01:49 < bridge> <kebscs> https://fxtwitter.com/TikTokInvestors/status/1969570959448498621
01:50 < bridge> <kebscs> :frozen:
02:36 < bridge> <Solly (bun bun)> One looks at those nails and I know she aint doing any work
02:36 < bridge> <Solly (bun bun)> It's near impossible to type with 2cm extra nail
04:53 < bridge> <Solly (bun bun)> Damn dead chat
07:50 < bridge> <chillerdragon> lol crypto bro scammers on GitHub xd never seen that before https://github.com/gitcoactionpromo/gitcoin.com/issues/127
07:50 < bridge> <chillerdragon> They pinged my ass
08:18 < bridge> <Assa> good morning ^^
08:18 < bridge> <Assa> @chillerdragon 10971 is free
08:32 < bridge> <Ewan> you underestimate flat finger laptop keyboard typing
08:32 < bridge> <Ewan> but I don't think anyone at Netflix does a whole lot of anything lol
08:32 < bridge> <Ewan> so I don't disagree
08:56 < bridge> <0xdeen> I don't think most workplaces will be happy if you're filming inside the office, unless you have approval.
08:57 < bridge> <0xdeen> Netflix does cool stuff with performance on FreeBSD last I checked (but that's a while ago)
08:57 < bridge> <Assa> I think I found out when this debug assertion can trigger:
08:57 < bridge> <Assa> ```
08:57 < bridge> <Assa> dbg_assert(Visuals.m_vTilesOfLayer[y * Visuals.m_Width + XR].IndexBufferByteOffset() >= Visuals.m_vTilesOfLayer[y * Visuals.m_Width + X0].IndexBufferByteOffset(), "Tile count wrong.");
08:57 < bridge> <Assa> ```
08:57 < bridge> <Assa> 
08:57 < bridge> <Assa> XR can be smaller than X0, time to build a map for this case and crash the client
08:57 < bridge> <filoqcus> https://tenor.com/view/pipotam-baby-hippo-hippo-wash-shower-gif-3406421565683572068
08:58 < bridge> <Assa> !remind 480 crash client
08:58 < bridge> <0xdeen> https://papers.freebsd.org/2021/eurobsdcon/gallatin-netflix-freebsd-400gbps/
09:04 < ws-client> **<ChillerDragon>** bsd <:poggers2:1008007455936094328>
09:19 < bridge> <Assa> @chillerdragon do you have a better idea for a function called `ShowClip`? It doesn't show anything, it just returns true or false if the clip region says it should render
09:22 < bridge> <Assa> hmm chat gpt suggests `IsVisibleInClipRegion`
09:37 < ws-client> **<ChillerDragon>** @Teero help in https://github.com/ddnet/ddnet/pull/10951
09:37 < ws-client> **<ChillerDragon>** sos
09:37 < bridge> <Assa> AMA
09:38 < bridge> <Assa> just merge, can only be less broken then before tbh
09:38 < ws-client> **<ChillerDragon>** @Assa i feel like i need more context but random guess woule be `ShouldClip`
09:38 < bridge> <Assa> and I wish I was joking ._.
09:41 < bridge> <Assa> I give you some explanation, because this is a bit non-trivial: The hook line actually simulates hook ticks. In the old behavior it was doing it as well. That is also why we need quantization
09:41 < bridge> <Assa> If you don't simulate it, your hook coll line length will be off and your direction as well
09:43 < bridge> <Assa> what is new? I fixed behavior with hook-teleports ( 3 bugs), I stop the simulation after max 250 ticks, added proper handling for tuning values and fixed a small bug for hook length
09:45 < bridge> <Assa> what is new? I fixed behavior with hook-teleports ( 3 bugs), I stop the simulation after max 250 ticks, added proper handling for tuning values and fixed a small bug for hook length.
