00:09 <+bridge> <deen> not sure, seems like someone is running a weird bot
07:57 <+bridge> <Ryozuki> someone doing automated tests on gentoo guru found this
07:57 <+bridge> <Ryozuki> https://bugs.gentoo.org/888875
07:57 <+bridge> <Ryozuki> idk how much of a problem it is
07:57 <+bridge> <Ryozuki> but i think its an issue with cmake itself not the ebuild
07:58 <+bridge> <Ryozuki> `Bug 888875 - [guru] games-action/ddnet-16.6 overwrites CFLAGS/CXXFLAGS or adds uncommon ones `
07:58 <+bridge> <Ryozuki> here is the ebuild if ur curious https://gitweb.gentoo.org/repo/proj/guru.git/tree/games-action/ddnet/ddnet-16.6.ebuild
07:59 <+bridge> <Ryozuki> their build log https://888875.bugs.gentoo.org/attachment.cgi?id=845660
07:59 <+bridge> <Ryozuki> ```
07:59 <+bridge> <Ryozuki> * Tinderbox QA Notice: This package seems to overwrite CFLAGS/CXXFLAGS or add uncommon ones:
07:59 <+bridge> <Ryozuki>  *   The following flags (probably added by the build system) should be stripped out
08:00 <+bridge> <Ryozuki>  * 
08:00 <+bridge> <Ryozuki>  *   1) -D_FORTIFY_SOURCE=2
08:00 <+bridge> <Ryozuki>  *   2) -fstack-protector-all
08:00 <+bridge> <Ryozuki>  * 
08:00 <+bridge> <Ryozuki>  * Final size of build directory: 175416 KiB (171.3 MiB)
08:00 <+bridge> <Ryozuki>  * Final size of installed tree:   66044 KiB ( 64.4 MiB)
08:00 <+bridge> <Ryozuki> ```
08:01 <+bridge> <Ryozuki> @heinrich5991 what do u think?
09:53 <+bridge> <deen> Yes, we do that, for security
09:53 <+bridge> <deen> I didn't know it's bad style to add security related flags
09:54 <+bridge> <deen> It's a bit like complaining our source code has runtime check for array out of bounds access because they are too expensive
10:16 <+bridge> <Learath2> I think the gentoo stance is to make all security flags optional if they impact perf
10:31 <+bridge> <Ryozuki> they are probable controlled by portage
10:31 <+bridge> <Ryozuki> probably
10:31 <+bridge> <Ryozuki> its more that the package forces this and the user cant choose with their portage flags or smth
10:36 <+bridge> <Ryozuki> this is a QA by a random user btw
10:37 <+bridge> <Ryozuki> i did run pkgcheck on the ebuild when making it and it didnt throw any warnings
11:00 <+bridge> <chillerdragon> Who does out of bounds check? Ubsan? Where is that active?
11:00 <+bridge> <chillerdragon> (@deen)
11:01 <+bridge> <chillerdragon> Oh wait or do you mean actual handwritten code that checks stuff like if (ClientID \> MAX\_CLIENTS) return
11:04 <+bridge> <Ryozuki> ye
14:39 <+bridge> <deen> Alright, added an option for them @Ryozuki ^
14:59 <+bridge> <Ryozuki> tyty
14:59 <+bridge> <Ryozuki> not just any random user, its a gentoo dev lol
16:32 <+bridge> <Ryozuki> https://blog.cloudflare.com/opaque-oblivious-passwords/
16:35 <+bridge> <heinrich5991> I thought there were such protocols before OPAQUE
16:46 <+bridge> <Learath2> Kerberos, ms chap?
16:50 <+bridge> <HahahoupasGR> ad
16:53 <+bridge> <Jupstar ✪> Tl;dr whats the alternative ?
16:54 <+bridge> <Ryozuki> https://github.com/jbangert/trapcc
16:54 <+bridge> <Ryozuki> lol
16:54 <+bridge> <Ryozuki> its zero proof password auth iirc
16:54 <+bridge> <heinrich5991> you can prove that you have the password without sending it to the server
16:54 <+bridge> <Ryozuki> u send a proof u know the password
16:54 <+bridge> <Ryozuki> so the password is never sent
16:55 <+bridge> <Jupstar ✪> Ok but u still save ur pw in a key manager which u host online in case ur house burns. Does it solve this too
16:55 <+bridge> <Ryozuki> dont host it online
16:55 <+bridge> <Ryozuki> just replicate it 3 times locally
16:55 <+bridge> <Ryozuki> invent a fire proof usb
16:55 <+bridge> <Ryozuki> kek
16:55 <+bridge> <Learath2> Well when you host it online you usually host it completely encrypted
16:56 <+bridge> <Learath2> So your plaintext pw never makes it anywhere of your computer, theoretically
16:57 <+bridge> <Jupstar ✪> Yes. So the blog is about the pw being plain text on the server?
