DDraceNetwork - developer

it's over 9000

in 15.5 I can spam refresh and create tons of tasks. and sometime trigger some io error.

[ddnet/ddnet] Issue opened: #3856 15.5rc1 browser refresh button can be spammed to fill up the job queue

In case of impatience and master is responding slowly, it might create more unnecessary load on the master. Plus, it makes the list seemingly more slow since the list only updates after the last http request is done.

[2021-05-28 12:08:54][http]: http https://info2.ddnet.tw/info?name=fokkonaut
[2021-05-28 12:08:54][http]: http https://master2.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:54][http]: task done https://info2.ddnet.tw/info?name=fokkonaut
[2021-05-28 12:08:55][http]: task done https://master2.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:55][http]: http https://master2.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:57][http]: task done https://master2.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:57][serverbrowse_http]: found master, url='https://master2.ddnet.tw/ddnet/15/servers.json' time=2437ms
[2021-05-28 12:08:57][http]: http https://master4.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:57][http]: task failed. libcurl error: Could not resolve host: master4.ddnet.tw
[2021-05-28 12:08:57][http]: http https://master3.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:57][http]: task failed. libcurl error: Could not resolve host: master3.ddnet.tw
[2021-05-28 12:08:57][http]: http https://master1.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:57][http]: task done https://master1.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:57][http]: http https://master1.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:58][http]: task done https://master1.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:58][serverbrowse_http]: found master, url='https://master1.ddnet.tw/ddnet/15/servers.json' time=201ms
[2021-05-28 12:08:58][serverbrowse_http]: determined best master, url='https://master1.ddnet.tw/ddnet/15/servers.json' time=201ms
[2021-05-28 12:08:58][http]: http https://master1.ddnet.tw/ddnet/15/servers.json
[2021-05-28 12:08:58][http]: task done https://master1.ddnet.tw/ddnet/15/servers.json

That is a bit spammy on startup, no?

It's a bit spammy on IRC

1

10:11

it seems ok for me

def-

[ddnet/ddnet] Pull request opened: #3857 [http]: http -> [http]: fetching

Checklist

[ ] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test if it works standalone, system.c especially
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--u...

def-

[ddnet/ddnet] Pull request opened: #3858 Reduce spamminess of http server browser

As reported by fokkonaut on Discord. Not sure if we lose relevant information. Previously: ``` [2021-05-28 12:08:54][http]: http https://master2.ddnet.tw/ddnet/15/servers.json [2021-05-28 12:08:55][http]: task done https://master2.ddnet.tw/ddnet/15/servers.json [2021-05-28 12:08:55][http]: http https://master2.ddnet.tw/ddnet/15/servers.json [2021-05-28 12:08:57][http]: task done https://master2.ddnet.tw/ddnet/15/servers.json [2021-05-28 12:08:57][serverbrowse_http]: found master, ur...

Http requests should get an id so we can associate results to requests

13:51

(In the log)

CMake Error (dev) at CMakeLists.txt:236 (message):
  DATA does not contain every file from directory data

14:58

15:02

oh ok i got it

bors[bot]

[ddnet:master] 2 new commits

badaae5 [http]: http -> [http]: fetching - def- 67611a5 Merge #3857 - bors[bot]

15:13

bors[bot]

[ddnet/ddnet] Pull request closed: #3857 [http]: http -> [http]: fetching

bors[bot]

[ddnet:master] 2 new commits

bcdb3e0 Reduce spamminess of http server browser - def- 7e9b0fc Merge #3858 - bors[bot]

15:23

bors[bot]

[ddnet/ddnet] Pull request closed: #3858 Reduce spamminess of http server browser

@heinrich5991 could we add something to the protocol to let the server signal that it's experiencing connectivity issues?

@Learath2 atomic integer counter?

One thing I've been missing with the new serverlist is the ability to see that a server is under ddos

heinrich5991

@Learath2 atomic integer counter?

Yeah, that's what I was thinking too for the ID

add a field to the server info once we have the true http master?

Doesn't the underlying collection server have some idea of the ping?

maybe also improve the client connecting dialog

15:29

@Learath2 does the ping increase for servers under attack? I'd guess at most the packet loss rate increases

hm, I've seen GER2 spike to double the ping before during an attack, but not all attacks do that yeah

I should have used a template parameter for the alignment...

15:56

I should stop committing quickly

Ah, I didn't because I didn't want to move the implementation of Allocate and AllocateFromChunk into the header file

16:05

When are they going to get rid of that silly requirement anyway?

or of header files in general. are C++ modules a step into that direction?

not really afaik

mh, I guess I'll have to leave it like this, a shame since aligning to 16 optimizes extremely well