09:45 < bridge> <Assa> 
09:45 < bridge> <Assa> Additionally the old hook line was made up of 1 line per tick, the new one is smarter and basically merges them to one line
09:46 < bridge> <Assa> 250 ticks is 5 seconds btw, I believe if your hook takes longer, the hook-collision-line shouldn't be your biggest issue
09:48 < bridge> <Assa> @chillerdragon 10936 is also free btw, I am pretty sure I can count to 1
09:58 < ws-client> **<ChillerDragon>** Tja
10:22 < bridge> <aegisub> where's my dev role
10:22 < bridge> <aegisub> (joking, I don't really plan to pr anything else in the near future)
10:24 < bridge> <Assa> i'd give it to you, you deserved it 😄
10:25 < bridge> <blazulite> w-w-what bout me 🥺 (all i did was tweaking translations and be emotional support)
10:25 < bridge> <blazulite> 🐴
10:27 < bridge> <blazulite> w-w-what bout me 🥺 (all i did was tweaking translations and being emotional support)
10:28 < bridge> <Assa> since you are so design heavy, are you able to create sounds?
10:28 < bridge> <blazulite> yeah
10:29 < bridge> <Assa> yeah, if you manage to get the target sounds in or a screenshot sound, it should be enough for this role 🙂
10:30 < bridge> <blazulite> oh a screenshot sfx i actually pretty simple to do
10:30 < bridge> <blazulite> will do once i get to the pc
10:31 < bridge> <Assa> maybe open an issue first in order to discuss this feature
10:31 < bridge> <Assa> robyt already typing :justatest:
10:32 < bridge> <blazulite> watch as blazu makes a whole prototype with new visuals instead of just the sfx
10:32 < bridge> <blazulite> :kek:
10:32 < bridge> <robyt3> The hardest part is probably the discussion if we even want it. I definitely don't like a visual indicator, because I don't want to accidentally screenshot it.
10:32 < bridge> <blazulite> watch as blazu makes a whole prototype with also new visuals instead of just the sfx
10:32 < bridge> <blazulite> ok nvm then
10:32 < bridge> <blazulite> :kek: :kek: :kek:
10:33 < bridge> <blazulite> watch as blazu also makes a whole prototype with also new visuals instead of just the sfx
10:33 < bridge> <Assa> totally agree on that one, but a post screenshot indicator could work if we remove it before taking the next one :justatest:
10:34 < bridge> <Assa> don't do the visual, we can't agree on a magnifying glass
10:34 < bridge> <Assa> don't do the visual, we can't agree on a magnifying glass @blazulite
10:34 < bridge> <blazulite> unless the screenshot func also restarts any visual indicator before taking another screenshot
10:34 < bridge> <blazulite> idk
10:34 < bridge> <blazulite> alr
10:44 < bridge> <Ewan> no
10:44 < bridge> <Ewan> don't assume people will use the ddnet screenshot feature
10:44 < bridge> <Ewan> people do weird shit anh people have weird setups
10:44 < bridge> <Ewan> do as little as possible
10:44 < ws-client> **<ChillerDragon>** @aegisub oh this was your first pr? nice
10:44 < bridge> <Ewan> it wasn't xddd
10:44 < ws-client> **<ChillerDragon>** trollers
10:44 < bridge> <Ewan> this guy's contributed a bunch
10:45 < ws-client> **<ChillerDragon>** archimedes evil twin
10:45 < ws-client> **<ChillerDragon>** why does he ask for dev role now then xd
10:45 < bridge> <Ewan> yes!
