04:47 <+bridge> [ddnet] <Smetanolub> Me
04:50 <+bridge> [ddnet] <Smetanolub> Imagine, you race hard long map, and after 3 hours walkthrough you both get timeout and can't reconnect(doesn't matter why)
04:50 <+bridge> [ddnet] <Smetanolub> Why there are no autosave on ddnet servers?
04:54 <+bridge> [ddnet] <Alexander> Tbf if you were in a team and got timeout then an autosave would be great
08:53 <+bridge> [ddnet] <Vy0x2> Because we have no accounts, so how do we assure that the correct person gets the save password?
10:50 <+ChillerDragon> is that a real story? imo it does matter why you can not reconnect. Timeout usually works great for me :) 
12:30 <+bridge> [ddnet] <Anime.pdf> I usually cant reconnect cuz electricity goes down :justatest:
12:55 <+bridge> [ddnet] <Ryozuki> me and bois building the pc
12:55 <+bridge> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1046031501487243304/image.png
15:06 <+bridge> [ddnet] <Smetanolub> make autogenerated password before player's internet dies, then when both did timeout-save and write that save to ddnet-saves.txt
15:19 <+bridge> [ddnet] <Smetanolub> imagine ddos and server dies, or your internet fells off for half a day. you cant recconect always
15:21 <+bridge> [ddnet] <Vy0x2> So timeout but its forever and the tees disappear instantly?
15:24 <+bridge> [ddnet] <Vy0x2> I guess the reasons against this are the same like for /save in t0.
15:24 <+bridge> [ddnet] <Vy0x2> Skips..
15:45 <+bridge> [ddnet] <maria_rodriguez> Start small with a minimum of $50 for BITCOIN MINING, answer YES to be added 👇👇👇
15:45 <+bridge> [ddnet] <maria_rodriguez> 
15:45 <+bridge> [ddnet] <maria_rodriguez> https://t.me/Stevecoldham001
17:04 <+bridge> [ddnet] <Alexander> Doesn't /save already show everyone the save password
17:19 <+bridge> [ddnet] <Learath2> https://blog.cloudflare.com/speeding-up-linux-disk-encryption/
17:20 <+bridge> [ddnet] <Learath2> Quite funny how the dm-crypt mailing list told them they don’t understand
17:24 <+bridge> [ddnet] <Learath2> I might try this, apparently it made it into the kernel. It may help explain why I only get like 30M/s writing into a luks partition on an ssd
17:36 <+bridge> [ddnet] <Ryozuki> @Learath2 just skimmed but are they putting the improvements to upstream?
17:36 <+bridge> [ddnet] <Ryozuki> also as always excelent blog
17:36 <+bridge> [ddnet] <Learath2> Already did, it’s on 5.9
17:36 <+bridge> [ddnet] <Ryozuki> nice
17:37 <+bridge> [ddnet] <Learath2> Tldr is a lot of outdated legacy code to accomodate for the limitations of the past
17:37 <+bridge> [ddnet] <Ryozuki> yeah
17:37 <+bridge> [ddnet] <Ryozuki> kernel stack or smth
17:37 <+bridge> [ddnet] <Ryozuki> 4 queues
17:37 <+bridge> [ddnet] <Ryozuki> xd
17:37 <+bridge> [ddnet] <Learath2> E.g. bad random read/write speed of traditional drives. Kernel stack size limits etc
18:43 <+bridge> [ddnet] <heinrich5991> meh, bad mailing list apparently 😦
18:44 <+bridge> [ddnet] <Learath2> I expect a lot of these older mailing lists full of kernel hackers and the compiler people are all similar
18:44 <+bridge> [ddnet] <heinrich5991> telling others you know better when you do is already iffy
18:44 <+bridge> [ddnet] <heinrich5991> but if you don't, that's just bad
18:46 <+bridge> [ddnet] <heinrich5991> > But for the sake of stability and further benchmarking we ended up not removing the actual code, but rather adding yet another dm-crypt option, which bypasses all the queues/threads, if enabled.
18:46 <+bridge> [ddnet] <heinrich5991> yay, options \o/ :p
18:46 <+bridge> [ddnet] <heinrich5991> I also had problems with dm-crypt devices in the past
18:46 <+bridge> [ddnet] <heinrich5991> never debugged them though
18:46 <+bridge> [ddnet] <Learath2> In this case options might actually be good, since their approach might not be strictly better for all workloads
18:47 <+bridge> [ddnet] <Learath2> E.g. a traditional spinning rust drive does indeed have a significant difference in sequential vs random performance
18:49 <+bridge> [ddnet] <Learath2> Does anyone know how I can fix an ext4 partition to patch the size?
18:49 <+bridge> [ddnet] <heinrich5991> to expand or shrink?
18:49 <+bridge> [ddnet] <Learath2> I imaged a partition and the new one I created seems to be like a couple bytes shorter for some reason
18:50 <+bridge> [ddnet] <Learath2> Now when I write the image onto the partition fsck complains that the physical size doesnt match the fs size
18:51 <+bridge> [ddnet] <heinrich5991> `resize2fs` sounds interesting?
18:58 <+bridge> [ddnet] <Learath2> Yep, that worked
18:58 <+bridge> [ddnet] <Learath2> I’m never imaging a drive like this again, such a bad idea 😄
19:01 <+bridge> [ddnet] <heinrich5991> why? ^^
19:01 <+bridge> [ddnet] <heinrich5991> why is it a bad idea?
20:06 <+bridge> [ddnet] <Learath2> @heinrich5991 well it doesn't check bounds of the partition, what I did wouldn't have worked with a completely full partition e.g.
20:54 <+bridge> [ddnet] <Patiga> no, only you see the password
20:54 <+bridge> [ddnet] <Alexander> Ah okay then
21:02 <+bridge> [ddnet] <heinrich5991> it shows everyone from the team the password, no?
21:03 <+bridge> [ddnet] <Patiga> true, not the entire server though
21:03 <+bridge> [ddnet] <Patiga> ah I might've not read enough context
21:21 <+bridge> [ddnet] <Izanagi> me and my mate trying to figure out koule maps
22:23 <+bridge> [ddnet] <Portak Abi> @noby