klange changed the topic of #osdev to: Operating System Development || Don't ask to ask---just ask! || For 3+ LoC, use a pastebin (for example https://gist.github.com/) || Stats + Old logs: http://osdev-logs.qzx.com New Logs: https://libera.irclog.whitequark.org/osdev || Visit https://wiki.osdev.org and https://forum.osdev.org || Books: https://wiki.osdev.org/Books
[_] has joined #osdev
[itchyjunk] has quit [Ping timeout: 260 seconds]
tixlegeek has quit [Quit: tixlegeek]
Burgundy has quit [Ping timeout: 240 seconds]
kof123 has joined #osdev
kof123 has quit [Quit: Leaving.]
kof1231 has joined #osdev
goliath has quit [Quit: SIGSEGV]
pretty_dumm_guy has joined #osdev
hrberg has quit [Remote host closed the connection]
hrberg has joined #osdev
pretty_dumm_guy has quit [Quit: WeeChat 3.5]
heat has quit [Read error: Connection reset by peer]
heat has joined #osdev
Matt|home has joined #osdev
netbsduser`` has quit [Ping timeout: 258 seconds]
[Kalisto] has joined #osdev
gog has quit [Ping timeout: 246 seconds]
dza has quit [Remote host closed the connection]
dza has joined #osdev
gbowne1 has quit [Read error: Connection reset by peer]
netbsduser`` has joined #osdev
netbsduser`` has quit [Ping timeout: 258 seconds]
heat has quit [Ping timeout: 260 seconds]
netbsduser`` has joined #osdev
netbsduser`` has quit [Ping timeout: 245 seconds]
xenos1984 has quit [Quit: Leaving.]
netbsduser`` has joined #osdev
netbsduser`` has quit [Ping timeout: 255 seconds]
pg12 has quit [Remote host closed the connection]
pg12 has joined #osdev
masoudd has joined #osdev
duderonomy has joined #osdev
masoudd has quit [Ping timeout: 255 seconds]
duderonomy_ has joined #osdev
duderonomy has quit [Ping timeout: 264 seconds]
GeDaMo has joined #osdev
[_] has quit [Read error: Connection reset by peer]
Left_Turn has joined #osdev
Matt|home has quit [Quit: Leaving]
masoudd has joined #osdev
Starfoxxes has quit [Ping timeout: 258 seconds]
Starfoxxes has joined #osdev
masoudd has quit [Quit: Leaving]
flom84 has joined #osdev
gog has joined #osdev
GeDaMo has quit [Ping timeout: 240 seconds]
GeDaMo has joined #osdev
pretty_dumm_guy has joined #osdev
tixlegeek has joined #osdev
GeDaMo has quit [Read error: Connection reset by peer]
tixlegeek has quit [Quit: tixlegeek]
Burgundy has joined #osdev
flom84 has quit [Ping timeout: 240 seconds]
heat has joined #osdev
[itchyjunk] has joined #osdev
GeDaMo has joined #osdev
netbsduser`` has joined #osdev
phoooo has joined #osdev
phoooo has quit [Client Quit]
vdamewood has joined #osdev
<Ermine> Time to install openbad
<gog> openbad
<zid> ovenbsd
<gog> in icelandic BSD would be BHBD
<gog> HBDB*
<zid> does bsd mean something then
<gog> hugbúnaðardreifings Berkeley
<zid> berkley.. so mething?
<gog> berkeley software distribution i thought
<zid> ah, in english that'd be BSD
* vdamewood gives gog a Berkeley Fish Distribution
* gog chomps fishe
<zid> I'm a BSD, Bastard Stack Developer
<vdamewood> Interesting thing about BSD, the first BSD release was just a collection of userland tools for AT&T Unix.
<zid> makes sense
<vdamewood> I don't know when BSD first included a kernel patch.
<Ermine> hence 'software distribution'?
<vdamewood> Ermine: Pretty much, yep.
<netbsduser``> what's charming about bsd is that they hack away on a piece of proprietary software for over a decade, tremendously improving it, they release the resultant ship of theseus as open source, USL gets upset, but they effectively counter-sue them and come to a settlement that both sides make some changes and their software is regarded as golden
<vdamewood> I have the BSD source code on CD, but I'm too lazy to find my Optical drive, plug it in, and hunt through the sources for the first BSD kernel.
<GeDaMo> It might be on archive.org
<nortti> 3bsd had kernel modifications at least, since it included paging virtual memory support. dunno if 2bsd releases before 3bsd had made any
<vdamewood> Though, the Ship of Theseus wasn't completely replaced at Berkely, 4.4BSD came in two flavours, one still had AT&T source code in it, and the other wasn't a complete system.
<vdamewood> I have no idea why I spelled flavor with a U there. I'm a Usian.
<zid> That there's no exonym for usaian boils my piss
<vdamewood> We don't really have an endonym either.
