07:22 < bridge> That just sounds like the average llvm experience xD 07:22 < bridge> Utilizing it for your lang? 07:22 < bridge> Or learning it just for the sake of it 07:33 < bridge> just building it 07:33 < bridge> @blaiszephyr 08:59 < bridge> :brownbear: 09:13 < bridge> i remember trying to build llvm for risc v shit took way too long 09:16 < bridge> @robyt3 should I fix #10307 first or should I keep working on the pipeline instead, since it fixes it? 09:16 < bridge> https://github.com/ddnet/ddnet/issues/10307 10:03 < bridge> Pipeline 10:55 < bridge> morning 11:02 < bridge> morning ruyozuki 13:37 < bridge> :what: 13:37 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1379786082761314414/f0e2e081c7483056.png?ex=6841816d&is=68402fed&hm=36a033ade584b4aef80defca28e7ebe34f8f159df0fbdbc62522430828a83f1b& 13:40 < bridge> Probably because the path contains a backslash instead 13:41 < bridge> I think the check also doesn't work if you open the file from externally 13:42 < bridge> Only works if you open the map with the file dialog in the editor directly 13:43 < bridge> yeah i just checked the code 13:43 < bridge> maybe should instead call some storage api to check it 13:52 < bridge> Did you open the map file by double clicking it in the file explorer? 15:53 < bridge> I FINALLY GOT IT TO HAPPEN ON A BUILD WITH DEBUG INFO 15:53 < bridge> ``` 15:53 < bridge> void CUuidManager::RegisterName(int Id, const char *pName) 15:53 < bridge> { 15:53 < bridge> dbg_assert(GetIndex(Id) == (int)m_vNames.size(), "names must be registered with increasing ID (got=%d want=%d)", GetIndex(Id), (int)m_vNames.size()); 15:53 < bridge> CName Name; 15:53 < bridge> Name.m_pName = pName; 15:53 < bridge> Name.m_Uuid = CalculateUuid(pName); 15:53 < bridge> dbg_assert(LookupUuid(Name.m_Uuid) == -1, "duplicate uuid"); 15:53 < bridge> ``` 15:53 < bridge> duplicate 15:53 < bridge> duplicate uuid for some reason 15:54 < bridge> ```cpp 15:54 < bridge> #0 dbg_assert_imp (filename=filename@entry=0x5555559aed78 "/home/solly/Code/ddnet/src/engine/shared/uuid_manager.cpp", line=line@entry=125, fmt=fmt@entry=0x5555559c6a2c "otal") 15:54 < bridge> at /home/solly/Code/ddnet/src/base/system.cpp:153 15:54 < bridge> #1 0x000055555560cf26 in CUuidManager::RegisterName (this=this@entry=0x555555b062c0 , Id=Id@entry=65546, pName=pName@entry=0x5555559c6362 "invalid") 15:54 < bridge> at /home/solly/Code/ddnet/src/engine/shared/uuid_manager.cpp:125 15:54 < bridge> #2 0x00005555556052cd in RegisterUuids (pManager=pManager@entry=0x555555b062c0 ) at /home/solly/Code/ddnet/src/engine/shared/protocol_ex_msgs.h:33 15:55 < bridge> #3 0x00005555555aa759 in CreateGlobalUuidManager () at /home/solly/Code/ddnet/src/engine/shared/global_uuid_manager.cpp:10 15:55 < bridge> #4 __static_initialization_and_destruction_0 () at /home/solly/Code/ddnet/src/engine/shared/global_uuid_manager.cpp:16 15:55 < bridge> #5 _GLOBAL__sub_I_g_UuidManager () at /home/solly/Code/ddnet/src/engine/shared/global_uuid_manager.cpp:16 15:55 < bridge> #6 0x00007ffff38377e4 in call_init (argc=1, argv=0x7fffffffe098, env=) at ../csu/libc-start.c:145 15:55 < bridge> #7 __libc_start_main_impl (main=0x5555555aa880 , argc=1, argv=0x7fffffffe098, init=, fini=, rtld_fini=, stack_end=0x7fffffffe088) 15:55 < bridge> at ../csu/libc-start.c:347 15:55 < bridge> #8 0x00005555555bd535 in _start () 15:55 < bridge> ``` 15:56 < bridge> no clue how this is being caused 15:57 < bridge> ` Name = {m_Uuid = {m_aData = { }}, m_pName = 0x5555559c6362 "invalid"}` 15:57 < bridge> so 2 invalids caused a duplicate? 