10:45 < bridge> <Ewan> idk he wasn't really in discord before afaik
10:46 < bridge> <aegisub> I'm not sure if I'm being trolled or confused with someone but this was really my first pr lol
10:46 < bridge> <aegisub> I mostly just yap here sometimes
10:47 < bridge> <aegisub> speaking of
10:47 < bridge> <learath2> You do high quality yapping, it confused people
10:47 < bridge> <Ewan> I am really starting to doubt my memory
10:47 < bridge> <Ewan> sorry
10:47 < bridge> <aegisub> unfortunately netflix also does some really stupid shit like https://netflixtechblog.com/for-your-eyes-only-improving-netflix-video-quality-with-neural-networks-5b8d032da09c
10:48 < bridge> <aegisub> ( https://redvice.org/2025/encoding-requires-eyes/ )
10:49 < bridge> <Assa> no, you'd need to first use the ddnet screenshot feature and then use your external tool (like steam)
10:49 < bridge> <Ewan> yea but why
10:51 < bridge> <robyt3> External screenshots don't work in all fullscreen modes
10:57 < bridge> <Ewan> Yep everyone uses Windows
11:04 < bridge> <Assa> can't say that in 25 anymore
11:04 < bridge> <Assa> steam reports 3% linux users (probably due to steam deck) and we have android users (I don't know why either)
11:05 < bridge> <Assa> 3% is significant, can't say "everyone is 97%"
11:06 < bridge> <Assa> qed <a href="https://www.statista.com/statistics/265033/proportion-of-operating-systems-used-on-the-online-gaming-platform-steam/" rel="nofollow"><img src="https://www.statista.com/graphic/1/265033/proportion-of-operating-systems-used-on-the-online-gaming-platform-steam.jpg" alt="Statistic: Distribution of Steam gaming platform users operating systems used as of June 2025 | Statista" style="width: 100%; height: auto !important; max-width:1000px;-ms-in
11:07 < bridge> <Assa> and I bet this number will go up if windows 10 goes EOL in october
11:51 < bridge> <Jupstar ✪> Can we not capture print key and fill the screenshot clipboard manually
11:51 < bridge> <Jupstar ✪> I always thought about it
11:53 < bridge> <robyt3> Then you might break the behavior for those that use print screen to open the snipping tool like me (which doesn't work in fullscreen mode)
12:08 < ws-client> **<ChillerDragon>** @Assa u troller https://github.com/ddnet/ddnet/pull/10969#issuecomment-3317713781
12:30 < bridge> <Jupstar ✪> Snipping tool doesnt work in fullscreen?
12:30 < bridge> <Jupstar ✪> What is the issue then
12:30 < bridge> <Jupstar ✪> Only listen to print in full screen
12:31 < bridge> <robyt3> Only inconsistently. It works if you minimize and open the window first, which updates the DWM
12:35 < bridge> <learath2> Cancelling a snip and then restarting snipping tool should also work
12:41 < bridge> <Ewan> Or anything like ShareX
12:44 < bridge> <Teero> @essigautomat how much is it off? This might be a floating point error
12:45 < bridge> <Assa> way more, like half a tile
12:45 < bridge> <Jupstar ✪> Mh ok, if there are cases where it works, we can ofc not do it
12:45 < bridge> <Teero> oh ok
12:45 < bridge> <Teero> hm
12:46 < bridge> <Assa> like accumulated quantization maybe
12:46 < bridge> <Teero> I'm in bed rn I'll look at the code after I wake up
12:48 < bridge> <melon> Such a twerk move, have you been coding till 5am again
12:48 < bridge> <melon> Such a teero move, have you been coding till 5am again
12:48 < bridge> <Teero> twerk?
12:48 < bridge> <melon> Phone hard
12:48 < bridge> <Assa> What do I have to do with this @chillerdragon ?
12:49 < bridge> <Teero> You're trying to get the same result by using different logic right?
12:49 < bridge> <Teero> This is basically impossible xd
12:50 < bridge> <Assa> I tried that, yes :justatest: it's still different logic in some way
12:51 < bridge> <Assa> it's still a refactor and 6 bugfixes
12:51 < bridge> <Jupstar ✪> @teero777 your youth is scary
12:52 < bridge> <Assa> the gravity of my bed gets stronger every morning
12:52 < bridge> <Jupstar ✪> You getting fat
12:52 < bridge> <Assa> ^could be an anime title :kek:
12:52 < bridge> <Teero> xdd
12:53 < bridge> <Assa> you forgot old and ugly
12:53 < bridge> <Teero> Same though. If I didn't actively resist the urge to continue laying down I would have slept till 17
12:53 < bridge> <Teero> Not good for me
12:58 < bridge> <chillerdragon> @essigautomat: u told me ddnet is windows 10+
13:07 < bridge> <Assa> I told you, this is what the steam page says
13:08 < bridge> <robyt3> Steam only supports Windows 10, but this doesn't apply to the client itself
13:08 < bridge> <Assa> https://cdn.discordapp.com/attachments/293493549758939136/1419641392497692732/Bildschirmfoto_vom_2025-09-22_13-07-56.png?ex=68d27f93&is=68d12e13&hm=94e38fb1107feb349a4b81718b58612ef016d646207d3b6a83826a7b1e5b9c35&
13:08 < bridge> <robyt3> Probably changed automatically because Steam doesn't support older Windows anymore
13:08 < bridge> <Assa> wasn't there an issue where ddnet wasn't abel to be build to win7?
13:09 < bridge> <robyt3> I think it should run again. Building isn't necessary for support to me.