<zid> merican, said in english
<nikolar> Is usonian better
<zid> but you can't use it as an exonym because it also includes canada and all of south america, technically
<vdamewood> Esperanto uses Usonano.
<vdamewood> In Spanish (at least in Mexico), they use Estadounidense.
<zid> hehe, dense
<nikolar> Wasn't there a movement at some point in us to change the endonym to usonian or something similar
<vdamewood> nikolar: I think it was a somewhat fringe movement.
<zid> I'm fine with yank
<nikolar> Yeah that one
<netbsduser``> vdamewood: what i like about it especially is that the result of the rewriting is still in clear continuity with the historic unix
<vdamewood> zid: Alas, within the US, 'yank[ee]' refers to a specific regeion.
<zid> that's fine
<zid> it's an exonym
<zid> they're always weird
<zid> like, japanese
<netbsduser``> you can find struct user, struct proc, etc and feel quite comfortable
<zid> because of a game of telephone through the portuguese
<vdamewood> So, like calling all of the Netherlands 'Holland'.
<zid> that's how "nihongojin" ends up
<zid> err not go
<vdamewood> Yeah, -go is language.
<zid> they're synonyms in english damnit
<zid> I brained the wrong one
<vdamewood> -jin is person.
<nikolar> netbsduser``: yeah bsd history is quite interesting
<zid> yes
<zid> japanese language person :P
<zid> rather than japanese person
<vdamewood> And really, Nihon is a derived form itself, since the word comes from Chinese.
<zid> english doesn't discriminate
<zid> Wá is the correct name for japan, obvs
<netbsduser``> it is basically a rejoinder and counterblast to people who have exaggerated beliefs on what copyright protects
<vdamewood> Yeah, but they only use 'wa' in compounds thesedays.
<zid> 倭
<zid> The submissive dwarf country :p
<gog> zid: bandaríkingar
<vdamewood> Yeah, but they don't use that character these days, either.
<zid> is that yank in iceland
<zid> (I've decided that all nationalities + languages should be the same word, like chinese and chinese, japanese and japanese)
<zid> iceland is now the name of the language.
<gog> böndur - union, rík - state, ing - originating from -ar plural
<zid> unistatian
<vdamewood> gog: Looks like German if you squint.
<vdamewood> Bundes, reich, ing, er.
<gog> germanic languages do like their compound words
<zid> most pie languages that aren't strongly romantic do
<zid> to be fair
<zid> latin, fucking up languages since a long ass time ago.
<vdamewood> I'm trying ot think of Persian compounds, but most are in the form X-e Y.
<vdamewood> Where that's the normal construction for Y of X.
<zid> synthetic languages for life
* vdamewood hands zid some Inuktitut.
* zid hands vdamewood a rock with some lines carved into its corner
<vdamewood> zid: Ogham?
<zid> ya
<vdamewood> "Seamus was here"
<zid> irish orthography still a mess
<vdamewood> It's a fascinating mess!
<zid> 'qurickrie', pronounced, 'duck'
<vdamewood> It's like someone took two confusing English rules, and applied them everywhere!
<zid> iaghrichet -> irish
<zid> hey it was the irish who decided on our bloody shared orthography in the first place
<vdamewood> That thing about g and c being pronounced differently when followed by e or i, and the -h digraphs.
<zid> I should learn to read welsh
<vdamewood> zid: I hear it has the same effect as reading the Necronomicon.
<nortti> welsh seems like a prettz simple orthographz
<zid> yea welsh is nice and simple
<zid> just a couple of weird y and ll things
<zid> other than that it's fairly usual
<nortti> note to self: switch out of hungarian keyboard when you're done with it
<zid> or switch to a compromise, qwertx
<vdamewood> I like macOS's ABC - Extended layout.
<zid> abcdef layout?
<vdamewood> No, it's QERTY, but I have all the diacritics with just a couple of key strokes.
<vdamewood> QWERTY*
<nortti> us qwerty?
<zid> oh that's a bad name then
<zid> I hate windows' layouts
<vdamewood> Ugh, why is that so hard to type? I just have to slide my finger across the row.
<netbsduser``> irish orthography became too simplified
<gog> qwerty
<zid> I have altgr but it only lets me type éúíóᦀ
<zid> all other keys do nothing
<gog> @ſ€¶
<netbsduser``> it's sometimes hard to match it up with the common words in scottish gaelic
<gog> æß„đ
<gog> huh i hae two ways to make æ
<gog> altgr-a and æ
<zid> wtf, altgr also lets me type \
<vdamewood> zid: In the case of mine, Alt (either of them), plus many letters gives me a dead key of a diacritic, then I just type the base letter.