16:02 < bridge> I dragged and dropped it to a open client 16:14 < bridge> how on earth did it get "invalid"??? 16:15 < bridge> anyone good at debugging? 16:21 < bridge> I'd say I'm pretty decent. What are we debugging? 16:22 < bridge> debug assert after exit by ctrl-z then kill -9 16:22 < bridge> sometimes 16:22 < bridge> i cant really go from the top since idk how the error is happening but ill try 16:23 < bridge> should i dm you 16:23 < bridge> or spam in #developer some more 16:23 < bridge> Spam more 16:23 < bridge> okay 16:23 < bridge> ` dbg_assert(LookupUuid(Name.m_Uuid) == UUID_INVALID, "duplicate uuid %s", Name.m_pName);` is failing 16:23 < bridge> in `void CUuidManager::RegisterName(int Id, const char *pName)` 16:24 < bridge> for some reason pName is "invalid" here (i cant find any disasembly or where this "invalid" comes from) 16:24 < bridge> but for it to be a duplicate it would have to already be registered 16:25 < bridge> ``` 16:25 < bridge> int CUuidManager::LookupUuid(CUuid Uuid) const 16:25 < bridge> { 16:25 < bridge> CNameIndexed Needle; 16:25 < bridge> Needle.m_Uuid = Uuid; 16:25 < bridge> Needle.m_Id = 0; 16:25 < bridge> auto Range = std::equal_range(m_vNamesSorted.begin(), m_vNamesSorted.end(), Needle); 16:25 < bridge> if(std::distance(Range.first, Range.second) == 1) 16:25 < bridge> { 16:25 < bridge> return GetId(Range.first->m_Id); 16:25 < bridge> } 16:25 < bridge> return UUID_UNKNOWN; 16:25 < bridge> } 16:25 < bridge> ``` 16:25 < bridge> im not sure how this determines duplicates 16:25 < bridge> but its searching `m_vNamesSorted` 16:25 < bridge> which i can print 16:25 < bridge> ``` 16:25 < bridge> $29 = std::vector of length 10, capacity 16 = {{m_Uuid = {m_aData = "\bL&\250\r\3426\032\247\333\026\234\024\262F\236"}, m_Id = 9}, {m_Uuid = {m_aData = "\"bd\210r\3075#\226\230\032\034\221\343#J"}, 16:25 < bridge> m_Id = 5}, {m_Uuid = {m_aData = "#\321m\276x\255:\203\213\265o\223\222)\377", }, m_Id = 1}, {m_Uuid = {m_aData = "1\243\320\333R\3156\005\205\f-\275\r\363|\246"}, m_Id = 7}, { 16:25 < bridge> m_Uuid = {m_aData = "3\352\226\340z\206;\003\227K\316\343\341o\245\241"}, m_Id = 3}, {m_Uuid = {m_aData = "q\225\206\351\024\245;{\257(N\365\240\252\002\\"}, m_Id = 6}, {m_Uuid = { 16:25 < bridge> m_aData = "\277\367\313\334\303&>퀡J+\a\224p\003"}, m_Id = 4}, {m_Uuid = {m_aData = "\305L\f]6\3333\353\254/\374\375%", }, m_Id = 8}, {m_Uuid = { 16:25 < bridge> m_aData = "\3141\034\337\314T4ӑ\005b\264\005[\215H"}, m_Id = 2}, {m_Uuid = {m_aData = "ӫc\3138\237?