Re: Schneier, Data and Goliath: no hope for privacy

On Thu, 20 Feb 2025, Salvador Mirzo wrote:

Now... do you seriously think anyone would ever be interested in that? ;)

>
LOL!  It turns out I'm *highly* interested in your leafnode.  Please,
can you put a package somewhere and let me look at it, try it out et
cetera?  I've been thinking about doing something like that myself.  I
can probably just live with your changes.

Of course!
Please grab a copy here: https://send.vis.ee/download/749384a5de2f4f33/#6ZgrL_j_qwmhqBNuXenoJA
It is very rough, not documented, but in the package you have the leafnode
source. My mod is in the file validatefqdn.c and if memory serves, I just
inverted a test for the fqdn from false to true, and that got rid of the check
for a fqdn. Before the mod leafnode would complain about the domain name and now
I think anything goes.
Also note that in order to run it in "server mode" you need to put the .socket
and .service files in /etc/systemd/system and start them.
In addition, there's my python script nntp2imap.py which takes care of copying
the downloaded news posts to my local Maildir directory "News" and names the
files and puts them in the right directories.
I found leafnode to be very simple to compile, and easy to get going.

I think I once packaged noffle (and not leafnode) for myself, but I
think I stopped using noffle for a news server, too.  I think I use
leafnode now, but I use it only as a local news server.  I'd like to
have an NNTP server that's simple to use, easily hosts local news server
and also easily peers with a USENET server for just a non-huge list of
groups.  I actually would like to write this server myself.

Why did you not use noffle in the end? Never heard of it before, but seems to be
similar perhaps to leafnode.

Let me share with you an upload script to https://0x0.st.  This is a no
nonsense temporary file uploader.  It works well as a paste bin, too.
The script below takes a series of files from the command line and
uploads each one to 0x0.st.  If you specify no file, then it will read
from the standard in.  The webserver at 0x0.st will print a URL for each
file uploaded, which is the address where you file is then stored.

Ahh... got it. Thank you very much!

--8<-------------------------------------------------------->8---
#!/bin/sh
args="$@"; test $# -lt 1 && args='-'
>
for f in $args; do
 curl --silent --show-error -X POST -F "file=@$f" https://0x0.st
done
--8<-------------------------------------------------------->8---
>
The next script uses upload to invoke scrot, which is a screen shot
taker.  It will take the URL printed by upload and store it in X's
primary selection area, which you can paste anywhere with your mouse or
with a keyboard shortcut.  (This is done by xsel.)
>
--8<-------------------------------------------------------->8---
#!/bin/sh
usage()
{
 printf 'usage: %s [options]\n' $(basename "$0")
}
test "$1" = '-h' && usage && exit 0
scrot --border -F- "$@" | upload | tr -d '\n' | xsel
--8<-------------------------------------------------------->8---
>
So if you call this script as /screenshot/ then all you need to do is to
invoke it and wait a moment then press the mouse middle button to paste
the URL somewhere.  The bug in it is that you don't know when /upload/
finishes.  It would be nice if somehow we could get a sign somewhere
that the URL is already in X's primary selection.
>
For the upload script, I would also like to write a GNU EMACS procedure
that takes a region of text and feeds upload's standard input.  That's a
nice way to share a bit of code or output or something.  I'm sure these
things exist already somewhere.  But it's nice to do them and perhaps
it's easier to do then to actually find where they are and then learn to
use them.
>