💽Programming Chat v2

[spare quartz]

this code finally has a use again!!

FlipBytesBE for network ordered data ..

kinda feels weird to know this code is already 4 months old

[spare quartz]

this looked normal until i saw ia32

ohhh

this image is from 2017

that explains it

plotting relations

mmm

might look into more data saving for extensible_http (do not send unicode support if all files are within ISO/IEC 8859-1)

mmm nevermind not worthwhile for processing costs

[lyric mesa]

it seems i found the cause of my storage problem

why is docker using so much space 😭

[spare quartz]

what's the biggest files

[spare quartz]

... huh

either initialized with that much data or a lot of dynamic allocation if i had to guess

i don't really use docker though

[lyric mesa]

eghhhhh

its like barely scratched it too

why must this wipe the entire disk 😭

i changed the limit to 8 gb and its... using 33.63 gb now???

it just jumped up to 68.72 gb???????

[spare quartz]

cccrappp gotta deal with file locking now... no wonder it's erroring

[spare quartz]

ada does have character ops after all!

I HATE HTTP

[timid quartz]

How many images do you have pulled

So basically just use unchecked conversion

https://bsd.network/@locnar/113507709126580936

[spare quartz]

savve me

it won't cache even with all i' mgiving it

💔

i've been awake throughout the entire night trying to add caching to this server btw

i think it might only cache on https...

[timid quartz]

uhh just

have a hash set

[spare quartz]

que

[timid quartz]

actually maybe hash map

[spare quartz]

que ...

are you telling me to cache server-side?? are you insane??

[timid quartz]

well where else are you supposed to cache

[spare quartz]

THE BROWSER

[timid quartz]

ok and why is it not caching

[spare quartz]

i don't know :<

firefox isn't clear on when their browser caches

[timid quartz]

try chrome

[spare quartz]

you-

it cached immediately

i hate firefox

[timid quartz]

https://developer.mozilla.org/en-US/docs/Web/HTTP/Caching

Read this

[spare quartz]

i have

multiple times

firefox just doesn't cache it

[timid quartz]

I only see max age here, try the example from MDN Cache-Control: no-store, no-cache, max-age=0, must-revalidate, proxy-revalidate

[spare quartz]

you are explicitly disabling caching

[timid quartz]

Oh that’s for a proxy cache mb

Try Cache-Control: private max-age=3600

[spare quartz]

no need for private as this does not carry user data

[timid quartz]

do it anyways

[spare quartz]

won't change anything

they even show an example with my setup

im 99% convinced it just wants to cache only on https

[timid quartz]

check ff settings?

[spare quartz]

doesn't have anything but clearing cache

i think it might be in the developer flags

but i dont wanna touch that

this will be a pain

i really need to improve extensible_http, it's very very low level (http wise) at the moment

[timid quartz]

Low level is ok

[spare quartz]

it's very tedious

especially since there's no intrinsic header parsing (except for Content-Length)

[timid quartz]

ughhhhhh I don’t wanna write 2 more pages of Japaneseeeeeeeeee

[spare quartz]

do it

[timid quartz]

I’m lazyyyyyy

[spare quartz]

yeah well uh

have you thought about that

[timid quartz]

https://tenor.com/view/cat-stare-catstare-cat-stare-sus-catglare-cat-glare-gif-14942558849944709546

[spare quartz]

yeah bet you didn't

god if only http wasn't catered to super simple languages

[timid quartz]

do I go to networking class…

[spare quartz]

yes......... you need to learn things idot...........

okay

caching. finally "done"

[timid quartz]

Nuh uh

I just need to learn how to force myself to write 2 pages of Japanese

[spare quartz]

in other news

i finally took the advice of moving some of th ecode outside the task loop

[timid quartz]

yipee

[spare quartz]

just one more conditional to account for

vleh my eyes hurt

oh huh

added Connection:close and now requests are super fast (<1ms)

don't know why i didn't do this before

AND NOW FIREFOX CACHES IT

🥳

[timid quartz]

because long-lived tcp connections are better

[spare quartz]

NO THEY'RE NOT!!

[timid quartz]

yuh huh'

imagine making 50 requests and having to tcp handshake for each one.

[spare quartz]

oh whatever they're like 0.3 bits anyways

[timid quartz]

thats 100 RTT wasted!!!!

actually

more than that

[spare quartz]

1 baud connection.....

[timid quartz]

200 RTT

well

hm

nah just 100 rtt

[spare quartz]

stop doing math i haven't studied tcp yet :<

[timid quartz]

client: syn
server: ack
client: synack + data

yeah only 100 RTT

no only 50 RTT

RTT is round trip so syn and ack are done in 1 RTT

still though if your rtt is like 20ms

that's 20 wasted milliseconds

20!!

[spare quartz]

and there

[timid quartz]

Connection: keep-alive ftw

you should uh respect what the client wants in terms of the connection...

so if client says keep-alive you should keep-alive..

[spare quartz]

okay yeah but im a bad task manager

[timid quartz]

[spare quartz]

if a client wants a keep alive connection they'll have to deal with it... because otherwise i need to actually use GNAT.Socket selectors

[timid quartz]

LET ME OPEN 500 KEEP-ALIVE CONNECTIONS ON YOUR SERVER

[spare quartz]

(and i have no idea how!!!)

[timid quartz]

just make normal tcp socket but don't close it

[spare quartz]

yes but that'll take up an entire thread

you need a selector to... not do that

[timid quartz]

wah wah wah who cares about that

[spare quartz]

ME cause you'll use tcp spammer or whatever and block all of them

[timid quartz]

[spare quartz]

im so grateful i added an ip argument though

-ip 127.0.0.1... testing in peace 🙏

[timid quartz]

trying to find docs for the selector

[spare quartz]

there are zero

[timid quartz]

there never are any docs..

[spare quartz]

like

theres a few stack overflow posts on them but

nobody has bothered to show an example

[timid quartz]

even in the code type Selector_Type is limited private;

[spare quartz]

that's intentional

[timid quartz]

still that means no reading the source code

[spare quartz]

anything with private; is totally implementation defined

you have to use the functions attached to it in the public part as your source

the legend:
.ads (public) -> use this for determining what to do with library code
.ads (private) / .adb -> do not use this for determination unless it's unclear what the specification entails

[timid quartz]

gay

[spare quartz]

you wouldn't read a .c file...

[timid quartz]

rust socket2::Socket::set_linger

[spare quartz]

oh no thing is i can set keep alive super easily

[timid quartz]

pub fn set_linger(&self, linger: Option<Duration>) -> Result<()>
Set value for the SO_LINGER option on this socket.

If linger is not None, a close(2) or shutdown(2) will not return until all queued messages for the socket have been successfully sent or the linger timeout has been reached. Otherwise, the call returns immediately and the closing is done in the background. When the socket is closed as part of exit(2), it always lingers in the background.

Notes
On most OSs the duration only has a precision of seconds and will be silently truncated.

On Apple platforms (e.g. macOS, iOS, etc) this uses SO_LINGER_SEC.

[spare quartz]

again the problem is tasking

hopefully this doesn't count as spamming

 type Option_Name is
     (Generic_Option,
      --  Can be used to set/get any socket option via an OS-specific option
      --  code with an integer value.

      ------------------
      -- Socket_Level --
      ------------------

      Keep_Alive,      -- SO_KEEPALIVE
      --  Enable sending of keep-alive messages on connection-oriented sockets

      Reuse_Address,   -- SO_REUSEADDR
      --  Enable binding to an address and port already in use

      Broadcast,       -- SO_BROADCAST
      --  Enable sending broadcast datagrams on the socket

      Send_Buffer,     -- SO_SNDBUF
      --  Set/get the maximum socket send buffer in bytes

      Receive_Buffer,  -- SO_RCVBUF
      --  Set/get the maximum socket receive buffer in bytes

      Linger,          -- SO_LINGER
      --  When enabled, a Close_Socket or Shutdown_Socket will wait until all
      --  queued messages for the socket have been successfully sent or the
      --  linger timeout has been reached.