\237\267/\332 SK\000\266"}, m_Id = 0}} 16:25 < bridge> ``` 16:25 < bridge> sadly it doesnt store the names 16:26 < bridge> ``` 16:26 < bridge> print CalculateUuid("invalid") 16:26 < bridge> $30 = {m_aData = "\bL&\250\r\3426\032\247\333\026\234\024\262F\236"} 16:26 < bridge> ``` 16:26 < bridge> but if there were a duplicate it would be this 16:26 < bridge> id = 9 16:26 < bridge> ``` 16:26 < bridge> 16:26 < bridge> UUID(NETMSG_WHATIS, "what-is@ddnet.tw") 16:26 < bridge> UUID(NETMSG_ITIS, "it-is@ddnet.tw") 16:26 < bridge> UUID(NETMSG_IDONTKNOW, "i-dont-know@ddnet.tw") 16:26 < bridge> 16:26 < bridge> UUID(NETMSG_RCONTYPE, "rcon-type@ddnet.tw") 16:26 < bridge> UUID(NETMSG_MAP_DETAILS, "map-details@ddnet.tw") 16:26 < bridge> UUID(NETMSG_CAPABILITIES, "capabilities@ddnet.tw") 16:26 < bridge> UUID(NETMSG_IAMTATER, "iamtater@sjrc6.github.io") 16:26 < bridge> UUID(NETMSG_TATER_CHECKSUM_REQUEST, "checksum-request@sjrc6.github.io") 16:26 < bridge> UUID(NETMSG_TATER_CHECKSUM_RESPONSE, "checksum-response@sjrc6.github.io") 16:26 < bridge> UUID(NETMSG_CLIENTVER, "clientver@ddnet.tw") 16:26 < bridge> UUID(NETMSG_PINGEX, "ping@ddnet.tw") 16:26 < bridge> UUID(NETMSG_PONGEX, "pong@ddnet.tw") 16:26 < bridge> ``` 16:27 < bridge> (pname should be each one of these) 16:27 < bridge> the id starts from 0 and goes up 16:27 < bridge> so "9" is `UUID(NETMSG_TATER_CHECKSUM_RESPONSE, "checksum-response@sjrc6.github.io")` 16:27 < bridge> ```cpp 16:27 < bridge> (gdb) print CalculateUuid("checksum-response@sjrc6.github.io") 16:27 < bridge> $31 = {m_aData = "\305L\f]6\3333\353\254/\374\375%", } 16:27 < bridge> ``` 16:27 < bridge> which is this... im not sure what incomplete sequence means 16:27 < bridge> but it might be something something memory corruption 16:27 < bridge> and thats where i am at 16:27 < ws-client> @avolicious any plans for fixing the crashes? GER3 #14 45.141.57.45:8313 just crashed ._. 16:28 < bridge> incomplete sequence means its not clean utf-8, i dont expect an md5 hash to, so this is fine 16:29 < bridge> when 9 was called, similar to now (it fails on "pong@ddnet.tw") it must have been "invalid" somehow 16:29 < bridge> when 9 was called, similar to now (it fails on "pong@ddnet.tw") it must have been magiced into "invalid" somehow 16:30 < bridge> sorry ping@ddnet.tw 16:30 < bridge> is where it crashed 16:31 < bridge> brrp no i cant count 16:31 < bridge> its `UUID(NETMSG_CLIENTVER, "clientver@ddnet.tw")` 16:32 < bridge> (pro tip never try to autocomplete in gdb) 16:32 < bridge> (its frozen...) 16:32 < bridge> oh im an idiot 16:33 < bridge> this doesnt collide aswell 16:33 < bridge> when i do print &"invalid" it gives me the same address 16:33 < bridge> so somewhere it is defined 16:33 < bridge> but its not a symbol 16:34 < bridge> oh but theres multiple "invalids" 16:35 < bridge> ``` 16:35 < bridge> (gdb) print &"invalid" 16:35 < bridge> $39 = (char (*)[8]) 0x555555dac120 16:35 < bridge> ``` and ``` 16:35 < bridge> m_pName = 0x5555559c6362 "invalid"} 16:35 < bridge> ``` 16:36 < bridge> i can dissasemble le function and so ive got the address for every one 16:36 < bridge> thanks mr gdb for commenting what the value in memory is 16:37 < bridge> idk why that isnt inlined 16:37 < bridge> load the value directly into the register 16:38 < bridge> so the address which is being passed in is correct with the one the bt shows for the argument 16:38 < bridge> ```asm 16:38 < bridge> 0x00005555556052bc <+204>: mov $0x1000a,%esi 16:38 < bridge> 0x00005555556052c1 <+209>: lea 0x3c109a(%rip),%rdx # 0x5555559c6362 16:38 < bridge> 0x00005555556052c8 <+216>: call 0x55555560cb10 <_ZN12CUuidManager12RegisterNameEiPKc> 16:38 < bridge> => 0x00005555556052cd <+221>: mov %rbx,%rdi 16:38 < bridge> ``` 16:38 < bridge> but when i look at the