<zid> but.. that's just a regular key
<gog> µ
<zid> maybe it's a broken port of the composite key
<vdamewood> zid: For example, é is Option-E e (option what Alt is called in the mac world)
<zid> that should absolutely not do that, to be consistent with all other keys
<zid> vdamewood: That's called a compose key
<zid> it mimics using a typewriter
<zid> to not advance the carriage
<vdamewood> I've heard it called a 'dead key' for the same reason.
<zid> yea
<vdamewood> It's 'dead' in that it doesn't advance the carriage.
<zid> the / thing is super weird..
<zid> why do I have two / keys
<vdamewood> / thing?
<gog> which keys make / for you
<zid> altgr + literally any key = does absolutely nothing
<gog> mine is shift 7
<zid> altgr + # = \
<vdamewood> / is a dedicated key on my layout.
<zid> I have a dedicated \ key
<gog> £
<gog> that's what altgr + # does for me
<gog> gross
<nortti> zid> I hate windows' layouts < zid> I have altgr but it only lets me type éúíóᦀ ← finnish keyboard layout used to be the same as swedish back in the early days of computing, but in 2006 (designed) to 2008 (ratified as a standard) we finally got a new keyboard layout that included, y'know, all the letters required for writing finnish (you could not type š and ž on the old swedish-based
<zid> that's shift 3 for me, which I think is what # is on enus
<nortti> keyboard layout). the new layout is a superset of the old one, so there's really no reason not to switch to it, and all the linux distributions et al did so by early 2010s. microsoft, on the other hand, still ships the old keyboard layout and the most official place to get the modern one is still someone's personal homepage https://jkorpela.fi/sfs5966.html
<bslsk05> ​jkorpela.fi: Multilingual Finnish keyboard (SFS 5966)
<vdamewood> ð́
* gog screams in icelandic
<zid> nortti: yea windows is dumb as fuck wrt keyboard layouts
<vdamewood> ^ I cal also do that
<zid> I keep meaning to make a decent layout, a bit like the linux one
<zid> but it's a pain in the dick
<vdamewood> gog: þðððððððððð?
<zid> put yen on altgr-y etc
<vdamewood> ¥ :P
<vdamewood> Option-Y for me.
<zid> yea, it is on linux
<zid> just.. not windows
<zid> because windows refuses to be useful
<vdamewood> zid: Which layout do you have?
<zid> uk
<gog> ¥
<gog> algtr-Y
<vdamewood> Microsoft: "No one in the UK will ever use east-Asian money"
<zid> `1234567890-= ¬!"£$%^&*()_+
<gog> altgr-y is ←
<zid> ¦€
<zid> neutral, shift, altgr of top row
<zid> if you've never seen uk layout
<zid> Have you seen the portugoose layout? it's mental
<vdamewood> My physical keyboard is a US layout.
<zid> ewww ansi
<zid> I mean
<zid> Is it physically american
<zid> or american on a physically ISO
<vdamewood> Well, yeah. I have the one-row Return key and not that turned-L one.
<zid> ansi, bleh
<nortti> zid: you should check out the hungarian keyboard layout: https://en.wikipedia.org/wiki/File:HunWinKB.png
<zid> I can't type on ansi
<zid> nortti: yea I'd rather die than try to program on that
<nortti> I love how it puts ö on the usual spot for 0 and then moves 0 all the way to the left
<zid> main difference between us and uk layout is " and @ are swapped, and # is on its own key, and shift-3 is £
<vdamewood> zid: Yeah, in the US, Shift-3 is instead the pound sign.
<zid> windows doesn't let you remove enUS layout at all so occasionally I will accidentally switch into it, often enough that I know the US layout differences too
<nortti> also ~ is shift-# and not shift-` going by the labers on my computer
<nortti> (I have a UK physical layout on this for reasons™)
<zid> uk layout has the old logical not and or symbols for some reason
<zid> on `
<zid> ¬A ¦ ¬B is a thing I can type, for some reason
<zid> Can't type french though
<zid> may or may not be intentional
<mcrod`> hi
<vdamewood> zid: What are you missing for French?
<zid> two sets of accents and the c
* vdamewood hands zid a ç
<zid> I can *only* type accute accents, no graves
<zid> and no hats
<zid> acute
<vdamewood> è & ô?
<zid> and definitely no umlauts
<zid> or sharp S
<vdamewood> ü ß!
<zid> and my keyboard has never even heard of a slashed o
<vdamewood> ß̈
<zid> much less a thorn or eth
<vdamewood> ð̧!
<vdamewood> I wonder if these are rendering properly on others' systems.
<zid> yes
<vdamewood> I'm typing htem all.
<zid> the world settled on utf-8 a long time ago, anyone using shift-jis or ucs2 can go suck a dog
<vdamewood> Well, some of these characters aren't single code-points in Unicode, and use combining diacritical marks, which some fonts or systems have trouple rendering.
<vdamewood> Like the Eth with Cedilla for example.