      Error,           -- SO_ERROR
      --  Get and clear the pending socket error integer code

      Send_Timeout,    -- SO_SNDTIMEO
      --  Specify sending timeout until reporting an error

      Receive_Timeout, -- SO_RCVTIMEO
      --  Specify receiving timeout until reporting an error

      Busy_Polling,    -- SO_BUSY_POLL
      --  Sets the approximate time in microseconds to busy poll on a blocking
      --  receive when there is no data.

      Bind_To_Device,  -- SO_BINDTODEVICE
      --  Bind to a specific NIC (Network Interface Controller)

      -------------------------------
      -- IP_Protocol_For_TCP_Level --
      -------------------------------

      No_Delay,            -- TCP_NODELAY
      --  Disable the Nagle algorithm. This means that output buffer content
      --  is always sent as soon as possible, even if there is only a small
      --  amount of data.

      Keep_Alive_Count,    -- TCP_KEEPCNT
      --  Maximum number of keepalive probes

      Keep_Alive_Idle,     -- TCP_KEEPIDLE
      --  Idle time before TCP starts sending keepalive probes

      Keep_Alive_Interval, -- TCP_KEEPINTVL
      --  Time between individual keepalive probes

      ------------------------------
      -- IP_Protocol_For_IP_Level --
      ------------------------------

      Add_Membership_V4,   -- IP_ADD_MEMBERSHIP
      --  Join a multicast group

      Drop_Membership_V4,  -- IP_DROP_MEMBERSHIP
      --  Leave a multicast group

      Multicast_If_V4,     -- IP_MULTICAST_IF
      --  Set/Get outgoing interface for sending multicast packets

      Multicast_Loop_V4,   -- IP_MULTICAST_LOOP
      --  This boolean option determines whether sent multicast packets should
      --  be looped back to the local sockets.

      Multicast_TTL,       -- IP_MULTICAST_TTL
      --  Set/Get the time-to-live of sent multicast packets

      Receive_Packet_Info, -- IP_PKTINFO
      --  Receive low-level packet info as ancillary data

      --------------------------------
      -- IP_Protocol_For_IPv6_Level --
      --------------------------------

      Add_Membership_V6,   -- IPV6_ADD_MEMBERSHIP
      --  Join IPv6 multicast group

      Drop_Membership_V6,  -- IPV6_DROP_MEMBERSHIP
      --  Leave IPv6 multicast group

      Multicast_If_V6,     -- IPV6_MULTICAST_IF
      --  Set/Get outgoing interface index for sending multicast packets

      Multicast_Loop_V6,   -- IPV6_MULTICAST_LOOP
      --  This boolean option determines whether sent multicast IPv6 packets
      --  should be looped back to the local sockets.

      IPv6_Only,           -- IPV6_V6ONLY
      --  Restricted to IPv6 communications only

      Multicast_Hops       -- IPV6_MULTICAST_HOPS
      --  Set the multicast hop limit for the IPv6 socket
     );

set via

[timid quartz]

it does

jk

[spare quartz]

uploading entire jdks again 🙏

[timid quartz]

can you not just uh

use a condvar

to put the task to sleep

[spare quartz]

a what what

[timid quartz]

a condition variable

smh

doesn't even know the basics of multithreading

[spare quartz]

that's what select does...

what do you think a selector is.....

bro doesn't even know the basics of tasking

[timid quartz]

static pthread_mutex_t mtx = PTHREAD_MUTEX_INITIALIZER;
static pthread_cond_t cond = PTHREAD_COND_INITIALIZER;

static int avail = 0;

while (TRUE) {
    s = pthread_mutex_lock(&mtx);

    while (avail == 0) {   /* Wait for something to consume */
       s = pthread_cond_wait(&cond, &mtx);
    }

    while (avail > 0) {   /* Consume all available units */ 
        avail--;
    }

    s = pthread_mutex_unlock(&mtx);
}

[spare quartz]

Ada RM 9.7 pilled 🙏

[spare quartz]

urrrgghh

[timid quartz]

pain

ok is there

nvm I could just do a lookup table

[spare quartz]

what

[timid quartz]

or actually can I change the repr of an enum

[spare quartz]

yes

use a representation clause

e.g.

[timid quartz]

im reading the machine id bits from the exe fiel

and I didn't want to make a giant switch-case

[spare quartz]

make a giant switch case

[timid quartz]

noooo

there must be a better way...

[spare quartz]

(technically, a switch is just compiler mnemonic for a lookup table (more appropriately a jump table) but who cares)

(yes use a lut)

[timid quartz]

there mus tbe a better wayyyyyyyyyyyyy

[timid quartz]

so..hash table..that maps

the things

to the things

[spare quartz]

i don't knowwwwwwww

you might need an ordered table if you can't cough up a hashing function

that reminds me

[timid quartz]

ugh stupid

why doesnt it just provide a hashing function by default

switch-case it is

[spare quartz]

it's totally valid to just make a function that returns anything as a normal hash in java

[spare quartz]

USE AN ORDERED TABLE

!!

wait

nnnnnnooooo you can't ignore that thought

[spare quartz]

becuase you can make an table (map?) of any type

how would you make a hash function for... i dunno, an access-to-record

[timid quartz]

idk

hash the pointer

[spare quartz]

for the record it does provide some hashing functions for you

like Ada.Strings.Hash

[timid quartz]

there's no way to instantiate it with a literal

I'd have to like

[spare quartz]

wdym

[timid quartz]

add it to a read function or smth

[spare quartz]

nah why is bayachao so cute

[timid quartz]

do like

type Machine_Type is (All_Of_Them);
--  ...
procedure Parse_Machine_Id (H : out Header; Id : Unsigned_16) is
  package La_Mapa is blah;
  M : La_Mapa;
begin
  M.Include (16#0#, Machine_Type.Unknown);
  --  ALL THE REST OF THEM
  --  ...
  H.Machine_Id := M (Id);
end Parse_Machine_Id;

@spare quartz

[spare quartz]

wait why are you just not using an enum

is he stupid

[timid quartz]

wait................................................

i can just....................

[spare quartz]

yes

[timid quartz]

read the enum...........................

[spare quartz]

igiot

[timid quartz]

yeah

............

IDGUIT

[spare quartz]

HERE

[timid quartz]

wait but

I have to go number -> enum

not enum -> number

[spare quartz]

'Val

[timid quartz]

'Val...

stupid language

stupid furries

stupid dod

[spare quartz]

ugh darpa should make ada 2035 to beat the crap out of rust

[timid quartz]

no

rust should beat the crap out of ada

[spare quartz]

your language is literally half baked

[timid quartz]

NUH UH

[spare quartz]

our language is mega baked.............. in one single oven

[timid quartz]

at least my language isnt uh...verbose
and uh...has a community
and uh.........doesnt have exceptions

and uh........borrow checker

[spare quartz]

me when i act silly and become a maintainer in a super used rust project like tokio or whatever

and insert panic!

o h yaeh

i need to look into the GNAT security features again

they have this one flag that makes return oriented programming practically impossible

super cool (but it can heavily affect perf)

[timid quartz]

me when I act silly and become a maintainer in a super used ada project like idk nobody uses ada
and insert Abort_Task!!!!

or GNAT.OS_Lib.OS_Exit

[spare quartz]

bro found the one ... i dont evne know what that is

[spare quartz]

me when i export an ada function to c exit (its just a noop

cat get offmy key bot

[timid quartz]

look at all those chickens

[spare quartz]

i bet i could beat their library...

[timid quartz]

Nope

[spare quartz]

what

[timid quartz]

You couldn’t

[spare quartz]

i could ......

my cat is curhign me so hard

[timid quartz]

based cat

stop atp from working

[timid quartz]

Nuh uh

[spare quartz]

okay true

i only have http 1.1..

also i got no idea how they have this many files

[timid quartz]

A lot are probably conditionally not compiled due to feature flags

When you were benchmarking your server, what did you have it doing?

[spare quartz]

im probably at the same level of functionality to them if im gonna be honest (excl http/2)

i just delegate a lot more to the end programmer

[timid quartz]

Just serving a static file?

