Re: [ELPA] New package: dired-duplicates

[Eli Zaretskii]

> Date: Wed, 8 Nov 2023 21:29:46 +0100
> Cc: philipk@posteo.net, visuweshm@gmail.com, emacs-devel@gnu.org
> From: Harald Judt <h.judt@gmx.at>
> 
> > It is IMO okay to fail when sha256sum is not available and a file is
> > larger than the available VM, so Emacs runs out of memory, if this is
> > the situation that worries you.  But these cases should be relatively
> > rare, and so there's no reason to fail to have this feature in the
> > much more frequent case that the files are not as large as the
> > available VM.
> > 
> > If the file can be read by Emacs, even if it's large, then killing the
> > buffer after computing the hash should not have any adverse effects on
> > memory usage of that Emacs session.
> 
> I have started to implement the fallback to internal functions, here are my 
> results - it does even have size-limiting to avoid getting Emacs killed, 
> which 
> I managed to do trying with a big 4 GiB ISO file:
> 
> https://codeberg.org/hjudt/dired-duplicates/compare/main...fallback-to-internal-checksumming
> 
> Eli, is that how you imagined it? I would be glad if someone could give it a 
> quick review.

Yes, that was the idea I had, thanks.

The size limitation should have its default value dependent on whether
the build is a 32-bit (which we still support) or 64-bit.  You can
look at how we compute treesit-max-buffer-size, to figure out how to
express the conditions for the default value.