<zid> they're actually all apparently just part of fixedsys, even
<zid> the font substition hasn't kicked in for me
<zid> if you type something REALLY weird, windows will pull it out of a different font
<zid> that does have that glyph
<vdamewood> Fixedsys? Really? It must have been updated since 1995.
<nortti> alternatively it's able to compose diacritics and the base into a glyph
<zid> you just typed really normal ass chararacters is all
<zid> fixedsys supports most european language characters, and squared and cubed, but that's it
<vdamewood> zid: Got a screen shot of that Eth with cedilla?
<zid> it only spans 0-255, but it has all the eurolang characters packed into the top half
<zid> oh maybe it didn't render right then? :P
<vdamewood> zid: Yeah, I typed two odd characters, a sharp S with umlaut, and an Eth with cedilla.
<zid> ð̧! ah I see it if I copy paste it
<zid> it's invisible when rendered though
<zid> if I ctrl-c that line, I get eth, black box, exclam
<nortti> what terminal are you using?
<zid> when it renders, it goes back to eth exclam
<nortti> slash client
<zid> "EDIT"
<zid> CreateWindowEx("EDIT", ...
<vdamewood> zid: That black box is a combining diacritic.
<zid> yea you already said it's a cedilla
<zid> I'd rather it showed the black box to me *all* of the time
<vdamewood> Like I said, some fonts don't do combining diacritics all that well.
<zid> for some reason the display doesn't show it, but the input box does
<zid> not sure why there's a difference
<zid> I can at least tell you sent something I couldn't decode if it shows the box
<vdamewood> It takes a little bit of scripting in the font itself to support combining diacritics.
<zid> fixedsys is a bitmap font, not truetype
<zid> so it was never going to support it
<zid> I'm just mad that it silently failed
<vdamewood> That's too bad. I rather like the look of Fixedsys. It reminds me of programming and web development in 1996.
<zid> It's still there, because I copy pasted *your* message to get it, it's just invisible
<vdamewood> hunter2?
heat_ has joined #osdev
heat has quit [Ping timeout: 240 seconds]
craigo has joined #osdev
craigo has quit [Read error: Connection reset by peer]
heat_ has quit [Remote host closed the connection]
heat_ has joined #osdev
Burgundy has quit [Ping timeout: 260 seconds]
termin has joined #osdev
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 260 seconds]
antranigv has quit [Quit: ZNC 1.8.2 - https://znc.in]
kof1231 has quit [Quit: Leaving.]
antranigv has joined #osdev
kof123 has joined #osdev
rustyy has quit [Quit: leaving]
netbsduser`` has quit [Remote host closed the connection]
netbsduser`` has joined #osdev
heat_ is now known as heat
<heat> kern_return_t mach_vm_remap(vm_map_t target_task, mach_vm_address_t *target_address, mach_vm_size_t size, mach_vm_offset_t mask, int flags, vm_map_t src_task, mach_vm_address_t src_address, boolean_t copy, vm_prot_t *cur_protection, vm_prot_t *max_protection, vm_inherit_t inheritance);
<heat> how many params do yall want
<definitely-mjg> that's a page from a bsd book
<heat> the bsd book would have obfuscated names
<heat> int vmremap(struct map *tmap, void *taddr, ulong_t sz, ulong_t mask, int flags, struct map *stask, void *saddr, int copy, int cprot, int mprot, int inherit)
<heat> wait, u_long? or u_long_t
<heat> something
rustyy has joined #osdev
definitely-mjg is now known as vaxuser
netbsduser`` has quit [Ping timeout: 272 seconds]
<kof123> /* XXX: what in tarnations? */
Turn_Left has quit [Ping timeout: 240 seconds]
<heat> or in ufs code /* XXX ass shit fuck ass */
<kof123> it was meant as "yall" style. i'm not sure what the "yall" style for the latter is, but it exists
<vaxuser> heat: i find the comment to plausibly exist
<vaxuser> does it?
netbsduser`` has joined #osdev
<heat> 522 * XXX I hate the world.
<heat> ufs_rename.c has 30 XXX's
<heat> is this the sign of really rock solid code?
<vaxuser> that's netbsd?