16:57 <+bridge> <Ryozuki> http://mainisusuallyafunction.blogspot.com/2014/02/x86-is-turing-complete-with-no-registers.html
16:58 <+bridge> <Learath2> I didn't read the entire thing yet, but yes, that's the idea
16:58 <+bridge> <heinrich5991> yes, basically about preventing that
16:59 <+bridge> <Jupstar ✪> I see thanks for saving me time xd
17:00 <+bridge> <Ryozuki> so u dont read everything i post
17:00 <+bridge> <Ryozuki> noted
17:00 <+bridge> <Ryozuki> :feelsbadman:
17:01 <+bridge> <Ryozuki> @Learath2 @Jupeyy_Keks https://startafuckingblog.com/
17:01 <+bridge> <Ryozuki> :justatest:
17:02 <+bridge> <Ryozuki> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1027108
17:02 <+bridge> <Ryozuki> its happening
17:03 <+bridge> <Ryozuki> (removing python 2 from debian)
17:03 <+bridge> <Ryozuki> ddnet-scripts needs fixing
17:03 <+bridge> <Ryozuki> it uses python2 xD
17:03 <+bridge> <Learath2> Rip python2, you won't be missed
17:03 <+bridge> <heinrich5991> I tried updating it once
17:04 <+bridge> <heinrich5991> not sure what problems I encountered
17:04 <+bridge> <Learath2> Veery big problem with ddnet-scripts is the disk caches we use
17:05 <+bridge> <Learath2> The upgrade needs to be instant and recalculate everything from scratch. Python3 version of diskcache chokes hard on the python2 saved caches
17:05 <+bridge> <deen> @Jupeyy_Keks hm, do we really want to have translations for the 50 or so error messages for graphics? 😄
17:05 <+bridge> <deen> How about we use materialized views for the entire website and cache nothing? I might be motivated to do that
17:06 <+bridge> <heinrich5991> what does "materialized views" mean?
17:06 <+bridge> <Learath2> Now this is a group of words idk. Materialized views?
17:06 <+bridge> <Learath2> Ah sql ones
17:06 <+bridge> <Learath2> Yeah, I had something similar in mind
17:06 <+bridge> <deen> Yes, SQL, a view stores the query to access a table, then you can easily select from it and the more complex query is executed
17:07 <+bridge> <deen> A materialized view doesn'T just store the query but also the result, so you can actually get it faster
17:07 <+bridge> <HahahoupasGR> fokkonaut
17:07 <+bridge> <HahahoupasGR> fokkonaut you nice server
17:07 <+bridge> <heinrich5991> is that always up-to-date? or can you tell it to lag behind to save processing?
17:07 <+bridge> <HahahoupasGR> Roolpo on me
17:07 <+bridge> <heinrich5991> is that always up-to-date? or can you tell it to lag behind to save processing time?
17:07 <+bridge> <HahahoupasGR> 😄
17:07 <+bridge> <HahahoupasGR> 🙂
17:07 <+bridge> <heinrich5991> @HahahoupasGR use DMs
17:07 <+bridge> <deen> It lags behind in the current implementation in all common databases (mysql, postgres)
17:07 <+bridge> <heinrich5991> nice
17:08 <+bridge> <Ryozuki> oh thats nice
17:08 <+bridge> <deen> There is work in postgres to auto-update it
17:08 <+bridge> <heinrich5991> I like that
17:08 <+bridge> <Ryozuki> so this is like a on-db cache?
17:08 <+bridge> <heinrich5991> sounds like a good use case for a db ^^
17:08 <+bridge> <HahahoupasGR> ^ ^
17:08 <+bridge> <HahahoupasGR> a
17:08 <+bridge> <Ryozuki> stop spamming
17:08 <+bridge> <HahahoupasGR> 🙂
17:08 <+bridge> <Ryozuki> ur annoying
17:08 <+bridge> <HahahoupasGR> ❤️
17:08 <+bridge> <deen> And I'm going to start working on a product that does these materialized views for an existing database: https://materialize.com/
17:08 <+bridge> <deen> So I might start implementing something, it would be a postgres backend though
17:09 <+bridge> <Ryozuki> postgres best
17:09 <+bridge> <deen> (auto-updating instantly)
17:09 <+bridge> <Ryozuki> as long as what we use is open source
17:09 <+bridge> <Ryozuki> i dont mmind
17:09 <+bridge> <Learath2> I never checked how these work. When does it get updated?
17:09 <+bridge> <Learath2> Manually?
17:09 <+bridge> <deen> Yes, you run a `refresh` command
17:09 <+bridge> <Learath2> And does refresh lock?
17:09 <+bridge> <heinrich5991> then it drops everything and the DB lags for 10 minutes?
17:10 <+bridge> <deen> https://github.com/sraoss/pg_ivm is the repo for the in-progress work for incremental mat views
17:10 <+bridge> <Ryozuki> ohh incremental too
17:10 <+bridge> <Jupstar ✪> I dunno. Better than answering the question in question every time
17:10 <+bridge> <Learath2> Oh and yes that, is refresh incremental currently?