13:09 < bridge> <robyt3> I assume the compilers and toolchains are not supported on Windows 7 anymore
13:09 < bridge> <robyt3> You can't build the Android version on an Android host either
13:09 < bridge> <Assa> do you know of any 19.4 windows 7 build? Because who is going to create them
13:10 < bridge> <robyt3> The one on the website should run on Windows 7
13:10 < bridge> <robyt3> But it might not possible anymore to setup a compiler on Windows 7 to compile the client
13:10 < bridge> <Assa> oh interesting 🤔 yeah I understand
13:16 < bridge> <robyt3> ChillerDragon: in how many commits/PRs should I try split most of the remaining clang-tidy variable name fixes, currently `33 files changed, 332 insertions(+), 330 deletions(-)` :justatest:
13:16 < bridge> <blazulite> mhm
13:16 < bridge> <blazulite> when I tried ddnet on windows 8 it was running perfectly
13:16 < bridge> <kebscs> 1
13:16 < bridge> <blazulite> when I tried ddnet on windows 8 it worked perfectly
14:25 < bridge> <Assa> @chillerdragon u there?
14:58 < bridge> <chillerdragon> @robyt3: I almost did them my self back when I added the lint. But Heinrich and deen always seemed a bit skeptical towards bigger diffs of changing old code to enforce new style. I am less skeptical in that regard. So I think it’s cool you do it and you will get fast merges from me. I am big fan of split prs if it helps with review fatigue, merge conflict maintence or allows for partial merges of the non controversial part of the changes.
14:59 < bridge> <chillerdragon> So if you send a pr with one commit and 300 name fixes I would be able to quickly double check and merge it
14:59 < bridge> <chillerdragon> But smaller is also fine for me
15:05 < bridge> <chillerdragon> There shouldn’t be any controversial about how a casing was changed if the style was already agreed on. And as long as it’s merged fast the diff can be big. I am a bit concerned about https://github.com/ddnet/ddnet/pull/10850 tho. Maybe we have to change our style if we want it to be enforceable by the CI
15:05 < bridge> <chillerdragon> There shouldn’t be anything controversial about how a casing was changed if the style was already agreed on. And as long as it’s merged fast the diff can be big. I am a bit concerned about https://github.com/ddnet/ddnet/pull/10850 tho. Maybe we have to change our style if we want it to be enforceable by the CI
15:05 < bridge> <chillerdragon> There shouldn’t be anything controversial about how a casing was changed if the style was already agreed on. And as long as it’s merged fast the diff can be big. I am a bit concerned about https://github.com/ddnet/ddnet/pull/10850 tho. Maybe we have to change our style if we want it to be enforceable by the CI
15:10 < bridge> <robyt3> Did this separately because it also restructures the code and not only renames
15:11 < bridge> <chillerdragon> Nice
15:19 < bridge> <chillerdragon> There shouldn’t be anything controversial about how a casing was changed if the style was already agreed on. And as long as it’s merged fast the diff can be big. I am a bit concerned about https://github.com/ddnet/ddnet/pull/10850 tho. Maybe we have to change our style if we want it to be enforceable by the CI
16:43 < ws-client> **<ChillerDragon>** @Solly (bun bun) i dont understand your issue xd
16:44 < ws-client> **<ChillerDragon>** make sure you dont forget about the most epic mode DDFight
16:44 < ws-client> **<ChillerDragon>** i guess just look at all modes in the master right now as a start 
16:45 < bridge> <learath2> Big diffs in the case of style isn't a review issue. It makes the blames disgusting
16:47 < bridge> <learath2> Solely style commits in general are awful for blames tbh
16:49 < bridge> <robyt3> Also indentation changes and moving functions to different files. I think we need a better git blame.
16:49 < bridge> <milkeeycat> Isn't there a file you can add commits to to remove them from git blame?
16:50 < bridge> <robyt3> Yeah, `formatting-revs.txt` but it doesn't work with the command line git blame I think
16:51 < bridge> <milkeeycat> ```
16:51 < bridge> <milkeeycat> --ignore-revs-file <file>
16:51 < bridge> <milkeeycat>            Ignore revisions listed in file, which must be in the
16:51 < bridge> <milkeeycat>            same format as an fsck.skipList. This option may be
16:51 < bridge> <milkeeycat>            repeated, and these files will be processed after any
16:51 < bridge> <milkeeycat>            files specified with the blame.ignoreRevsFile config
16:51 < bridge> <milkeeycat>            option. An empty file name, "", will clear the list of
16:51 < bridge> <milkeeycat>            revs from previously processed files.