source code 16:39 < bridge> ah i wish i could see source and dissasembly at same time 16:43 < bridge> i can,,, but i cant scroll very easily 16:47 < bridge> ``` 16:47 < bridge> 0x5555559c62c3 ----------- 16:47 < bridge> 0x5555559c62d4 hot data size invalid 16:47 < bridge> 0x5555559c62e3 nvalid 16:47 < bridge> 0x5555559c62f8 items 16:47 < bridge> 0x5555559c630b large 16:47 < bridge> 0x5555559c6320 nge 16:47 < bridge> 0x5555559c6336 item types 16:47 < bridge> 0x5555559ae128 checksum-request@sjrc6.github.io 16:47 < bridge> 0x5555559ae150 checksum-response@sjrc6.github.io 16:48 < bridge> 0x5555559c634f invalid 16:48 < bridge> 0x5555559c6362 invalid 16:48 < bridge> ``` 16:48 < bridge> here is what each passed in arg is 16:48 < bridge> what the absolute fuck 16:50 < bridge> ``` 16:50 < bridge> 0x5555559c62c3 ----------- != what-is@ddnet.tw 16:50 < bridge> 0x5555559c62d4 hot data size invalid != it-is@ddnet.tw 16:50 < bridge> 0x5555559c62e3 nvalid != i-dont-know@ddnet.tw 16:50 < bridge> 0x5555559c62f8 items != rcon-type@ddnet.tw 16:50 < bridge> 0x5555559c630b large != map-details@ddnet.tw 16:50 < bridge> 0x5555559c6320 nge != capabilities@ddnet.tw 16:50 < bridge> 0x5555559c6336 item types != iamtater@sjrc6.github.io 16:50 < bridge> 0x5555559ae128 checksum-request@sjrc6.github.io == checksum-request@sjrc6.github.io 16:50 < bridge> 0x5555559ae150 checksum-response@sjrc6.github.io == checksum-response@sjrc6.github.io 16:50 < bridge> 0x5555559c634f invalid != clientver@ddnet.tw 16:50 < bridge> 0x5555559c6362 invalid != ping@ddnet.tw 16:50 < bridge> ``` 16:50 < bridge> they line up??? 16:50 < bridge> what did checksum request and response do to deserve not being butchered 17:02 < bridge> IT HAPPENS IN ASAN 17:37 < bridge> can you give tldr? 17:37 < bridge> 17:37 < bridge> if it crashes in an assert you should have assert logs in your configdir/dumps folder 17:37 < bridge> ~~IT HAPPENS IN ASAN~~ 17:37 < bridge> i know what the assertion is 17:37 < bridge> theres no tldr, its something memory corruption 17:38 < bridge> ah k, then why is asan cancelled out xd 17:39 < bridge> because my attempt at code to debug it failed, and caused the issue 17:39 < bridge> it doesnt happen in asan 17:39 < bridge> not that ive made it 17:39 < bridge> so the assert is not part of a memory corruption? 17:39 < bridge> 17:39 < bridge> how do you assume a memory corruption otherwise? 17:40 < bridge> the assert is correct as far as i can tell 17:40 < bridge> me reading the dissasembly versus what it should be 17:40 < bridge> the last 2 cause the "duplicate assert" 17:41 < bridge> whats suprising is that theyre all ascii 17:42 < bridge> i can give u the binary and the core dump 17:42 < bridge> does it happen reproducably? 17:42 < bridge> nope 17:43 < bridge> it happens sometimes when i exit uncleanly (normally by ctrl-c spam.. it not exiting, me ctrl-z and kill -9 %1) 17:46 < bridge> mh 17:47 < bridge> i'd suspect independent code at the moment tbh, makes no sense to me 17:47 < bridge> `CreateGlobalUuidManager` creates an instance in a local variable.. i don't see how this should ever lead to duplicated entries 17:47 < bridge> (reproducable) 17:57 < bridge> you reproduced it? 17:57 < bridge> but i get no errors under asan 18:05 < bridge> i dont even know which code you working on, tater master? 