17:10 <+bridge> <deen> It shouldn't when you use `concurrently`: https://www.postgresql.org/docs/current/sql-refreshmaterializedview.html
17:11 <+bridge> <deen> Right now every refresh in mysql/postres is a full recalculation, thus expensive
17:12 <+bridge> <Learath2> I can imagine that being a pretty hard problem to solve for dbms ppl
17:12 <+bridge> <deen> Yes, that's the hope with Materialize, that it's too complex and better solved in an on-top-solution where your changes are streamed (via Change Data Capture)
17:13 <+bridge> <deen> This blog post on the topic is a nice motivation: https://lironshapira.medium.com/data-denormalization-is-broken-7b697352f405
17:14 <+bridge> <Ryozuki> medium meh
17:14 <+bridge> <Ryozuki> it always looks so uggly
17:16 <+bridge> <deen> But when we get the error message in Russian it will be harder for us to understand 😄 Can the user do anything with `Submitting to graphics queue failed.` or `No vulkan queue family properties found.`?
17:17 <+bridge> <deen> The ones with an action like `Failed during initialization. Try to change gfx_backend to OpenGL or Vulkan from settings_ddnet.cfg in the config directory and try again.` seem helpful
17:18 <+bridge> <Jupstar ✪> Mh hard part is just to distinguish between both
17:18 <+bridge> <Jupstar ✪> Then I've to extend the error strings by a is translatable string
17:19 <+bridge> <Jupstar ✪> Open am issue and tag me then i do it soon: tm:
17:23 <+bridge> <heinrich5991> alternatively, include an error code
17:23 <+bridge> <heinrich5991> rustc does it
17:23 <+bridge> <heinrich5991> V-0001
17:23 <+bridge> <heinrich5991> to V-0015
17:27 <+bridge> <heinrich5991> did you notice it's just a couple of numbers short of 888888?
17:27 <+bridge> <heinrich5991> newest bug rn is https://bugs.gentoo.org/888885
17:40 <+bridge> <Ryozuki> true lol
17:46 <+bridge> <Jupstar ✪> Yeah could do that too^^
17:47 <+bridge> <Jupstar ✪> Let's create some ddnet issue xdd
17:48 <+bridge> <deen> You are too late, some issues are not visible because they might have been deleted or are security relevant
17:48 <+bridge> <murpi> I've got a sample of a failed map download, anyone interested in debugging this?
17:48 <+bridge> <deen> They are higher already @heinrich5991 https://bugs.gentoo.org/888890
17:49 <+bridge> <deen> Got full logs? You could put them in an issue
17:50 <+bridge> <gerdoe> https://bugs.gentoo.org/888888 ?
17:50 <+bridge> <gerdoe> 404 for me
17:50 <+bridge> <murpi> I just got the map and the temp file, that's it
17:50 <+bridge> <deen> Yes, because the issue is deleted or you don't have permissions to see it
17:50 <+bridge> <gerdoe> got it
17:50 <+bridge> <gerdoe> thanks xd
17:51 <+bridge> <deen> They might have a policy where a crash report is always marked as possibly security relevant initially, and only unmarked when a dev took a look
18:13 <+bridge> <HahahoupasGR> a
18:13 <+bridge> <HahahoupasGR> hey
18:13 <+bridge> <HahahoupasGR> hi
18:14 <+bridge> <Ryozuki> @heinrich5991 this guy has only been spamming now
18:14 <+bridge> <Ryozuki> maybe do smth
18:14 <+bridge> <HahahoupasGR> sry it was my littlbe brother
18:15 <+bridge> <HahahoupasGR> I'm gonna remove his disc 😄
18:19 <+bridge> <Mr.Gh0s7> @heinrich5991 Can you ban him, it's ok if not? It's my little brother. He won't let me remove his disc rn. I forgot to uninstall it from his windows partition. I have done so from his linux one. The above messages are the proof you need tbh.
18:21 <+bridge> <Mr.Gh0s7> @heinrich5991 Can you ban him, it's ok if not? It's my little brother. He won't let me remove his disc rn. I forgot to uninstall it from his windows partition. I have done so from his linux one. The above messages are the proof you need tbh (to know his my lil bro).
18:28 <+bridge> <gerdoe> i can start spreading misinformation to troll but the situation is already too funny xd
18:29 <+bridge> <Mr.Gh0s7> D: I could take his hardrive or his whole pc but banning him should appear to him as it should. He did smth wrong and was punished for it by a 3rd party.
19:30 <+ChillerDragon> @Mr.Gh0s7 omg is that roolpo? is that ur lil brota? can you please unbind his chat key or teach him some english bruu none of his sentences make any sense 
19:31 <+bridge> <Mr.Gh0s7> lol sry. Ok will do. Idk if he knows how to binds keys tho
21:21 <+bridge> <Robyt3> Should we update this list of releases: https://github.com/ddnet/ddnet/blob/6d3af742dd1aec107cffcd2fe49d80eaea973e19/other/ddnet.appdata.xml#L40-L83 ? Or should we just get rid of the list so we don't have to maintain it?
22:15 <+bridge> <deen> not sure, what is it good for?
22:20 <+bridge> <Robyt3> It would show a list of releases with their release dates in package managers that use this type of file
22:21 <+bridge> <Robyt3> Not really useful to maintain this separately IMO. We can just link to the downloads page that contains all releases.