<vaxuser> the infamous file
<vaxuser> #if 0 /* XXX */ goto whymustithurtsomuch;
<vaxuser> #endif goto arghmybrainhurts;
<vaxuser> hateyou: return error;
<vaxuser> erm
<vaxuser> ihateyou:
<vaxuser> return error;
<vaxuser> actual fucking code
<heat> 546 arghmybrainhurts:
<heat> 547 UFS_WAPBL_END(mp);
<kof123> the freebsd channel on freenode used to have a vin diesel bot. so later generations have their own interpretation
heat has quit [Remote host closed the connection]
<vaxuser> i don't know if that channel was "official"
<vaxuser> there is at least one which was name-squatted
<vaxuser> for example, this channel is named osdev, but all we do here is pet gogs
heat has joined #osdev
zxrom has quit [Quit: Leaving]
<vaxuser> heat: ey mofo, check this out
<vaxuser> heat: flamegraphs from sampling on instructions which also print the lock the thread is contending on
<vaxuser> ok, don't got me wrong, but you know what would be the shit
<vaxuser> flamegraphs in ascii art
Arthuria has joined #osdev
Turn_Left has joined #osdev
<zid> heat can you teach musl about tabs
<bslsk05> ​git.musl-libc.org: exp_data.c\math\src - musl - musl - an implementation of the standard library for Linux-based systems
<puck> zid: Arm, you mean :p
<puck> https://github.com/ARM-software/optimized-routines has the same file but with even more macros
<bslsk05> ​ARM-software/optimized-routines - Optimized implementations of various library functions for ARM architecture processors (82 forks/483 stargazers/NOASSERTION)
<gog> very professional coding
<zid> they stole it from arm yes
<zid> but why did they not format it
<zid> I'd have rejected that patch due to "what the fuck is this shit? I can't maintain this, can you? if you can, write a fucking commit message explaining it"
vdamewood has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<puck> zid: because it means dealing with annoying merge conflicts when you want to update it
<puck> i suspect exp_data.c was also Generated in some way
<zid> possibly yea
<puck> also let's be fair
<puck> indentation would not have saved this file
<zid> no but it's just very.. musl
<puck> i'd probably not see the use in indenting "0xbc905e7a108766d1, 0x3fefe315e86e7f85," either
<heat> musl libm isn't an awful read because a lot of it was lifted from BSD/Sun
CaCode has joined #osdev
termin has quit [Read error: Connection reset by peer]
<kof123> somehow i was reading that the other day https://libre-soc.org/HDL_workflow/#index10h1 > do not commit autogenerated output. write a shell script and commit that, or add a Makefile to run the command that generates the output, but do not add the actual output of any command to the repository. ever. this is really important. even if it is a human-readable file rather than a binary object file.
<bslsk05> ​libre-soc.org: HDL workflow
<heat> that's silly, now you're adding extra tooling to your build for meaningless static data
<kof123> well, i think that project wants that, i.e. they want to keep it 'bootstrappable' was not really arguing either way, just hadn't seen an explicit rule before either way perhaps
<kof123> i tend to do opposite depending on goal. if the goal is NIH ...then i probably don't want any "external tools" unless i wrote them already. 'bootstrappable' is relative "from what?"
<kof123> s/NIH/dogfooding/
<sortiecat> Somebody highlighted me in here but it scrolled off my scrollback, I assume it was just heat praising me :)
<heat> praise be sortiecat
<heat> <heat> if you want a cloak just ask sortie
<heat> <heat> i rep sortix colors in linux channels
<heat> maybe this?
<sortiecat> praise the cat
* heat pets sortiecat
<zid> I said '... coprocessor tie cats to different resources'
<Ermine> heat: are you doing pets now?
<zid> maybe that was it
<heat> Ermine, no, one time only
<heat> this is *not* becoming a thing. i also won't hi
Burgundy has joined #osdev
<gog> hi
antranigv has quit [Quit: ZNC 1.8.2 - https://znc.in]
antranigv has joined #osdev
zxrom has joined #osdev
<mcrod`> hi
antranigv has quit [Quit: ZNC 1.8.2 - https://znc.in]
antranigv has joined #osdev
kof123 has quit [Ping timeout: 252 seconds]
antranigv has quit [Quit: ZNC 1.8.2 - https://znc.in]
<mcrod`> heat: hi
antranigv has joined #osdev
goliath has joined #osdev
kfv has joined #osdev
kfv has quit [Quit: Textual IRC Client: www.textualapp.com]
* geist yawns
<zid> quick, geist is tired and not paying much attention, steal everything not nailed down
<vaxuser> vax is nailed down tho
<vaxuser> the fuck i need damaged 3.5 floppies for
<gog> the vax isn't nailed down it's just too bulky to steal
<gog> unless it's one of those microvax jobbies
<zid> I was about to say "aww, all I got was a stupid vax"
<geist> gezus you're still going on about that?
<geist> omg developer time is 100% fungible!!!! why they work on vax!
<geist> someone nedes a project manager!
<zid> who going on about what
<geist> everyone should use scrum
<vaxuser> i'm just rolling with standup improv
<vaxuser> ye i justo pened an itanium ticket on jira
<vaxuser> completion: 1%
<heat> itanium: remove
<heat> sadness.jpeg
<geist> heat: you can still port your os
<geist> do it!
<heat> the ia64-elf toolchain i have installed is kinda borked
<heat> i need to try out ia64-linux-gnu
<zid> oh did they pull the trigger
<zid> through mm tree? okay then
<zid> or not, mm just having to deal with breakage?