[spare quartz]

directory listing

which is actually the only dynamic thing in the server now that i think about it

directory list has to be computed for each request, but it usually does it in under half a milisecond (haven't benchmarked that part yet)

function Get_Directory_Listing (For_Search : in out Search_Type; Request : HTTP_11_Request_Message; Composed_Actual : String) return Directory_Listing_Result is
   begin
      declare

         Result      : Directory_Listing_Result;
         Padded_Kind : String (1 .. Ordinary_File'Image'Length);
         Padded_Size : String (1 .. 12);

         Search_Item         : Directory_Entry_Type;
         Send_Human_Friendly : Boolean := Request.Fields.Contains ("Accept-Language");

      begin
         For_Search.Start_Search (Directory => Composed_Actual, Pattern => "");

         while For_Search.More_Entries
         loop
            For_Search.Get_Next_Entry (Search_Item);

            if Search_Item.Modification_Time > Result.Highest_Modified
            then
               Result.Highest_Modified := Search_Item.Modification_Time;
            end if;

            if (Search_Item.Simple_Name = ".." and then Composed_Actual /= Current_Directory) or else (Search_Item.Simple_Name /= "." and then Search_Item.Simple_Name /= "..")
            then
               if Send_Human_Friendly
               then
                  Ada.Strings.Fixed.Move (Search_Item.Kind'Image, Padded_Kind, Justify => Ada.Strings.Right);

                  if Search_Item.Kind = Ordinary_File
                  then
                     Ada.Strings.Fixed.Move (Search_Item.Size'Image & 'B', Padded_Size, Justify => Ada.Strings.Right);

                  else
                     Ada.Strings.Fixed.Delete (Padded_Size, 1, Padded_Size'Length);
                  end if;
                  Result.Contents.Append (Padded_Size & ' ' & Padded_Kind & ' ' & Search_Item.Simple_Name & ASCII.LF);

               else
                  Result.Contents.Append
                    (Search_Item.Simple_Name & ASCII.NUL & Search_Item.Size'Image
                     & (if Search_Item.Kind = Ordinary_File then 'F' elsif Search_Item.Kind = Special_File then 'S' else 'D') & ASCII.LF);
               end if;
            end if;
         end loop;
         For_Search.End_Search;

         return Result;
      end;
   end Get_Directory_Listing;

directory listing code which is then passed into GET_Procedures (minimal writing)

Highest_Modified : Time := GNAT.Calendar.No_Time; -- interestingly this could give a cache result from "1901" if no files are in the directory and it's the root ...

[spare quartz]

reading about serde

rust really is that much library oriented huh

[timid quartz]

yeah

rust is incredibly library-oriented

[spare quartz]

kinda surprising to see for me honestly

kotlin has the entirety of the JDK toolkit which is massive
and ada has... a lot of things (gnat on top too)

i'm just not used to that

hm

should probably add a -rom argument to basic_http_server

like -d just doesn't allow PUT

[timid quartz]

@spare quartz this uses tokio instead of plain rust threads but benchmark this

[spare quartz]

should i compile the ada one in fast mode or keep it in debug

[timid quartz]

uh whichever you want

just compile both in the same mode

so if you do ada in debug do rust in debug

if you do ada in fast do rust in release

[spare quartz]

uhhh

wow this compiler really is slow...

[spare quartz]

how do i use this

the program, when running cargo run

[timid quartz]

localhost:3000

cargo run is debug by default

[spare quartz]

top is mine, bottom is the one you sent

both compiled in unoptimized w/ debug

i also got this flooded ...

[timid quartz]

huh interesting

I wonder what that is

[spare quartz]

oh crap

i forgot, when running my server, without an Accept-Language provided by a browser it'll send binary data of the directory instead of human readable stuff

unsure if that contributes much but just letting you know

[timid quartz]

I’ll try to fix the errors

But I need to write Japanese first

[spare quartz]

nerd..

WAIT

I KNOW WHAT YOU SHOULD WRITE!!

[lyric mesa]

i heard japanese

[spare quartz]

@timid quartz https://fixupx.com/i/status/1857876458896765230

[spare quartz]

specifically <name>\0<size><F/D/S descriptor>\n

what a weird name

[timid quartz]

hm

[timid quartz]

https://github.com/hyperium/hyper/issues/3669#issuecomment-2120480121

Likely the combination of wrk tossing the connection quickly and being over localhost, the hyper server notices the connection is no longer connected when trying to shutdown. It returns the error from the operating system, letting the user decide whether to care based on the context. In your case, you can likely just ignore that error.

hm

[spare quartz]

oh

neat

[timid quartz]

maybe it would be a better comparison to implement http without a library

because hyper kinda relies on tokio

you can't just use normal threads

[spare quartz]

i'd imagine it also does other processing of the request

[timid quartz]

yeah

[spare quartz]

oh yeah i also modified how the timeout works for my http

reading the HTTP message itself is under the time limit, but the body isn't

[timid quartz]

bro wtf...I was writing this japanese thing on my pc and it came out to almost a page, but now that I'm on my laptop it's like a little over 3/4 of a page???

[spare quartz]

dementia

[timid quartz]

what is the difference between the windows and macos font...

[spare quartz]

such a weird type!!

[timid quartz]

me when I flip your bytes

[spare quartz]

okay but how

[timid quartz]

procedure Flip_Bytes (A : in out Atp)

that's how

[spare quartz]

when in doubt

uncheck it

[spare quartz]

cooooooooooooll

[spare quartz]

very nice

(RECORD_TYPE => HANDSHAKE,
 RECORD_PROTOCOL_VERSION => TLS_10,
 RECORD_LENGTH =>  512,
 MESSAGE_TYPE => CLIENT_HELLO,
 MESSAGE_LENGTH =>  64513,
 CLIENT_PROTOCOL_VERSION => TLS_12,
 CLIENT_RANDOM => "&�ޞ���o� �����mͤ!����W��AP�",
 CLIENT_SESSION_ID => 
 [":�&kt��i��Q�r*��벴���\��
                           �&"],
 CLIENT_CIPHER_SUITES => 
 [
  [TLS_AES_256_GCM_SHA384, TLS_CHACHA20_POLY1305_SHA256, TLS_AES_128_GCM_SHA256,
   UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN,
   UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN,
   UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN,
   UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, UNKNOWN, TLS_EMPTY_RENEGOTIATION_INFO_SCSV]],
 CLIENT_COMPRESSION_METHODS => 
 [
  [NONE]],
 CLIENT_EXTENSIONS => 
 [
  (EXTENSION_TYPE => SERVER_NAME,
   DATA => 
   ["
        localhost"]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   [""]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   [""]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   [""]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   ["
     hhttp/1.1"]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   [""]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   [""]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   [""]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   ["(

"]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   [""]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   [""]), 
  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   ["$ �y���(�}��grXM��+����ٝ���v��Ֆ7"]),

  (EXTENSION_TYPE => UNKNOWN,
   DATA => 
   [""])])

the entire CLIENT_HELLO packet as a record

i am very tired

[timid quartz]

[timid quartz]

me too..

I'm probably around a page and a third of my japanese

for some reason, windows makes the text take slightly more space

even though both fonts are 12 point and the spacing and the font itself should be the same

[spare quartz]

Tomorrow will be a dark day…

• need to create X25519 keys
• need to deal with 256 bit numbers
• need to securely generate random
• need to read certificate files

[timid quartz]

wait the for thing for the enum has to be ordered???

[spare quartz]

Yes

[timid quartz]

ughhhh

that's so stupid

[spare quartz]

The compiler should’ve given you an error — did it?

[timid quartz]

order it yourself stupid compiler

yeah it did

[spare quartz]

write your code better then!!

unorganized …

[spare quartz]

Oh yeah

I also gotta fix all of these unknowns

Except for Cipher suites since TLS 1.3 only defines 5

[timid quartz]

rrrr

how do I order thiss

it starts with 16#0#

[spare quartz]

It’s just

Ordered 1:1 with the enum def

[timid quartz]

WHAT

[spare quartz]

You don’t need to sort the numbers iirc

[timid quartz]

THAT'S HOW IT WAS ORIGINALLY

[spare quartz]

What did you do…….