18:06 < bridge> yeah 18:10 < bridge> Then I cannot reproduce it, TSan & ASan both clean on latest tater master 18:10 < bridge> i know 18:10 < bridge> Do you press ctrl+c in terminal? 18:10 < bridge> yeah 18:10 < bridge> Where you connected to a game? 18:10 < bridge> server 18:10 < bridge> doesnt matter 18:10 < bridge> ive had it in and out 18:12 < bridge> did you try valgrind? 18:12 < bridge> to find unintialized values 18:12 < bridge> not yet 18:15 < bridge> No 18:21 < bridge> The only thing I found was this 18:21 < bridge> ``` 18:21 < bridge> Syscall param socketcall.sendto(msg) points to uninitialised byte(s) 18:21 < bridge> ==274962== at 0x85579EE: __syscall_cancel_arch (syscall_cancel.S:56) 18:21 < bridge> ==274962== by 0x854C667: __internal_syscall_cancel (cancellation.c:49) 18:21 < bridge> ==274962== by 0x854C6AC: __syscall_cancel (cancellation.c:75) 18:21 < bridge> ==274962== by 0x85CF263: sendto (sendto.c:27) 18:21 < bridge> ==274962== by 0x302B5A: net_udp_send(NETSOCKET_INTERNAL*, NETADDR const*, void const*, int) (in TaterClient-ddnet/build/DDNet) 18:21 < bridge> ==274962== by 0x2E2838: CNetBase::SendPacket(NETSOCKET_INTERNAL*, NETADDR*, CNetPacketConstruct*, int, bool, bool) (in TaterClient-ddnet/build/DDNet) 18:21 < bridge> ==274962== by 0x2E3C6B: CNetTokenCache::SendPacketConnless(CNetChunk*) (in TaterClient-ddnet/build/DDNet) 18:21 < bridge> ==274962== by 0x2E4758: CNetClient::Send(CNetChunk*) (in TaterClient-ddnet/build/DDNet) 18:21 < bridge> ==274962== by 0x3BF3F4: CServerBrowser::Refresh(int, bool) (in TaterClient-ddnet/build/DDNet) 18:21 < bridge> ==274962== by 0x444805: CMenus::RefreshBrowserTab(bool) (in TaterClient-ddnet/build/DDNet) 18:21 < bridge> ==274962== by 0x44688A: CMenus::RenderMenubar(CUIRect, IClient::EClientState) (in TaterClient-ddnet/build/DDNet) 18:21 < bridge> ==274962== by 0x446FCB: CMenus::Render() (in TaterClient-ddnet/build/DDNet) 18:21 < bridge> ==274962== Address 0x1ffeffbbbc is on thread 1's stack 18:21 < bridge> ==274962== in frame #5, created by CNetBase::SendPacket(NETSOCKET_INTERNAL*, NETADDR*, CNetPacketConstruct*, int, bool, bool) (???:) 18:21 < bridge> ==274962== Uninitialised value was created by a stack allocation 18:21 < bridge> ==274962== at 0x2E3B01: CNetTokenCache::SendPacketConnless(CNetChunk*) (in TaterClient-ddnet/build/DDNet) 18:22 < bridge> ``` 18:22 < bridge> 18:22 < bridge> And that sounds more like a ddnet bug 18:22 < bridge> If it's not false positive, but it only showed this for connless packets 18:22 < bridge> that is after the uuid crash 18:23 < bridge> reproduce it with asan on 😉 18:43 < bridge> @avolicious: didn’t you say the root cause was OOM without memory leak? Doesn’t that mean you just started too many game servers on one node? Can’t you turn some of them off and it’s fixed? 18:44 < bridge> No, other system resources takes more memory than usual, so it kills the servers in the first place 18:45 < bridge> Do you have a memory limit on the pods? Can’t you increase that? 18:46 < bridge> We cant increase 18:47 < bridge> But doesn’t the same logic apply? 