<zid> I'm too lazy to look for actual merges
<geist> i think we were just talking about the impending removal yeah
<zid> I saw the original discussion thread and me and heat talked about it
<zid> but it didn't actually get any definitive conclusion
<zid> and now it's apparently gone, but I can't find that email offhand, I assume it came in through some subsystem
<zid> rather than being dropped on main lkml
<zid> so I'd have to find the linux-next merge for it and meh
<vaxuser> removing stuff is hard
<heat> it's in asm-generic
<heat> .git
<zid> it is? okay
<zid> I saw the merge conflicts with it and mm tree
<zid> but couldn't figure out which of the two had merged it
<bslsk05> ​lore.kernel.org: [GIT PULL] Remove Itanium support - Ard Biesheuvel
<vaxuser> there is always someone claiming they use it
<zid> which provides a fix for the ia64 build (which
<zid> is currently broken), and subsequently removes it
<zid> wow, harsh
<zid> It's like the ending to 1984
<heat> 65kloc of great amazing work down the drain
<zid> I bet there's sitll like anotehr few thousand lines of simplifications that can happen in -mm now too
<heat> nah
<zid> ia64 had lots of edge cases that maybe people don't realize existed only because of ia64
<heat> all architectures have em
<heat> x86 has some weird inverted-PTE shit they're doing, for some reason
<zid> all architectures have edge cases that people don't realize exist because of itanium? wat?
<heat> for speculation I think
<zid> I'm saying there are features not *named* _IA64_ in there, that exist for IA64 to use
<heat> all architectures have weird mm edgecases
<zid> that have probably slipped through this big removal patchset
<heat> IA64 was probably fairly clean since they do not use radix-like page tables at all
<zid> nah ia64 had a bunch of shit people were always complaining about
<zid> a bunch of which came up in that thread before
<zid> "yea I'd like to see it gone, most of this file exists just because ia64 does something different" etc
<heat> i had the impression most of the problem was just patch overhead (patch for one arch, patch for all)
<heat> but maybe im wrong
<zid> Your kernel is as weird as your weirdest architecture
<zid> and ia64 had that crown in a lot of files
<heat> surely the weirdest is alpha
<heat> RCU is full of "this does not need memory barriers on every arch except alpha"
<vaxuser> alpha the greateterest
<zid> yea alpha is the new king
<netbsduser``> alpha or itanium seem like fun arches but they are not on my agenda
<netbsduser``> i am happy with my 68k port
<zid> what about xeon phi
<heat> vaxuser, sun really missed an opportunity by not naming the dtrace type format as DTF
<vaxuser> now that you mention it, how is htis not named ztrace
<heat> particularly given the popular woman-kisser bcantrill
<vaxuser> the last word in tracing innit
<netbsduser``> that took me a moment
<netbsduser``> but it is indeed the difference between linux virgins and solaris chads
<geist> yah alpha i missed. the only thing i never fully grokked was the whole PALcode thing. specifically which palcode you load for what, and precisely which version offloads what
<geist> thing though is alpha is somewhat special to linux
<geist> it's the second architecture it was ever ported to. linux 2.0
<geist> so it has probalby a fairly wide influence on core design
<geist> timeline lines up too, alpha was a Super Big Deal in the mid 1990s
<zid> me too
<heat> they lost their portability cherry to alpha
<heat> popped
<heat> whatever, the joke still lands
<geist> yep
<heat> i think i've read before that linus really does not want to ever delete i386 due to it being the first arch
<geist> gosh i remember seeing an alpha 200mhz or so running NT 3.51 about 1994. you could tell it was a unit of a cpu
<geist> compared to a P60 or P90 it was clearly kicking ass
GeDaMo has quit [Quit: That's it, you people have stood in my way long enough! I'm going to clown college!]
<geist> btw watched pretty uch the entire Dave Cutler interview last night, was pretty great
<geist> and dang he's still got it, even at 81
<geist> he has a heck of a memory, still dropping names for people he met 60 years ago
<heat> P90 rush B
<heat> does he talk about UNIX?
<sortiecat> I wish I could afford a real Unix
sortiecat is now known as sortie
<geist> not really. the general topic of the interviewer was mostly about windows and NT, didn't know much about DEC or other stuff
<sortie> sortiecat, psst need unix
<geist> so cutler mentioned xenix and whatnot, but mostly in the context of what MSFT was doing at the time
<heat> i hope they asked him about the horrendous coding style he cursed us with
<geist> hmm, what do yo umean?