[spare quartz]

Wait no… you do… or don’t… I have dementia

[timid quartz]

type Machine_Type is (Unknown, Alpha32, Alpha64, Am33, Amd64,
                         Arm32, Arm64, ArmNT, EfiBytecode,
                         I386, Ia64, LoongArch32, LoongArch64,
                         M32R, Mips16, MipsFpu, MipsFpu16,
                         PowerPc, PowerPcFp, R4000, RiscV32,
                         RiscV64, RiscV128, Sh3, Sh3Dsp, Sh4,
                         Sh5, Thumb, WceMipsV2) with Size => 16;
   for Machine_Type use
     (Unknown => 16#0#,
      Alpha32 => 16#184#,
      Alpha64 => 16#284#,
      Am33 => 16#1D3#,
      Amd64 => 16#8664#,
      Arm32 => 16#1C0#,
      Arm64 => 16#AA64#,
      ArmNT => 16#1C4#,
      EfiBytecode => 16#EBC#,
      I386 => 16#14C#,
      Ia64 => 16#200#,
      LoongArch32 => 16#6232#,
      LoongArch64 => 16#6264#,
      M32R => 16#9041#,
      Mips16 => 16#266#,
      MipsFpu => 16#366#,
      MipsFpu16 => 16#466#,
      PowerPc => 16#1F0#,
      PowerPcFp => 16#1F1#,
      R4000 => 16#166#,
      RiscV32 => 16#5032#,
      RiscV64 => 16#5064#,
      RiscV128 => 16#5128#,
      Sh3 => 16#1A2#,
      Sh3Dsp => 16#1A3#,
      Sh4 => 16#1A6#,
      Sh5 => 16#1A8#,
      Thumb => 16#1C2#,
      WceMipsV2 => 16#169#);

wait

why are there awo arm64

nvm im blind

one is amd64

[spare quartz]

and this causes an error?

[timid quartz]

yes

[spare quartz]

Sooorrrtt by the hex then and see what happen

[timid quartz]

coff_header.ads:27:15: error: enumeration value for "Am33" not ordered
coff_header.ads:29:16: error: enumeration value for "Arm32" not ordered
coff_header.ads:31:16: error: enumeration value for "ArmNT" not ordered
coff_header.ads:33:15: error: enumeration value for "I386" not ordered
coff_header.ads:38:17: error: enumeration value for "Mips16" not ordered
coff_header.ads:41:18: error: enumeration value for "PowerPc" not ordered
coff_header.ads:43:16: error: enumeration value for "R4000" not ordered
coff_header.ads:47:14: error: enumeration value for "Sh3" not ordered
coff_header.ads:52:20: error: enumeration value for "WceMipsV2" not ordered

[timid quartz]

I did

still error

[spare quartz]

Also I’d suggest padding the hex

[spare quartz]

Huhg send the sorted

[timid quartz]

coff_header.ads:25:22: error: enumeration value for "I386" not ordered
coff_header.ads:26:22: error: enumeration value for "R4000" not ordered
coff_header.ads:27:22: error: enumeration value for "WceMipsV2" not ordered
coff_header.ads:29:22: error: enumeration value for "Sh3" not ordered
coff_header.ads:33:22: error: enumeration value for "Arm32" not ordered
coff_header.ads:35:22: error: enumeration value for "ArmNT" not ordered
coff_header.ads:36:22: error: enumeration value for "Am33" not ordered
coff_header.ads:37:22: error: enumeration value for "PowerPc" not ordered
coff_header.ads:40:22: error: enumeration value for "Mips16" not ordered

[spare quartz]

Send the sorted enum!!!!!

[timid quartz]

ughhh

I have to re sort it

I changed the redo history

[spare quartz]

Just press CTRL z

Oh my odddd

[timid quartz]

   type Machine_Type is (Unknown, Alpha32, Alpha64, Am33, Amd64,
                         Arm32, Arm64, ArmNT, EfiBytecode,
                         I386, Ia64, LoongArch32, LoongArch64,
                         M32R, Mips16, MipsFpu, MipsFpu16,
                         PowerPc, PowerPcFp, R4000, RiscV32,
                         RiscV64, RiscV128, Sh3, Sh3Dsp, Sh4,
                         Sh5, Thumb, WceMipsV2) with Size => 16;

   for Machine_Type use
     (Unknown     => 16#0#,
      I386        => 16#14C#,
      R4000       => 16#166#,
      WceMipsV2   => 16#169#,
      Alpha32     => 16#184#,
      Sh3         => 16#1A2#,
      Sh3Dsp      => 16#1A3#,
      Sh4         => 16#1A6#,
      Sh5         => 16#1A8#,
      Arm32       => 16#1C0#,
      Thumb       => 16#1C2#,
      ArmNT       => 16#1C4#,
      Am33        => 16#1D3#,
      PowerPc     => 16#1F0#,
      PowerPcFp   => 16#1F1#,
      Ia64        => 16#200#,
      Mips16      => 16#266#,
      Alpha64     => 16#284#,
      MipsFpu     => 16#366#,
      MipsFpu16   => 16#466#,
      RiscV32     => 16#5032#,
      RiscV64     => 16#5064#,
      RiscV128    => 16#5128#,
      LoongArch32 => 16#6232#,
      LoongArch64 => 16#6264#,
      Amd64       => 16#8664#,
      M32R        => 16#9041#,
      Arm64       => 16#AA64#,
      EfiBytecode => 16#EBC#);

[spare quartz]

Send error

[timid quartz]

coff_header.ads:25:22: error: enumeration value for "I386" not ordered
coff_header.ads:26:22: error: enumeration value for "R4000" not ordered
coff_header.ads:27:22: error: enumeration value for "WceMipsV2" not ordered
coff_header.ads:29:22: error: enumeration value for "Sh3" not ordered
coff_header.ads:33:22: error: enumeration value for "Arm32" not ordered
coff_header.ads:35:22: error: enumeration value for "ArmNT" not ordered
coff_header.ads:36:22: error: enumeration value for "Am33" not ordered
coff_header.ads:37:22: error: enumeration value for "PowerPc" not ordered
coff_header.ads:40:22: error: enumeration value for "Mips16" not ordered

[spare quartz]

And also add the original enum on top

That’s not the enum…

[timid quartz]

wrong

fixed

[spare quartz]

That is

The enum isn’t sorted

[timid quartz]

hwo

do the NAMES have to be sorted?

[spare quartz]

annnndd values

[timid quartz]

this is so GAY

[spare quartz]

[timid quartz]

WHY

DO THEY HAVE TO BE SORTED

[spare quartz]

[timid quartz]

so I have to sort the names alphabetically

and then the values

[spare quartz]

Nooo

Not alphabetically

[timid quartz]

?????

hlep

[spare quartz]

They just have to appear in the same order in both the clause and the enum

[timid quartz]

THEY DO

[spare quartz]

And the values are to be sorted

[timid quartz]

THEY WERE

[spare quartz]

they were hmmm

[spare quartz]

What you’ve shown up here is a violation of the first…

[timid quartz]

in this, the names appear in the same order

do I have to arrange the enum such that the names appear in the same order as the values's ordering?

so like

Unknown has to come first in the enum because the value is 16#0#

[spare quartz]

type A is (A, B, C);
for A use (A => 0, B => 1, C => 2);

personally I use 16#FF_FF# for unknown enums but idk if that’s reserved by the format here