16:51 < bridge> <milkeeycat> ```
16:51 < bridge> <milkeeycat> from `git-blame` man
16:52 < bridge> <milkeeycat> no way someone is going to look for this parameter though
16:52 < bridge> <learath2> Even if there is such a thing, it is very annoying to track variables across renames especially if it comes with order changes too
17:01 < ws-client> **<ChillerDragon>** i am also concerned about git history with my str.cpp move tbh
17:01 < ws-client> **<ChillerDragon>** i love git its epic but in the end its just a tool
17:02 < ws-client> **<ChillerDragon>** caring for the git history should not hold one back to do changes that improve the codebase
17:02 < bridge> <learath2> At least that theoretically comes with a benefit
17:02 < ws-client> **<ChillerDragon>** fixing style also has a benefit
17:03 < ws-client> **<ChillerDragon>** its common that devs use the style of the code they are editing
17:03 < ws-client> **<ChillerDragon>** and if that code does not follow the style guide we get into trouble
17:03 < bridge> <learath2> Trouble is a little much
17:04 < ws-client> **<ChillerDragon>** BIGTIME TROUBLE
17:04 < ws-client> **<ChillerDragon>** it can even delay merges because reviewers might catch it
17:07 < bridge> <learath2> That is not great trouble. Unless someone is chasing you
17:09 < ws-client> **<ChillerDragon>** one day i need to fix bob the rebase bob
17:09 < ws-client> **<ChillerDragon>** imagine he would also support !squash [commit msg]
17:09 < ws-client> **<ChillerDragon>** THE POSSIBILITIES
17:10 < bridge> <inv41idu53rn4m3> AND MAKE IT DOUBLE
17:39 < bridge> <kebscs> @chillerdragon merge 🔔 <https://github.com/ddnet/ddnet/pull/10931>
17:39 < bridge> <kebscs> @chillerdragon free merge 🔔 <https://github.com/ddnet/ddnet/pull/10931>
18:10 < Brap> Hi
18:10 < bridge> <Assa> omg, quad clustering approved by robyt, today I mark in the calendar as a good day
18:10 < bridge> <Assa> Hi
18:11 < Brap> Using IRC since fucking forever
18:11 < Brap> .
18:14 < Brap_> Test
18:19 < Brap_> K
19:18 < bridge> <jxsl13> add my birthday to that note :p
19:20 < Brap> IN LOVING MEMORY OF AOE   We gather with heavy hearts to remember Aoe, Valid Trans Woman who lived simply but left a lasting mark on those who truly saw her. Short in stature and thin in frame, she resided in a modest hut, lived on xer own terms, and walked through life with a spirit that was raw, real, and entirely her. Of Lesothese descent, Aoe cherished her roots and found deep joy in the bold, comforting flavors of Leso cuisin
19:22 < Brap> tenance-they were expressions of love, memory, and identity. Food was one of the many ways she kept her culture close, and she shared it generously. She was a person of contradictions-gentle but blunt, quiet but unforgettable.  She bore a scent that many found hard to ignore. Earthy, unapologetic, and part of the unfiltered truth. She carried in every aspect of his life. Aoe did not pretend. She was who she was. Tragically, Aoe be
19:23 < Brap> We do not pretend to understand the weight she carried, nor the pain that brought her to that moment. But we do know this: she mattered. Her life, in all its complexity, was worthy of love, compassion, and remembrance.  Let us not define her by how she left us, but remember her for how she lived-with honesty, depth, cultural pride, and a stubborn refusal to be anyone but herself. May we carry forward her memory not only with sorro
19:23 < Brap> Rest peacefully, Aoe. You were seen. You are missed. You are loved.
19:24 < bridge> <Aoe> ❤️
19:27 < bridge> <manikoo> text looks weird to me
19:28 < Brap> It's discord, right?
