Tabs. Tabs are terrible.

Problem: Tabs are not morally obligated to be glued to the edge of the text file. This happenstance is theoretically fine, however tabsize is NOT DEFINED, & thus varies from everything from nil to however high you can count.

Solution: make Tabs thus: '\t' == '\r\n' + Site Tab Size * (Next Byte + 1) where the Next Byte is 0x01->0x1F (except \t, which is treated as one (+1) for ""Backwards Compatibility"").

> Note: Emil does not morally support the conception of \r\n, just \n should be fine on all flatforms [Sic.], & having O_BINARY (Linux doesn't even larp this as a macro) on windows disgusts me.

magic value 0x1F becaues 0x20 is space.

so a file:

#include <stdio>\n
// the tabs are formatted at the beginning of the next line for clarity, they would be physically at the end of the prior line.\n
int main( void ) {
\tputs( "Hello World!" );
\tfor( int i = 0; i < 100; ++i )
\t0x01puts( "Tabs!\n" );
\treturn 0;\n
}\n

// assuming tabsize is 10
int main( void ) {
          puts( "Hello World!" );
          for( int i = 0; i < 100; ++i )
                    puts( "Tabs!\n" );
          return 0;
}

This system fundamentally prevents intermitten tab issues.

You could also define a universal, objectively correct standard space width of 3. (if you think this is wrong email me at My Email.)

For alignment, if we wish to solve that too, lets use Vertical Tab \v as it's generally useless.

\v0x01<->0x7E

Magic value 0x7E which is ~, which is not an escape (but very well could be just fine, but y'know) & prevents leaking into the evil UTF-8 bit.

-- \v0x10hello\v0x10world\v0x10now
-- \v0x10world\v0x10hello\v0x10surely
-- \v0x10give hope\v0x10for free\v0x10aligned

-- hello           world           now             
-- world           hello           surely          
-- give            hope for free   aligned         

We can think of this like putting 16 spaces, & then putting overwrite mode on. When we encounter another \v, we just go to the end of the prior \v.

We could apply the count-if-in-range idea from \t to \n if we would like, with the same rules. Deduplicating long stretches of \n sounds useful, however this is almost certainly counterproductive.

The only problem with this bloog entry is that it has not been emailed (read: emiled) back in time 50 years. If it is the case that \t & \v work this way, then I have been killed by the time police.

Emacs Email Configuration (For The Most Based Alive Only):

for GMail, get your App Password: Here. (this worked at the time of publication, if it is broken, fuck Google.)

This password will be put in .authinfo (which can be encrypted with gpg, see later of article.)

machine imap.gmail.com login you@gmail.com password XXXXXXXXXXXXXXXX port 993
machine smtp.gmail.com login you@gmail.com password XXXXXXXXXXXXXXXX port 465
machine mail.cock.li login you@cock.li password "some stupid password"

(setq message-send-mail-function 'smtpmail-send-it
      ;; These details may be unneeded:
      smtpmail-smtp-server "mail.cock.li"
      user-mail-address "you@cock.li"
      user-full-name "You Yoyo"
      smtpmail-smtp-service 465
      ;; --
      smtpmail-stream-type 'ssl
      message-from-style 'angles
      message-alternative-emails "you@cock\\.li\\|you@gmail\\.com"
      gnus-posting-styles
             '((".*gmail.*"
                (address "you@gmail.com")
                (name "You Yoyo"))
               ("you@cock\\.li"
                (address "you@cock.li")
                (name "You Yoyo")))
      gnus-select-method
             '(nnimap "cock"
                      (nnimap-address "mail.cock.li")
                      (nnimap-server-port 993)
                      (nnimap-stream ssl))
      gnus-secondary-select-methods
             '((nnimap "gmail"
                      (nnimap-address "imap.gmail.com")
                      (nnimap-server-port 993)
                      (nnimap-stream ssl))))
;; This line is here so the HTML doesn't freak out, & below is a useful hook you need:
(add-hook 'message-send-hook (lambda ()
  (let ((from (message-fetch-field "from")))
      (cond
       ((string-match "@cock\\.li" from)
        (setq smtpmail-smtp-server "mail.cock.li"
              smtpmail-smtp-service 465))
       ((string-match "@gmail\\.com" from)
        (setq smtpmail-smtp-server "smtp.gmail.com"
              smtpmail-smtp-service 465))))))

(add-hook 'dired-mode-hook 'turn-on-gnus-dired-mode)

For encryption ensure pinentry is stalled with USE=+emacs, & app-emacs/pinentry is istalled & enabled with (pinentry-start), additionally install/append the following two files: ~/.gnupg/gpg-agent.conf

allow-emacs-pinentry
allow-loopback-pinentry

~/.gnupg/gpg.conf

pinentry-mode loopback

Past that setup, ensure the daemon is running ps aux | grep gpg-agent and restart it via your DM or by restarting your account/system.

Run M-x epa-encrypt-file and enter ~/.authinfo just click/ENT on [OK] and enter your passphrase.

Run srm .authinfo after the .gpg file exists and you're done.

(setq auth-sources '("~/.authinfo.gpg"))

Try tramp out with C-x C-f /sudo:root@localhost:/root/ ENT, sync account details if you'd like.

> Note: sudo is retarded, use doas instead. Or just use su like a REAL human being.

If in doubt, ask computer.

last updated on Sun, 17 May 2026 11:56:07 +0000

©opyright Emil $CURRENT_YEAR CC-0 <~ Public Domain