Ladybird uses a new browser engine called LibWeb that is being created from scratch by the development team. Unlike SerenityOS, it will also use other open source libraries for development.[3] An ad blocking feature is planned.[7]

mmmsmmashhhggghhs external libraries

I become a zombie the moment I see them

[timid quartz]

type Machine_Type is (Unknown, I386, R4000, WceMipsV2, Alpha32, Sh3,
                         Sh3Dsp, Sh4, Sh5, Arm32, Thumb, ArmNT, Am33, PowerPc,
                         PowerPcFp, Ia64, Mips16, Alpha64, MipsFpu, MipsFpu16,
                         RiscV32, RiscV64, RiscV128, LoongArch32, LoongArch64,
                         Amd64, M32R, Arm64, EfiBytecode) with Size => 16;
   for Machine_Type use
     (Unknown     => 16#0#,
      I386        => 16#14C#,
      R4000       => 16#166#,
      WceMipsV2   => 16#169#,
      Alpha32     => 16#184#,
      Sh3         => 16#1A2#,
      Sh3Dsp      => 16#1A3#,
      Sh4         => 16#1A6#,
      Sh5         => 16#1A8#,
      Arm32       => 16#1C0#,
      Thumb       => 16#1C2#,
      ArmNT       => 16#1C4#,
      Am33        => 16#1D3#,
      PowerPc     => 16#1F0#,
      PowerPcFp   => 16#1F1#,
      Ia64        => 16#200#,
      Mips16      => 16#266#,
      Alpha64     => 16#284#,
      MipsFpu     => 16#366#,
      MipsFpu16   => 16#466#,
      RiscV32     => 16#5032#,
      RiscV64     => 16#5064#,
      RiscV128    => 16#5128#,
      LoongArch32 => 16#6232#,
      LoongArch64 => 16#6264#,
      Amd64       => 16#8664#,
      M32R        => 16#9041#,
      Arm64       => 16#AA64#,
      EfiBytecode => 16#EBC#);

ok

now

just EfiBytecode is unordered

[spare quartz]

It’s correct

[timid quartz]

compiler says otherwise

coff_header.ads:50:22: error: enumeration value for "EfiBytecode" not ordered

[spare quartz]

That’s what I’m saying

The compiler is correct

[timid quartz]

how fix...

[spare quartz]

Efi needs to be before Riscv32

[timid quartz]

[spare quartz]

This is why you pad the hexadecimal!!!!!

makes it easier to spot

[timid quartz]

ugh this language is stupid and gay and dumb

[spare quartz]

Servo is an experimental browser engine designed to take advantage of the memory safety properties and concurrency features of the Rust programming language. It seeks to create a highly parallel environment, in which rendering, layout, HTML parsing, image decoding, and other engine components are handled by fine-grained, isolated tasks.

[spare quartz]

Rust user when they have to be consistent for once 👻

[timid quartz]

yeah but IMAGE_FILE_MACHINE_EBC | 0xebc | EFI byte code

not specified if it's 0x0EBC or 0xEBC0

[spare quartz]

Well, when there’s less digits in the hex that means it’s smaller

0xA000 > 0xFBA

0x0A00 < 0xF00

(left most is pad)

(but you should also add _ for separation between bytes)

[full berry]

very broad question
what are good ways to check permissions in an admin system (discord or roblox)
or generally like a permission system

[spare quartz]

What

Wdym permissions

[full berry]

like

lets say you have uhhhhhh

idfk member mod admin owner or something

[spare quartz]

In like, a Roblox group?

[full berry]

no

admin system

[spare quartz]

You don’t

That’s something you define

[full berry]

yes i know im

asking for ideas

[spare quartz]

Do anything

[full berry]

member has no perms
mod has member perms + kick perms
admin has mod perms + ban
owner has admin + blacklist
(idk i just made this up)

[spare quartz]

Like… permissions boil down to something being true or false, or a number

[full berry]

i was thinking of an inheritance system

[spare quartz]

Have a list of permission names or what have you and make them numbers or boolenas

[timid quartz]

@spare quartz constraint_error

H.Machine_ID := Machine_ID_Type'Val (Unsigned_16'Input (S));

[spare quartz]

Wdym

[spare quartz]

Okay, give more information

What’s Machine_ID_Type and what’s the error

[timid quartz]

index check failed

[timid quartz]

machine_id_type is the enum

[full berry]

liiike the higher rank your is the more perms you get
but perms wouldnt have to be redefined, you would just inherit all of the permissions of lower ranks

[timid quartz]

I just renamed it

[full berry]

if that makes sense

[spare quartz]

A table of permission names with numbers then

[full berry]

wait doesnt shield just use that

[spare quartz]

Do I look like a shield dev

[timid quartz]

kinda yeah

shield perms need to be redone though

[full berry]

yeah um...
why does @Roblox have level 10...

[timid quartz]

¯_(ツ)_/¯

[spare quartz]

Anyways uh you’re reading an invalid value obviously, you need to use an unchecked method

Then apply 'Valid

[timid quartz]

[spare quartz]

It’s fine

[timid quartz]

so..

1. Unsigned_16'Read
2. Unchecked_Conversion
3. ?

[spare quartz]

3. Use 'Valid on the enum, if it’s false, assign to Unknown, otherwise pass

Do not read the enum before assigning. Just use the Valid aspect on it

otherwise you get a spooky invalid daya error

(Are you sure the stream isn’t being misaligned btw?)

[timid quartz]

yes Im sure the stream isn't misaligned

if I read the machine id as plain u16 it outputs the correct thing

[spare quartz]

Mmmm

[timid quartz]

namely 16#8664#

[spare quartz]

Obama wants a mcdouble

[timid quartz]

      Machine_ID_Thing := Convert_Machine_ID (Unsigned_16'Input (S));
      if Machine_ID_Thing'Valid then
         H.Machine_ID := Machine_ID_Thing;
      else
         H.Machine_ID := Unknown;
      end if;

ugh this is so stupid and gay

but it works

[spare quartz]

Oh yknow what you’re supposed to use Enum_Val, a 2022 thing… but that won’t handle Unknowns for you

[timid quartz]

well...I'm operating under the assumption that the file isn't corrupt

and unknown is a defined value

[spare quartz]

You should inline if

:3

[spare quartz]

Oh

Then use enum_Val!

you might need to add Pragma Ada_2022 at the top of your file though assuming the project is in 2012 mode (default)

Omaneko

[timid quartz]

Compiling with -gnat2022

[spare quartz]

That works too

[timid quartz]

where is..

enum_val...

[spare quartz]

'Enum_Val

[timid quartz]

found it

[spare quartz]

https://learn.adacore.com/courses/whats-new-in-ada-2022/chapters/enum_val.html

[timid quartz]

it works

[spare quartz]

Nya

[timid quartz]

is it possible to use Put_Line to output an actual tab character rather than a literal "\t"

I know i can like

use the latin charset and do HT

but that's verbose because you have to "string 1" & HT & "string 2"

[spare quartz]

What like this https://gcc.gnu.org/onlinedocs/gcc-14.1.0/gnat_rm/String-interpolation.html ?

[timid quartz]

oh do you have to do f""

[spare quartz]

I’m pretty sure this is a GNAT extension so

probably

[timid quartz]

ugh gnat extension

whatever

[spare quartz]

GNAT is the only compiler that matters

You’re not gonna use like, HP Ada or whatever

[timid quartz]

I'll just type literal tabs myself then

[spare quartz]

Wha

[spare quartz]

You don’t wanna use the extension??

[timid quartz]

how do u even use the gnat extension

[spare quartz]

-gnatX

[timid quartz]

ughhh it's almost 3 almost time to go back and suffer by writing japanese

death by 3000 characters

[spare quartz]

or if you try using it in code it’ll tell you what flag to use

[timid quartz]

I am at 1500 characters

in this forsaken paper

[spare quartz]

Mr mrroowww

I’m so sleeeepyyy . I woke up at 3 pm yesterday

[timid quartz]

go eep

rn.