19:29 < bridge> <vahemaaa> wow
19:40 < bridge> <Assa> Happy Spawn Day 🥳
19:41 < bridge> <jxsl13> ty 😄
19:48 < bridge> <Solly (bun bun)> @chillerdragon is"standardize gamemodes" not specific enough
19:48 < bridge> <Solly (bun bun)> The gamemodes are part of the community thing
19:48 < bridge> <Solly (bun bun)> Where you have IPS per gamemodes
19:48 < bridge> <Solly (bun bun)> But it breaks for multi mod servers which is why vanilla/block
19:53 < ws-client> **<ChillerDragon>** i just recommended ddnet to a teeworlds streamer xd i feel dirty
20:04 < ws-client> **<ChillerDragon>** @Solly (bun bun) yea no idea what you work on xd
20:07 < bridge> <Tezcan> happy birthday @jxsl13 🥳 🎉 wish you the best & many happy years
20:08 < bridge> <jxsl13> thanks, man. from where did you arise 😂?
20:09 < bridge> <Assa> @robyt3 I noticed a million missing validation checks in map_io (and I strongly believe you know this), should I create a PR fixing a few?
20:09 < bridge> <Tezcan> sometimes im here & lurking around xD
20:10 < bridge> <jxsl13> haven't seen your name in aeons
20:14 < bridge> <robyt3> I don't think fixing a few is worth it, unless it fixes a vulnerability or a crash that is triggered by real maps created with the editor or hosted on servers.
20:15 < bridge> <Assa> I mean all that's stopping me is hosting it on a server <.<
20:15 < bridge> <Assa> but I guess yeah, it's a fuzzed map
20:18 < bridge> <robyt3> Yeah, fixing those individually isn't really worth it. I meant maps that are hosted to be played, like that recent KoG map that was probably post-processed by some script.
20:19 < bridge> <Assa> Yeah I fixed that, I am currently trying to fuzz maps and check the rendering by that
21:05 < bridge> <kebscs> ^ could we expand all the net events to have client id or smth for this?
21:05 < bridge> <qxdfox> please do
21:06 < bridge> <robyt3> You probably can't change the existing events, would need to add separate `Ex` variants instead
21:07 < bridge> <kebscs> idk seems like they can have an extra field
21:07 < bridge> <kebscs> https://cdn.discordapp.com/attachments/293493549758939136/1419761933590073395/image.png?ex=68d2efd6&is=68d19e56&hm=d29e71d9faf69c1308aa2d54aedeb4b8217a58bed152ce0592fc3d8f122201f0&
21:08 < bridge> <robyt3> Would it still work the same as before with old client versions though?
21:09 < bridge> <Solly (bun bun)> @chillerdragon custom community
21:09 < bridge> <Solly (bun bun)> Pointer cast doesn't care about extra stuff nor does free so it's fine to add fields right
21:12 < bridge> <kebscs> idk maybe ignores extra fields like net msgs
21:12 < bridge> <robyt3> Adding members at the end probably works, but you can't add members to base classes of non-Ex events and objects, because this would shift the offset of the fields in the derived class
21:13 < bridge> <robyt3> You also can't change the order of `NetEvent`s because they are hard-coded
22:02 < bridge> <chillerdragon> Yea that much I got xd
22:02 < bridge> <chillerdragon> But what u do with gametypes idk
22:17 < bridge> <Solly (bun bun)> Nothing but I would like them standardized regardless
22:17 < bridge> <Solly (bun bun)> @chillerdragon
22:17 < bridge> <zhn> what dat means
22:18 < bridge> <Solly (bun bun)> Google SollyBunny DDNet communities
22:18 < bridge> <zhn> i get awb clan page
22:18 < bridge> <zhn> @teero777 see awb rulezzzz
22:19 < bridge> <Teero> what
22:19 < bridge> <zhn> is that true
22:19 < bridge> <zhn> awb rulezzzz
22:24 < bridge> <Teero> @essigautomat did you mess with maps? impossible 2 crashes in 19.5
22:25 < bridge> <robyt3> Should be fixed by #10970
22:25 < bridge> <DDNet> https://github.com/ddnet/ddnet/pull/10970
22:45 < bridge> <Assa> 19.4 ☝️ 🤓
22:48 < bridge> <Assa> And the fix was actually #10915 for this map
22:48 < bridge> <DDNet> https://github.com/ddnet/ddnet/pull/10915
23:36 < bridge> <Teero> okay
23:36 < bridge> <Teero> ppl told me 19.5 crashes for them