<mcrod`> hm
<heat> actually that's not even the worst bit
<mcrod`> 64GB of RAM here I come
<heat> the whole coding style is very whitespacey, screenspace-wasteful
<heat> ULONG Var = 0; is afaik not allowed, for example
<heat> but rather ULONG Var; /* more decls ... */ Var = 0;
<heat> oh and the fucking // comments that need "warning" lines
<heat> //
<heat> // This is a comment
<heat> //
<geist> oh. well that doesn't seem too bad
<geist> i thought you were gonna say something like GNU style
<sortie> ULONG because unsigne dlong is too long
<mcrod`> does someone know where the hell to talk to someone about LLVM
<heat> and then of course the intel firmware people took the coding style and ran with it, now ACPICA and EFI all have it
<mcrod`> the official #llvm is consistently dead
<heat> 1) say something on IRC and pray 2) say something on their discourse thingy and pray 3) say something on discord and pray
<mcrod`> ok
<geist> yah my experience is even at google where a lot of the llvm folks are, things tend to be 'file a bug, maybe we'll look at it' for all llvm solutions
<heat> the true corporate experience
<geist> indeed
<mcrod`> wtf?
<mcrod`> that's awful
<geist> i had a dream last night that i just yelled 'i quit' and walked out of work. maybe a sign of things to come
<mcrod`> you had a good run geist
* mcrod` pet
<geist> but then the next scene of the dream i found myself back at work and a bit ashamed that i came back
<heat> mcrod`, llvm is largely a corporate project
<mcrod`> still awful
<geist> right
<heat> an example, with firmware: tianocore is completely corporate backed so there's really no community and very few people give a shit. coreboot has a very large community (apart from the few corps that do use it) and as such IRC/email is responsive
gbowne1 has joined #osdev
<geist> yeah that's the problem. even well meaning corporate teams that start contributing to a project i think have a chilling effect
<heat> as such for tianocore if you report a problem sometimes someone tells you to file a bug on the bugzilla, that someone will eventually look at
<geist> even if they mean well, think they're doing everyone a favor, it's like this big quiet elephant that comes in the room, sucks up all the air, and then everyone else feels left out
<heat> but few people will ever step out of their "zones" so to speak. I.e Intel isn't reviewing AMD/ARM code ("they're not paid to")
<geist> precisely
<geist> that sets in almost immediately
<geist> they're not in it for the general health of the project, they're in it for their needs
<geist> it may mean the overall project gets a lot of help, work, ends up being better than before, but stuff tends to be somewhat more specialized, because companies have their goals
<geist> i've been and am on the corporate side of that a few timse, and it's really disheartening (to me)
<geist> you see the damage you're doing
<heat> it gets help but if the project is corporate enough (or was born as a corporate thing) it'll end up affecting things in the long run
<heat> as soon as work sets in it seems that people get very... pragmatic
<geist> all this being said the llvm team on our project is quite good, and they actively try not to break things
<geist> but the tasks they choose to work on are clearly for teh benefit of us, and they're not paid to generically solve community problems
<geist> though i dont really speak for them
<mcrod`> well, I'm trying not to complain here
<mcrod`> the most I'll say is their documentation is sorely lacking
<mcrod`> at least for the cmake stuff
<heat> i know it is
<heat> have you seen the compiler option docs?
<geist> that is definitely true. it's surprising hjow much of the switches/pragmas/builtins/etc end up coming out of gcc docs
<heat> there are a good number of GCC switches that get silently eaten by the clang frontend and don't do jack shit because of it, silently
<mcrod`> heat: yes I have
<mcrod`> I found out that clang doesn't exactly specify what options it DOESN'T support
<mcrod`> you have to play
yoyofreeman has quit [Read error: Connection reset by peer]
yoyofreeman has joined #osdev
<mcrod`> -Wlogical-op isn't supported on clang, for example
Arthuria has quit [Remote host closed the connection]
CaCode_ has joined #osdev
CaCode has quit [Ping timeout: 255 seconds]
elastic_dog has quit [Ping timeout: 248 seconds]
<mcrod`> so, I think I see the problem
<mcrod`> you can't use LLVM_INSTALL_TOOLCHAIN_ONLY and get the CMake exports
<netbsduser``> the PALCODE for OpenVMS apparently implements the VAX's INSQUE/REMQUE instructions
<netbsduser``> or emulations thereof
<netbsduser``> these manipulate a tailqueue atomically
elastic_dog has joined #osdev
<heat> oh nice, insque
<heat> cpu-sanctioned data structures
<netbsduser``> openvms portability also consists of compiling vax macro assembly language to different target platforms
<netbsduser``> supposedly DEC was in disarray after microsoft poached most of their people and their grand VMS replacement, MICAH, was thereby dead