[spare quartz]

Ugghhh fineeeee

[timid quartz]

game ending screen

ending 9/20

character 1/20

oh yikes that's a failure ending too

[timid quartz]

@spare quartz hmm surely there's a better way to read a struct than a giant procedure

ughhh range check failed again

possible misalignment

[lavish dove]

misalignment crazy

[timid quartz]

mm no seems to not be misaligned

[timid quartz]

forgot to read a record member

lmao

[spare quartz]

Not that I know of

[timid quartz]

https://tenor.com/view/cringe-gif-22260074

[spare quartz]

Ada’s default RW procedures are usually sufficient for basic scalars, but when it comes to composite stuff like records it’s more geared towards like

Communication between Ada apps or the same app

You can see this with Input/Output, which sends stuff like discriminants over (which other apps can’t understand, but ada can)

[spare quartz]

Also you just reminded me I really need to make enum reading generic

[timid quartz]

https://tenor.com/view/profile-picture-gif-25124188

[spare quartz]

I have essentially three of the same function, just differing in unsigned length and the target enum

[timid quartz]

also im SO CLOSE to being done with this accursed paper

[spare quartz]

What is it about anyways

[timid quartz]

有名な少年主人公と日本社会

[spare quartz]

Ohhhh

So it’s a scientific fanfic

[timid quartz]

...no?

[spare quartz]

yyyeessss…

[timid quartz]

it's about how shonen anime protagonists relate to japanese soceity

specifically elements of japanese society reflected in said protagonists

don't tell me I'm wrong, I'm literally writing the damn paper

I have written 2.7k characters

over the past like 2.5 days

[spare quartz]

I’m not telling you you’re wrong that’s just how I interpret it

[timid quartz]

listen normally I think I'm right but in this case your interpretation is actually objectively wrong

this is not a fanfic at all and is a research paper

[spare quartz]

precisely

[timid quartz]

a very shoddy research paper

fucking windows ime can't even know to type 炭治郎 (tanjiro) correctly

and yes ik it's tanjirou (with a long o at the end) but windows ime doesn't

[spare quartz]

honestly no clue how that thing even works

I’ve seen it like once in action

[timid quartz]

romaji -> hiragana (or direct to katakana) -> katakana or kanji

"ookii" -> おおきい -> 大きい

"anime" -> アニメ

[spare quartz]

Oh so it’s a glorified LUT… with probably a lot more data than I could imagine

[timid quartz]

pretty much yeah

with also some temporal locality built in

[timid quartz]

it wasn't macOS's first guess either, BUT it learned pretty damn quickly that I wanted to type that

windows IME doesn't even display it as an option

[spare quartz]

We love windows

[timid quartz]

at least it's not fcitx5 or ibus

looks at kde and gnome respectively

fucking mozc

[proud creek]

https://discord.com/blog/how-discord-reduced-websocket-traffic-by-40-percent

pretty cool

[timid quartz]

wow acc kinda neat

[spare quartz]

Damn now I wanna make a discord api… again…

[timid quartz]

procedure Ugh is
  type Me is record
    Is_Alive : Boolean := True;
  end record;
  procedure Yeet (M : in out Me) is
  begin
    M.Is_Alive := False;
  end Yeet;
  Literally_Me : Me;
is
  Yeet (Literally_Me);
end Ugh;

[timid quartz]

ada time?

[spare quartz]

I need to finish TLS first soooo

Nooo not yet

[timid quartz]

gay

[spare quartz]

and I think I’d rather make a Minecraft server first but idk

if I do ever do that LSP I’ll probably fail

Hm you know i think it just occurred to me

Ada doesn’t free up memory for you because it’d probably be bad in embedded/RT environments

[timid quartz]

makes sense how it gets away without a borrow checker

[timid quartz]

wdym

you'd be fine

[spare quartz]

Semantic analysis of code is hard thougghhh

I don’t even know where I’d start

[timid quartz]

ugh

make abstract syntax tree

then analyze

[spare quartz]

that’s just saying “make a compiler” to me….

[timid quartz]

basically

honestly I'm taking a compilers course next semester so I might be equipped to do this after that class

[spare quartz]

That is one thing I’ve wanted to do for a while since GCC is kinda annoying to deal with but like

Code generation looks scary as hell

[timid quartz]

meh just generate assembly and feed it to gcc

or generate llvm ir

[spare quartz]

me when the assembly I generate makes the cpu die (I forgot to account for the 1 billion different external vars that determine the state of a processor)

[timid quartz]

naaah the processor handles that

[spare quartz]

if anything that’s probably why a lot of my VMs I’ve made corrupt themselves: I never account for CPU state right

[spare quartz]

Maybe

LLVMs giant though, I’d probably be sitting in the stability of something like Zig

I’d imagine the first step to making an LSP is figuring out that protocol they all use right

[timid quartz]

yeah I linked said protocol

[spare quartz]

Yeah I can do that

[timid quartz]

actually honestly

instead of making an lsp from scratch

taking ada_language_server and fixing parts of it would also be acceptable

that way you don't have to worry as much about semantic analysis

[spare quartz]

ALS doesn’t handle semantic analysis itself, that’s part of libadalang

there’s also ASIS (standardized semantic analysis) but that hasn’t been updated since like, Ada 05

[timid quartz]

ok well

ig you could latch onto libadalang for that then

[spare quartz]

Bark bark

[timid quartz]

wrong animal idiot

[spare quartz]

It is time to wake up

[spare quartz]

wow okay

looking at the wikibooks page for GNAT.Sockets

edited 12 times in the span of 19 years

[timid quartz]

@spare quartz it’s doone

[spare quartz]

what is

your paper?

[timid quartz]

Y. e. S

[timid quartz]

Thanks for the auto-inserted periods iPhone

[timid quartz]

blank

[spare quartz]

mmmmmmmmmmm how do i wanna do this

[timid quartz]

oh it says SPARQ but the text is black and my discord theme is black so I can’t see it

[spare quartz]

it says sparc

its the architecture

[timid quartz]

@wheat zinc I summon thee, Sun SPARC chip mentioned

[spare quartz]

secure random generator

• figure out all the math to make one myself
• pragma Import(C, "__gcc...")
• System.Machine_Code
• use the insecure random generator

[timid quartz]

WolfSSL

[spare quartz]

i'm not adding a set of dlls to my project

[timid quartz]

Import openssl then

[spare quartz]

no

an ada library already exists for that

and i already have the code done for handshaking, its just encryption that needs to be solved

god

this is what goes into reading a client handshake if you can see this file

[timid quartz]

When I was in college in the early 70s, I devised what I believed was a brilliant encryption scheme. A simple pseudorandom number stream was added to the plaintext stream to create ciphertext. This would seemingly thwart any frequency analysis of the ciphertext, and would be uncrackable even to the most resourceful government intelligence agencies. I felt so smug about my achievement.

Years later, I discovered this same scheme in several introductory cryptography texts and tutorial papers. How nice. Other cryptographers had thought of the same scheme. Unfortunately, the scheme was presented as a simple homework assignment on how to use elementary cryptanalytic techniques to trivially crack it. So much for my brilliant scheme.

From this humbling experience I learned how easy it is to fall into a false sense of security when devising an encryption algorithm. Most people don’t realize how fiendishly difficult it is to devise an encryption algorithm that can withstand a prolonged and determined attack by a resourceful opponent.

