DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.tw/irclogs/ Connected with DDNet's IRC channel, Matrix room and GitHub repositories — IRC: #ddnet on Quakenet | Matrix: #ddnet-developer:matrix.org GitHub: https://github.com/ddnet

Between 2020-05-27 00:00:00Z and 2020-05-28 00:00:00Z

weird that such a simple thing is optimized so inconsistently accross compilers

so when using bitset<128> it is basically the same as if __int128 was available?

00:29

I can just typedef the bitset<128> as int128_t and replace it all over the place?

bitset is an array of bools

00:52

and it has functions

00:52

its not the same

but it could do the same job?

bitwise some kind of yes

01:25

is it so horrible to replace it with std::bitset<128>? or just typedef it not int128_t so it doesnt collide with the gcc datatype

[ddnet-maps:master] 1 new commit

d11f443 M Samsara VI - ddnet-maps

Dont bother with gcc and clangs int128 anyway

07:33

For this purpose their emulated 128bit integers optimize to shit

Was reading papers for work and stumbled over this:

All our bug reports for MariaDB were verified quickly. However, only one of them was fixed, which is why we have stopped testing this DBMS.

https://www.manuelrigger.at/dbms-bugs/ Oh well...

Huh

i set my map parse repository to public, if anyone wants to take a look: https://gitlab.com/Patiga/twmap

Map parsing library

Nice @Patiga

[ddnet/ddnet] Pull request closed: #2213 Rename variables for consistency

15:12

[ddnet:master] 2 new commits

fab1cde Rename variables for consistency - Zwelf f670ced Merge #2213 - bors[bot]

[ddnet-web:master] 1 new commit

2cb79a7 Add Win64 client with videorecorder - def-

[ddnet-scripts:master] 1 new commit

95e9657 Add videorecorder build instructions - def-

15:29

[ddnet-web:master] 1 new commit

cc6e633 Update downloads page - def-

15:30

[ddnet/ddnet] Pull request opened: #2215 Version 13.2

Waiting for: https://github.com/ddnet/ddnet/pull/2212 https://github.com/ddnet/ddnet/pull/2202 https://github.com/ddnet/ddnet/pull/2204

[ddnet:master] 2 new commits

e1849ad Implement /practice for teams - def- 6d978a9 Merge #2204 - bors[bot]

15:48

[ddnet/ddnet] Pull request closed: #2204 Implement /practice for teams

[ddnet:master] 1 new commit

6b95bc2 Fix team save practice reset - def-

[ddnet:master] 6 new commits

7f14b4b Add glibc's PRNG - heinrich5991 2692f3c Use a consistent PRNG on all platforms - heinrich5991 b488be0 Actually seed the PRNG on each round - heinrich5991 c9c7f94 Use the PCG-XSH-RR PRNG instead - heinrich5991 9075bc4 Add a comment justifying the modulo for random numbers - heinrich5991 438dc51 Merge #2212 - bors[bot]

16:21

[ddnet/ddnet] Pull request closed: #2212 Use a consistent PRNG on all platforms

Does anyone remember how CNetMsg_Sv_PlayerTime is supposed to work?

16:39

It seems to modify the score, but isn't score part of CNetObj_PlayerInfo?

[ddnet/ddnet-scripts] Issue opened: #21 Python2.7 not suported contexmanager ?

https://github.com/ddnet/ddnet-scripts/blob/95e9657d2567513a51f908ac3b68478344939664/servers/scripts/update-points.py#L25

Traceback (most recent call last):
 File "./update-points.py", line 25, in 
   with con:
ttributeError: __exit__

Whats the solution ?

CNetMsg_Sv_PlayerTime is a relic from the first race client. I doubt any client still relies on it. The ddnet implementation even seems to be incompatible with itself. Server uses hundredths of a second while the client uses seconds.

we hit 10'000 commits!

