm, I can load this string locally. I have no idea yet why it fails to load on ger.ddnet.tw.

locally with sqlite or mariadb?

10:14

different mariadb versions? different default encoding

10:15

I guess I can add some debug printing and try it on GER

Using mariadb (on ubuntu 20.04)

@Deleted User https://github.com/ddnet/ddnet/pull/2553#issuecomment-672784157 pls take a look

Add OpenGL error handling by Jupeyy · Pull Request #2553 · ddnet/dd...

Don't merge yet, since i don't know if this should be merged at all. (only for OpenGL3 commands rn) Might help with #2551, if the driver knows the error(e.g. texture incompleteness)...

Zwelf: I tried dbg_msg("sql", aSaveString); and get a very cut off string:

[2020-08-12 12:19:52][sql]: 2   1       1       0
deen    1       0       34      0       1       0       -1      0       1       0       -1      0       1       0
       0       0       0       0       0       0       0       0       0       0       P

10:23

And I have no idea where that P in the end comes from, that's not in the DB

https://bugs.mysql.com/bug.php?id=94503

@deen you probably should have printed "%s", aSaveString to be sure

10:31

Though I doubt aSaveString contains %s unless someone saved with malicious intent

yeah, was my own save

Where did you add the print?

After the strcpy(pData->m_pResult->m_aMessage, "Unable to load savegame: data corrupted");

10:36

We have the exact same version of mysqlcppconnector, so I guess we have the same bug

10:36

have to compile my own mysqlcppconnector I guess, no newer one in Debian

10:37

Why didn't we get this before, didn't we use getString?

10:43

Confirmed, works

Someone pls increase timeout protection on ger2

10:58

@deen

def-

[ddnet-scripts:master] 1 new commit

4d020de Add mysqlcppconn workaround - def-

nope, can't connect to it either

give rcon, im on server

11:10

def-

[ddnet/ddnet] Issue closed: #2558 Can't load saves after SQLite update

@deen should we maybe whitelist master4 again?

what do you mean @Learath2 ?

I thought it was flooding with server ips, looks like randomly spoofed

ah, sure. where did we get the ips?

11:52

there was this favorites file somewhere, wasn't therE?

I think we just took all the servers ips

yeah, but where from?

hm, with the masters dead I actually am not sure

found it: https://status.tw/server/dump/6/

11:54

i will add it to whitelist

11:54

lis2 packages?

fw!!, bea2

pulic ddos mitigation discussion xd

whatever, it's obvious

11:56

imagine being so much of a cuck that you do spoofed attacks on an old games masterserver

yes 0.6 is very old

11:56

imagine using master server without tokens

11:56

:p

imagine being dimwitted enough to think 0.7 master servers are immune to spoofing attacks

:/

We should have migrated to http masters

1

11:58

Who cares if 0.7 follows us anyway? We can implement it and use it

sounds good

12:00

btw @Learath2 could you try to explain https://github.com/ddnet/ddnet/commit/444bd934377e3cc63dac45516a1220dcecb3209f#diff-1dd7d837c25bee6c3d340a8e8f388230 for a 6yo?

Fix recvmmsg again... · ddnet/ddnet@444bd93

recvmmsg doesn't use the same buffer as normal recv

12:01

pData points to the correct one in either case

Did I do it right?

nah you seem to have broken serverstatus too

12:01

oh wait

12:01

looks good actually

weird

hm i hate coding :/

12:02

i give 100 euros to the person who ports ddnet client to 0.7

12:03

200 if its good and merged into ddnet/ddnet

At 20e per hour that's only 5 hours of work

12:03

If it was only 5 hours of work I'd do it for free

then hurry up

i got it connecting ^

I don't like ports anyway

but merging with ddnet upstream becomes more annoying every time

We should clean it up and do it right while going up to 0.7

@Learath2 why is status not working?

@deen it wasn't working for a second then it started working again

12:05

ChillerDragon: I actually should have all the tools you need to make ddnet client able to connect to 0.7 servers in place

wdym by tools?

12:05

that current ddnet6 code base is 7ish enough?

The protocol is generated

yea i started working on my ddnet port before you did all the stuff and now i rage quitted the merge

12:06

anyways i dont get ur code

12:06

so much

:(

I cant build ddnet on windows

12:52

since the 0.7 layer got merged

did you recreate your build folder? there were issues with cmake not finding python3

python3 found, yes. folders all there, I use MSVS

Then show error

Not at home right now, but the error is something with the get_msgs and the generated maps I believe

12:53

something with compile.py i think

12:54

i will look for it later

Jupeyy

[ddnet/ddnet] Pull request closed: #2553 Add OpenGL error handling

@Deleted User i thought arb debug was widely used and available tho