• Phil Zimmermann, PGP creator

[spare quartz]

well yeah that's why im not making my own algo

i can't do that anyways -- its a violation of the TLS 1.3 spec

oh yeah

FFMPEG made a tweet today about ISO gatekeeping documents

i needed that post... a month ago

they even linked the exact document i was using for that mp4 decoder

[timid quartz]

heh

[spare quartz]

which wouldn't help much if it were free; that thing is horrible to read

[timid quartz]

They knew

[spare quartz]

https://fixupx.com/FFmpeg/status/1858654093985935740

[timid quartz]

waah waah look at me I need everything to be foss I can’t stand using proprietary software that actually just works

[spare quartz]

the structure of an MP4 file is incredibly basic to read as Apple basically invented it and they open source their crap

but holy crap AVC* as described in the ISO document is literal alien language

they just shove paragraph over paragraph that amounts to "these 2 bytes are the type of x"

https://fixupx.com/wavefnx/status/1846540695361790231

uuuuuughhhhh

1,337 lines of rust and all you made is a box decoder

[timid quartz]

all those ada files and all you made is a special box (png) decoder

[spare quartz]

pal the time i made an mp4 box decoder in roblox was at maximum 400/500 lines

[timid quartz]

Ok but how many features did you skimp on

[spare quartz]

zero

[timid quartz]

bet you didn’t decode the audio

[spare quartz]

neither did this guy

[timid quartz]

hm

could just be badly written then

He did say it could be minified more

[spare quartz]

i don't have their code so i can't really confirm nor deny

also rust's semantics might require more stuff per line than lua

should be noted that the PNG decoder is also dependency-less

but again: that's cause rust defers some language features ada has baked in to libraries

[timid quartz]

I think so yeah, also rustfmt likes to use more lines to be pretty in a lot of cases

[timid quartz]

Yeah

The rationale is a) not bloating the stdlib and b) not restricting things to the release pace of the stdlib (meaning old, bad interfaces get fossilized)

[spare quartz]

understandable

static ada programs tend to be quite large for what they do, where as

lemme go check your rust app

what

is this cause of the debug data? how do i turn that off

strip = "debuginfo"

[timid quartz]

cargo b --release

[spare quartz]

okay its still quite large...

i'll just chalk that up to the libraries having a lot of stuff in them

like tokio being built with full features

[timid quartz]

Tokio is

Yeah

[spare quartz]

i'd imagine a rust hello world program is smaller than an ada one though

[timid quartz]

Try it I’m in bed rn

I’ll also try to manually implement HTTP sometime soon so you can test

I just have the world’s busiest week

[spare quartz]

okay yeah somethings going on here that i can't see

13 MB for hello world doesn't sound right

[timid quartz]

Compile in release mode idk

[spare quartz]

thats what im doing

[timid quartz]

I might pull in a library for a proper thread pool tho

[spare quartz]

lemme tack on a few flags

there

-O3 with debug information (alr defaults)

[timid quartz]

Turn off debug info

Curious to see how much smaller it gets

[spare quartz]

with roughly the same arguments (except for "panic" since that'd require messing with gnat exception handling)

very minor

[timid quartz]

Hm do panic = "unwind" for Rust then

[spare quartz]

if i cheat and dynamically link the executable i instead get

[timid quartz]

DAM

that is cheating though

[timid quartz]

Try this

[spare quartz]

god this takes forever to compile

[timid quartz]

Nuh uh

[spare quartz]

323 kb

so about the same

[timid quartz]

neat

Now do Zig with .ReleaseSmall

jk

[spare quartz]

sure lemme install zig

[timid quartz]

I might write a thing in Zig for you to test though….

(Http thing)

[spare quartz]

should note dynamically linking will do some odd stuff to exception handling too

it won't inform you if you're missing a library either so it's a tradeoff

zig build (no parameters)

zig build (releasesmall)

about 2 kb smaller than ada shared

[spare quartz]

secure random done

[spare quartz]

lot of code

[spare quartz]

To create additional motivation, the NSA incorporated various gamification features. For instance, analysts who were especially good at using XKeyscore could acquire "skilz" points and "unlock achievements."

[spare quartz]

longest function name so far

[timid quartz]

Me when I compromise /dev/random

Or use windows

[spare quartz]

Accounted for by using an imported function

[spare quartz]

Also… not my problem! If your random is compromised you have bigger issues

That’s more of a laziness thing though, I could’ve imported linux random(3)… but I didn’t!

Rn the OpenSSL client recognizes I’ve sent it a Server Hello

But… it errors soon after about how I sent the wrong cipher…

Oh yeah TODO : need to patch my byte flip library to flip oddly sized Modular types better

Rn it’s getting 0xFC0100 and transforming it into 0xFC01…

[timid quartz]

https://tenor.com/view/erm-what-the-flip-what-the-flip-gif-1686926213572154405

what the flip

not the gif I wanted to send

[spare quartz]

This mf couldnt even do simple unchecked conversions 🙏

[timid quartz]

die

[spare quartz]

THE CAT

Idiot

[timid quartz]

smh

unchecked conversions are cringe anyways

[spare quartz]

Nuh uh

Well

For everything outside of streams yeah

But using a String as a byte buffer and reinterpeting is so useful

[timid quartz]

Ada should just read the record correctly, shouldn't need to do what is basically a C typecast

[spare quartz]

Well

What if it’s a tagged record

[timid quartz]

oop 🤢

[spare quartz]

Orrrr in my case a weirdly sized modular type

Besides you don’t use Unchecked convs when reading off records

[timid quartz]

hm question, do I make a thread pool from scratch or just use a crate

[spare quartz]

Use a crate?

Rust threading looks weird

[timid quartz]

eh not really

but a crate will prob be quicker

[timid quartz]

@spare quartz are you fine if I tailor this server to specifically serve get requests for the root lmao

so like nothing other than GET / HTTP/1.1

[spare quartz]

Sure that’s fine

[timid quartz]

or do you want me to make it more general

ok

[spare quartz]

Oh yeah

TLS is looking less scarier the more I look at ada big numbers

[timid quartz]

Noo it must be more scary

[spare quartz]

It is quite unfortunate I have to staple “WARNING THIS UNIT IS SUSCEPTIBLE TO TIMINGS SIDE CHANNEL ATTACKS” on the top of it though

Specifically the X25565 generator

[spare quartz]

I’m pretty sure this can’t be exploited in the way it’s used though

but… I suck at cryptography

[timid quartz]

just make it take a fixed time

1. figure out how long it takes to generate number
2. wait enough time to make it take constant time
3. ez

[spare quartz]

Ohhh but I don’t wannnaaa become sudoooo

[timid quartz]

@spare quartz ok try this

[spare quartz]

Aera

I have unfortunate news

[timid quartz]

hm

[spare quartz]

I’m in bed!!!!

If you want you can clone the basic http server and run it yourself

[timid quartz]

dies

[spare quartz]

Go onto my profile and click the bread block icon

That’s where the stuff is

[timid quartz]

urgh it keeps erroring why

like it says it has read errors

[spare quartz]

Huh?

I need details

[spare quartz]

Oh that

Honestly I have no idea what causes those

[timid quartz]

interesting

[spare quartz]

I just ignore them

[timid quartz]

well

that was the server

so not as fast as ada maybe

well...less throughput

[timid quartz]

Idk I’m probably doing something wrong

https://github.com/Xudong-Huang/may_minihttp

This was able to have a massive throughput

[spare quartz]

Their parameters are different it serms

Seems

天ガイン
010 809D

[timid quartz]

what is that

[spare quartz]

Trying the iPhone OCR

[timid quartz]

o

ngl I can barely read that second kanji

[spare quartz]

Yeah it totally minced the first two kanji

Oh well

I copied it from the sun facing side too so it had even less to work off of

[timid quartz]

Oh it just says Touyoko

[spare quartz]

Oh

[timid quartz]

Literally “Touyoko Inn”

[spare quartz]

my .com!!!

[timid quartz]

yeah I used the website url to figure out what it said

and it just says the name

lmao

バカ猫

[lavish dove]

I mean you can start really simple lol

When I was younger I made a compiler that converted some linear thing to python

It's just taking text then figuring out what to do with it

Very time consuming 😭

[spare quartz]

Well I know at a high level it’s just take structure A, turn it into instructions but like

Doing it well is the hard part!!

If I ever do make one I’d start with Ada -> JVM Bytecode

since it’s actually detailed on what to do

[spare quartz]

Also wouldn’t this technically be a transpiler?