<heat> did they ever consider writing equivalent x86 assembly
<heat> or did that not fit into the 90s equivalent of KPI
<netbsduser``> and they had no idea what to do now about an OS for Alpha, until nancy kronenberg interrupted a meeting by crying out "vms is NOT the vax, and the vax is NOT vms"
<netbsduser``> https://dl.acm.org/doi/pdf/10.1145/151220.151224 here's her article ont hat entire endeavour
Turn_Left has quit [Ping timeout: 272 seconds]
Turn_Left has joined #osdev
pretty_dumm_guy has quit [Ping timeout: 252 seconds]
pretty_dumm_guy has joined #osdev
Turn_Left has quit [Ping timeout: 258 seconds]
Turn_Left has joined #osdev
Starfoxxes has quit [Ping timeout: 264 seconds]
janemba has quit [Ping timeout: 272 seconds]
vai has joined #osdev
netbsduser`` has quit [Ping timeout: 240 seconds]
tixlegeek has joined #osdev
netbsduser`` has joined #osdev
janemba has joined #osdev
<geist> that being said i think alpha came into existance a bit after cutler left. they were working on prism, which was shelved, and caused him to quit
<geist> and then a few years later they revived it into alpha
Turn_Left has quit [Ping timeout: 255 seconds]
Turn_Left has joined #osdev
<geist> https://en.wikipedia.org/wiki/DEC_Alpha has a pretty great history
CaCode_ has quit [Ping timeout: 255 seconds]
vai has quit [Ping timeout: 272 seconds]
Turn_Left has quit [Remote host closed the connection]
Turn_Left has joined #osdev
Turn_Left has quit [Remote host closed the connection]
Turn_Left has joined #osdev
Turn_Left has quit [Read error: Connection reset by peer]
<kazinsal> geist: yeah, I'm still about halfway through the dave cutler interview and his ability to precisely remember the tiniest details of stuff he wrote half a century ago is just incredible. I can barely remember what I had for lunch yesterday
<geist> yeah even details about some project he worked on for some other company, and what the other company was doing, etc
<heat> I mentioned POSIX in the mailing list and linus recalled a tiny dispute between linux and POSIX where the API in POSIX was wrong and why
<heat> some people just do be like that
<heat> nothing compared to the high level chess players that recall games played 50 years ago by 2 randos
<geist> also unsurprising linus would remember some dispute too
<kazinsal> cutler: "that reminds me of this really cool thing my team and I built out of baling twine and two PDP-10s"
<heat> mofo recalled it and wrote a whole essay email about it in 15 minutes
<kazinsal> torvalds: "that reminds me of the time I told a standards body to fuck themselves with a rusty pitchfork"
<mcrod`> stupid question
<mcrod`> does anyone know how to rewrite history on git, in the sense that I don't want the 40 commits of "CI test" on my repo
<mcrod`> or am I, in fact, stuck with that
<heat> al viro: "that reminds me of the time I called someone an obnoxious cunt plonk"
<kazinsal> squash your commits
<heat> yeah rebase and squash, the git push --force
<heat> then*
<mcrod`> ok
<mcrod`> thanks
<geist> keep in mind anyone that may be following your repo will be upset at you
<mcrod`> no one is
<geist> but if you dont care, or these are CLs you haven't pushed upstream, go to town
<mcrod`> :p
<mcrod`> I'm entirely still fucking with github CI pipeline
<heat> fyi you could've just amended and git push --force'd back when you were testing things. or various git commit --fixup=hash, then git rebase -i --autosquash HEAD~40 would've auto squashed everything ez
<kazinsal> git just-make-it-look-nice
<mcrod`> i was so close to getting the tarball to upload
<mcrod`> but alas... can't find it for some reason
<geist> also you should try to get familiar with `git rebase -i`
<geist> very very powerful stuff
<mcrod`> yes it'd be cool
<heat> something I accidentally learned: always pass --autosquash
<heat> autosquash does not just mean "interpret !fixup and shit"
<mcrod`> also
<heat> wait, no
<heat> i think it was -i, you always want -i
<geist> right, -i
<geist> autosquash i dunno is a good idea
<mcrod`> i'm slightly convinced that shared libraries (wrt qt, sdl etc) are your best bet on linux and friends
<heat> anyway I always pass -i --autosquash
<mcrod`> because there is simply no way you can possibly keep up with the random breakage of shit
<heat> geist, why?
<geist> well, to be honest i dont know autosquash, so i'm talkin out my butt
* gog squashes heat
<geist> but in general auto- anything is a bit risky
<kazinsal> autosquash searched for commits with tags like fixup! and selects those
<heat> yeah
<kazinsal> if you pass -i it still goes through the interactive rebase
<kazinsal> so it's not always applicable, but it saves you some time cherry-picking commits
<geist> ah i see,
<geist> okay, that's kinda nice then
tixlegeek has quit [Quit: tixlegeek]
moon-child has quit [Ping timeout: 255 seconds]
yoyofreeman has quit [Read error: Connection reset by peer]
yoyofreeman has joined #osdev
Burgundy has quit [Remote host closed the connection]
netbsduser`` has quit [Ping timeout: 272 seconds]