18:47 < bridge> Your servers are packed too full 18:48 < bridge> Might be 18:48 < bridge> It works out for 95% of the time, downscaling is currently not planned 18:50 < bridge> I mean decrease 18:50 < bridge> Decrease the load by turning off things that take up ram 18:50 < bridge> increase swap file size 18:51 < bridge> How much ram do you have @avolicious? 18:52 < bridge> Those 5% are unacceptable imo 18:52 < bridge> I lost some pretty long good runs. Was hella frustrating. And now I play in constant fear. 18:53 < bridge> what's the whackommended amount of dedotaded wam 18:53 < bridge> Makes kog was less attractive. How much ram do you have available and what’s running? 18:54 < bridge> There should be a way to arrange it in a way to run some 20 year old Kirby game without crashing daily with what you have available 18:55 < bridge> Or that ye 18:55 < bridge> Snort 18:55 < bridge> Smort 18:56 < bridge> @avolicious: why did you decide against down scaling? 18:57 < bridge> Makes kog way less attractive. How much ram do you have available and what’s running? 18:58 < bridge> superkai64 Moment fr 18:58 < bridge> I was already scared you wouldn't get the reference 18:59 < bridge> having servers dying mid run sounds absolutely frustrating 18:59 < bridge> oh wow the matrix bridge must be really behind 18:59 < bridge> like \~5mins 18:59 < bridge> Common matrix L 19:01 < bridge> @fushi_gg: good meme dw 19:16 < bridge> Seems fine to fix it in the pipeline if we plan to merge it soon 19:33 < bridge> when cpp20 19:34 < bridge> #6552 19:34 < bridge> https://github.com/ddnet/ddnet/issues/6552 19:35 < bridge> darn 19:35 < bridge> very stale 19:41 < bridge> maybe you can revive i already got 1 edg-l revival 19:42 < bridge> which i need... to revive 19:42 < bridge> its just been sitting there 19:42 < bridge> anyway i said that cuz i thought regex was cpp20 but its 11 19:54 < bridge> w-why are exceptions disabled?? 19:54 < bridge> `Cannot use 'try' with exceptions disabledclang(exceptions_disabled)` 19:55 < bridge> When the oldest distro that we support also supports C++20 19:56 < bridge> unu 19:57 < bridge> ``` 19:57 < bridge> # Disable exceptions as DDNet does not use them. 19:57 < bridge> add_cxx_compiler_flag_if_supported(OUR_FLAGS -fno-exceptions) 19:57 < bridge> ``` 19:57 < bridge> b-but the standard library throws them 20:03 < bridge> just gotta be careful 20:03 < bridge> theres not an exceptionless way to use regex? 20:03 < bridge> theres not an exceptionless way to use std::regex? 20:03 < bridge> sure there is, don’t mess up and it won’t throw 20:03 < bridge> yeah but the user controls what goes into the regex 20:03 < bridge> validate 20:04 < bridge> btw std regex hella slow 20:04 < bridge> i have to do that with another regex library 20:05 < bridge> yeah 20:05 < bridge> ive had std regex code for ddnet before iirc and it was a point of contention 20:05 < bridge> the things its used for arent done every frame 20:05 < bridge> if you have exceptions off and you pass it a bad expression the thing will crash 20:06 < bridge> i know 20:06 < bridge> the STL exceptions go right to the top 20:06 < bridge> so you gotta use constants and test it or do smth else 20:06 < bridge> different regex library i guess 20:06 < bridge> hscan looks the best in terms of performance 20:06 < bridge> pcre2? 