Compiler - translates to machine code
Transpiler - translates to another language

At least as far as I understand it

[lavish dove]

Wouldn't Machine code just be a language the cpu supports

Intrinsics of the human language

[timid quartz]

urghhh microsoft why did you omit the value 0x0010

[timid quartz]

@spare quartz

MZ SIG          OK
COFF SIG        OK

(SIGNATURE =>  17744,
 MACHINE_ID => AMD64,
 SECTION_COUNT =>  3,
 CREATED_AT =>  1729522237,
 SYMBOL_TABLE_PTR =>  0,
 SYMBOL_COUNT =>  0,
 OPTIONAL_HEADER_SIZE =>  240,
 CHARACTERISTICS => 
 (STRIPPED_RELOCATIONS => FALSE,
  IS_EXECUTABLE => TRUE,
  STRIPPED_LINE_NUMS => TRUE,
  STRIPPED_SYMBOLS => TRUE,
  AGGRESSIVE_WS_TRIM => FALSE,
  LARGE_ADDR_AWARE => TRUE,
  RESERVED => FALSE,
  BYTES_REVERSED_LO => FALSE,
  IS_32_BIT => FALSE,
  NO_DEBUG_INFO => TRUE,
  IS_REMOVABLE => FALSE,
  IS_NET => FALSE,
  IS_FS => FALSE,
  IS_DLL => FALSE,
  ONLY_UNIPROCESSOR => FALSE,
  BYTES_REVERSED_HI => FALSE))

(MAJOR_LINKER_VER =>  2,
 MINOR_LINKER_VER =>  43,
 CODE_SIZE =>  1568768,
 INIT_DATA_SIZE =>  45056,
 UNINIT_DATA_SIZE =>  3026944,
 ENTRY_POINT_ADDR =>  4596080,
 BASE_OF_CODE =>  3031040,
 PREFERRED_BASE =>  5368709120,
 SECTION_ALIGNMENT =>  4096,
 FILE_ALIGNMENT =>  512,
 MAJOR_OS_VER =>  4,
 MINOR_OS_VER =>  0,
 MAJOR_IMG_VER =>  0,
 MINOR_IMG_VER =>  0,
 MAJOR_SUBSYS_VER =>  5,
 MINOR_SUBSYS_VER =>  2,
 RESERVED =>  0,
 IMAGE_SIZE =>  4644864,
 HEADER_SIZE =>  512,
 CHECKSUM =>  1637883,
 SUBSYSTEM => WINDOWS_GUI,
 DLL_CHARACTERISTICS => 
 (RESERVED1 => FALSE,
  RESERVED2 => FALSE,
  RESERVED3 => FALSE,
  RESERVED4 => FALSE,
  EMPTY1 => FALSE,
  HIGH_ENTROPY_CAPABLE => TRUE,
  DYNAMIC_BASE => TRUE,
  FORCE_INTEGRITY => FALSE,
  NX_COMPAT => TRUE,
  DONT_ISOLATE => FALSE,
  NO_SEH => FALSE,
  DONT_BIND => FALSE,
  USE_APPCONTAINER => FALSE,
  IS_WDM_DRIVER => FALSE,
  HAS_CONTROL_FLOW_GUARD => FALSE,
  TERMINAL_SERVER_AWARE => FALSE),
 STACK_RESERVE_SIZE =>  2097152,
 STACK_COMMIT_SIZE =>  4096,
 HEAP_RESERVE_SIZE =>  1048576,
 HEAP_COMMIT_SIZE =>  4096,
 LOADER_FLAGS_RESERVED =>  0,
 RVA_NUM_AND_SIZE =>  16)

lovely

[lavish dove]

cuz its something deprecated

average microsoft docs

[lavish dove]

wheres the thing that gives the link date

[timid quartz]

they could at least put it in the docs...

[lavish dove]

I mean its microsoft their docs are siht

[timid quartz]

uhhh, CREATED_AT probably

[lavish dove]

that was separate

i think

ohhhh

yes

[spare quartz]

Well I suppose so

A langauge has semantics to it though

Machine code is simply just “instructions”

[spare quartz]

Neat

[timid quartz]

Tbh I may just make an overarching PE record and have both a way to read it in and pretty-print it

[spare quartz]

Well thankfully Ada 2022 added custom ‘Images

[timid quartz]

…how

[spare quartz]

https://learn.adacore.com/courses/whats-new-in-ada-2022/chapters/image_redefine.html

[timid quartz]

function My_Image (Self : whatever) return Unbounded_String;?

[spare quartz]

Hold on I need to find my glasses

The procedure call is defined by the aspect Put_Image

[timid quartz]

My glasses are on rn so I’m reading

Neat

You get this in Rust via impl Display for Type_Name

You can also #[derive(Debug)] if you’re lazy but you don’t get .to_string() from that (you only get the debug formatter, normally used by format!("{:?}", whatever))

You can also use :#? instead of :? for a “pretty print”

[spare quartz]

i need to see if Put_Image is inheritable

the date formatter for HTTP is quite reliant on numbers not having a space for the sign (negative years don't make sense)

hear me out

what if we just use telnet

[timid quartz]

What if we just use netcat/nmap

[spare quartz]

nmap is just a mapping tool thats boring

netcat maybe but thats just telnet

[timid quartz]

hm nvm nmap can’t

netcat can though

https://nmap.org/ncat/guide/ncat-chat.html

[spare quartz]

DTLS/UDP chat server

back to the days of using a BBS

[timid quartz]

it needs encryption so the nsa can’t snoop

[spare quartz]

thats what DTLS is for

[timid quartz]

Knowing you that’s probably been backdoored

[spare quartz]

❌

[timid quartz]

Also no way you put a matrix screenshot here

[spare quartz]

shh

[spare quartz]

https://fixupx.com/i/status/1859187828086776165

[spare quartz]

im sure none of these are important

[timid quartz]

[spare quartz]

hi barosto

i really more computers

all i have is

• mk2 (current pc, 64gb/3060/5500gt)
• optiplex from early 2010s
• gateway laptop from 1999 (nonfunctional)

if only mk1 wasn't confiscated..!

[timid quartz]

[spare quartz]

i hate how openssl isn't clear at all

[timid quartz]

[spare quartz]

die

this cant be where the error message was raied

[timid quartz]

hmmm

I can't really override Dos_Header's 'Read

[spare quartz]

wdym

[timid quartz]

so the signature needs to be procedure My_Read (S : Stream_Access; Item : out T); (where T is Dos_Header.Header but eh)

but I also need F : File_Type so I can Set_Index

   procedure Read_Header (S : Stream_Access; F : File_Type; H : out Header) is
   begin
      Unsigned_16'Read (S, H.Magic);
      Unsigned_16'Read (S, H.Last_Page_Size);
      Unsigned_16'Read (S, H.Page_Count);
      Unsigned_16'Read (S, H.Relocation_Entries);
      Unsigned_16'Read (S, H.Header_Size);
      Unsigned_16'Read (S, H.Min_Paras);
      Unsigned_16'Read (S, H.Max_Paras);
      Unsigned_16'Read (S, H.Init_Stack_Seg);
      Unsigned_16'Read (S, H.Init_Stack_Ptr);
      Unsigned_16'Read (S, H.Checksum);
      Unsigned_16'Read (S, H.Init_Instr_Ptr);
      Unsigned_16'Read (S, H.Init_Code_Seg);
      Unsigned_16'Read (S, H.Relocation_Table_Addr);
      Unsigned_16'Read (S, H.Overlay_Num);

      Set_Index (F, Index (F) + (2 * 4));

      Unsigned_16'Read (S, H.Oem_Id);
      Unsigned_16'Read (S, H.Oem_Info);

      Set_Index (F, Index (F) + (2 * 10));

      Unsigned_32'Read (S, H.PE_Sig_Offset);
   end Read_Header;

there are reserved words in those spots

and Set_Index is just easier

@spare quartz

[spare quartz]

why not reading onto a trash string

like
discard : String (1 .. 8)

[timid quartz]

mmm could

[spare quartz]

i thinnnkk

if you needed the file_type in a 'Read overriding procedure, you'd have two types like

type A is abstract tagged private;
type B is new A with record ... end record;
private
type A is record
  File : ...;
end record;

[timid quartz]

gross

tagged records

oop

yknow what's even more gross though

[spare quartz]

mfs when they antagonize a design instead of its use

[spare quartz]

what

[timid quartz]

my pc....is booted into arch 🤢

[spare quartz]

linux user.. its over..

[timid quartz]

it's so icky

nothing works right

[spare quartz]

hmm

do i do tls or sleep

hmm

[timid quartz]

sleep

[spare quartz]

nnnnnnnnnooooooooo