I wonder if it might make more sense to allow users to define such
commands in the config instead of hardcoding arbitrary messages.


Friendly greetings,
Leon Henrik Plickat

On Tue Sep 20, 2022 at 12:44 PM CEST, Ferass EL HAFIDI wrote:
> > A reverse engineered Java application supposedly provides support 
> > for completing this process on Linux. However, this approach comes 
> > with the typical bullshit of setting up a working Java environment, 
> > and, crucially, Xiaomi appears to have sabotaged this effort
>
> Xiaomi didn't sabotage the effort. They simply now require 2FA
> authentication, which XMT doesn't support yet. At least, if I recall
> correctly.

Is there any practical, sane reason to require two factor auth. just
to unlock the bootloader? I can't see any. In other words, this
definitely smells like making the process as annoying as possible
while still technically complying with EU law. So yes, this pretty

Thanks!


Friendly greetings,
Leon Henrik Plickat

Thanks!


Friendly greetings,
Leon Henrik Plickat

Yeah I'd be fine with that.  A few notes on that:

* It would need to be a new command to make the intent and
  functionality clear. Like "run-background-pipe-marks" or something
  like that. Obviously also with format string like the other exec'ing
  commands. Take a look at "run-background", you can probably copy /
  adapt most of it.
* If there are no marks, abort and show error message.
* While I like symmetry, I don't think you need to bother yet with a
  "forground" equivalent to the new command.


Friendly greetings,
Leon Henrik Plickat

I mean, I get why you came to that conclusion: The iterator is pretty
messy :D


Friendly greetings,
Leon Henrik Plickat

The general idea of nfm is to have absolutely no file management code
in it at all, especially super specific things like this.

I think this could be done with a script. There is one thing missing
for that right now: A placeholder in format strings to insert the
path of all marked files. If that existed, a script would work well
for this I suppose:

    [keybinds]
    M = run-background(move-all-to-cwd.sh %m);

You'd need to modify format() in InputBuffer.zig to support a format
specifier for that. If command length becomes an issue, maybe we also
could have an additional format specifier that makes the command run

Thanks for the patch!

I'll take a proper look tomorrow, but I have a few questions before that:

* While this makes iterating over all marked files more efficient,
  doesn't this make checking if a specific file is marked considerably
  less efficient? We do the latter one a lot more often (every render)
  than the first, so it's efficiency has a higher priority.
* Is the efficiency of iterating over marked files right now really
  bad enough to make this necessary?


Friendly greetings,
Leon Henrik Plickat

Thanks!


Friendly greetings,
Leon Henrik Plickat

I wonder if you have seen rnote[1] yet. I think it comes close: It
features an infinite canvas, supports hand writing, shapes, pictures
and PDF.

Although it has a few caveats. Its startup time is way to slow for my
taste. And while I appreciate UI design that is not a windows 98 clone,
I am not sure if I can get used to rnotes interface. Diagrams, plots,
(flow-)charts are also missing (graphviz, latex, etc might be a
workaround, but not a particularly nice one).


Friendly greetings,
Leon Henrik Plickat