20:07 < bridge> regex for what 20:07 < bridge> https://github.com/HFTrader/regex-performance/blob/master/results_threadripper.png 20:07 < bridge> chat filter, conditionals 20:07 < bridge> std::regex is nice because it lets you just std::regex_replace 20:08 < bridge> i would not use regex for that in ddnet. what are u trying to match past words 20:08 < bridge> the user can supply a regex to filter chat 20:08 < bridge> what would they filter past words 20:08 < bridge> currently it uses reimu which is wonky, slow and missing features 20:09 < bridge> whatever they want, i used it to filter spam bots 20:09 < bridge> she’s fine lay off 20:09 < bridge> ? 20:09 < bridge> xD 20:09 < bridge> poor reimu 20:09 < bridge> https://github.com/wareya/Remimu 20:09 < bridge> i’d just add a words list tbh 20:09 < bridge> i cant spell reimu is easier to spell 20:09 < bridge> https://tenor.com/view/reimu-hakurei-reimu-hakurei-reimu-chibi-gif-10765846393329744520 20:09 < bridge> wait wot idt most players would want to use regex over a word list 20:09 < bridge> there’s already sorta behavior like that for the server side censoring 20:09 < bridge> it also doesnt have regex replace 20:10 < bridge> and you can iterate over that and check super ez if it’s a bad word 20:10 < bridge> more performantly than regex 20:10 < bridge> 1. its already there 20:10 < bridge> 2. i use it 20:10 < bridge> less surface for crashing 20:10 < bridge> 1. its already there 20:10 < bridge> 2. i use it 20:10 < bridge> 3. i have write perms i can do what i want 20:10 < bridge> i definitely agree w louis, people already struggle to make a file in their config directory 20:10 < bridge> oh is it for tclient or ddnet 20:10 < bridge> tclient 20:10 < bridge> oh go add it then 20:10 < bridge> xd 20:10 < bridge> the word list ban thing is already a feature in ddnet 20:11 < bridge> o i wasnt aware 20:11 < bridge> sorta... i remember something./.. something local ban list 20:11 < bridge> and server ban list 20:11 < bridge> that’s what im saying, use the same logic for the client to hide the words? 20:11 < bridge> but all that was in the master serv was nothing 20:11 < bridge> but that will be added to ddnet 20:11 < bridge> but im a power user and i want my regex 20:11 < bridge> it’s tclient whatever 20:11 < bridge> but that will be added to ddnet or is added to ddnet or ... its feature accepted atleast 20:11 < bridge> i just dont let words bodda me 20:11 < bridge> tclient sounds cool 20:12 < bridge> idk why it wouldn't be added 20:12 < bridge> spam... doesnt bother you? 20:12 < bridge> oh that might but 20:12 < bridge> breeze em out 20:12 < bridge> doesn't happen on USA 20:12 < bridge> lucky you 20:12 < bridge> xd 20:12 < bridge> well not lucky you 20:12 < bridge> no one plays usa 20:12 < bridge> 20:12 < bridge> not evennn we get a decent avg 20:12 < bridge> Lie 20:12 < bridge> but I wish it was double 20:13 < bridge> both prce2 and hscan are *big* 20:14 < bridge> i guess im sticking with reimu and doing the replace myself 21:35 < bridge> what should i do 21:35 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1379906480802627646/jV552MG.png?ex=6841f18e&is=6840a00e&hm=f3018fcd1aa0961a06e8bf7e7c3914528e0337496a42214ee84c83766f126711& 21:35 < bridge> i has other sv_map in my config 21:35 < bridge> i has another sv_map in my config 21:38 < bridge> dont run ddnet server as root 21:38 < bridge>