[ddnet:master] 2 new commits

0d7872c Send DDNet version early in the connection process - heinrich5991 92fc95d Merge #2202 - bors[bot]

17:55

[ddnet/ddnet] Pull request closed: #2202 Send DDNet version early in the connection process

@deen PRs got merged

[ddnet-web:master] 1 new commit

749c07c add Šøø™ - 12pm

18:38

[ddnet/ddnet] Pull request closed: #2215 Version 13.2

18:38

[ddnet:master] 2 new commits

0f02fc5 Version 13.2 - def- 33bc0af Merge pull request #2215 from def-/pr-version-13.2 - def-

18:39

[ddnet/ddnet] Pull request opened: #2216 Remove unused `Sv_PlayerTime` game message

https://www.youtube.com/watch?v=Aymrnzianf0

Making History: NASA and SpaceX Launch Astronauts to Space!

20:03

2 astronauts launching in 30 mins

20:03

justatest

Image attachment

^ cancelled

y

20:22

may 30

[ddnet-web:master] 1 new commit

c72418a Version 13.2 - def-

20:26

[ddnet-web:master] 1 new commit

1c6e96c videorecorder link update - def-

20:26

[ddnet-web:master] 1 new commit

441956e date - def-

20:28

[ddnet-web:master] 1 new commit

b2f1972 date - def-

[ddnet-web:master] 1 new commit

eb9b1a3 Switch markdown processor from redcarpet to kramdown - def-

[ddnet:master] 3 new commits

2097265 Remove unused Sv_PlayerTime game message - heinrich5991 941d3aa Some formatting - heinrich5991 ec6e6f9 Merge #2216 - bors[bot]

20:46

[ddnet/ddnet] Pull request closed: #2216 Remove unused `Sv_PlayerTime` game message

[ddnet-scripts:master] 1 new commit

4a7dd20 No autoupdate for videorecorder build - def-

21:06

[ddnet/ddnet] Pull request opened: #2217 Appdata: add new version

This PR just adds new version into the appdata manifest.

[ddnet:master] 2 new commits

11d3daa Appdata: add new version - ElXreno bf218a0 Merge #2217 - bors[bot]

21:16

[ddnet/ddnet] Pull request closed: #2217 Appdata: add new version

killing practice team leaves the team in practice

21:20

is it normal?

21:20

some people will get fucked because of that, they practice, restart the map with locked team

21:20

they end the map normal, and realize practice is still on

[ddnet/ddnet] Pull request opened: #2218 Make /rescue work in deepfreeze

Thanks to NANAMAR for report

@deen m_SetSavePos && !m_Super && !m_DeepFreeze

21:24

I doubt this will work

I wanted to say the same issue as @Chairn. There is no command to leave practice mod or no server command after crossing start (after killing) that you are still in practice mod to pay attention

@Chairn thanks, fixed in here: https://github.com/ddnet/ddnet/pull/2218

Make /rescue work in deepfreeze by def- · Pull Request #2218 · ddne...

Thanks to NANAMAR for report. Update: tested, works.

21:38

@heinrich5991 i noticed when testing locally

have run count in practice mode of /r is never used?

no

22:14

runs in practice mode do not count

how do you remove it though

22:15

i thought killing would do it

currently: join a different team

lame

rust most loved again

./practice broken on some maps

23:18

such as: Yo'ci (/r doesn't work with deepfreeze)

23:18

additionally, in parts where a powerup is changed, failing past the powerup could mean your whole team fails in practice mode

23:19

for example, if there was a hammerfly part through a freeze to unfreeze, followed by hammer off tile to hookfly, failing past the hookfly would mean your team dies because you no longer have the hammer for the hammerfly part

23:19

Finally, rescue on a lot of insane/brutal 4/5\ maps does not work very well if there are kill tiles in the map because killing would kill everyone and you would need to /r very fast

23:20

the practice mode is very good for some maps, but does not work for others. I think a good way to fix this would to use /setspawn or allow players to /disablekill

@heinrich5991 this looks interesting: https://github.com/calccrypto/uint128_t

calccrypto/uint128_t

C++ unsigned 128 bit integer type. Contribute to calccrypto/uint128_t development by creating an account on GitHub.

23:20

or @Deleted User

For the powerup changes, i think a good way to fix it would to give commands like /enable hammeroff, etc.

[ddnet/ddnet] Issue opened: #2219 TeamMask in pickup.cpp

All team masks in pickup.cpp for the sounds are without passing the asker, so pChr->Teams()->TeamMask(pChr->Team()) should be pChr->Teams()->TeamMask(pChr->Team(), -1, pChr->GetPlayer()->GetCID())

Exported 84 message(s)