GNU bug report logs - #17663
24.4.50; arc-mode.el has unrar-free hard-coded

Previous Next

Full log

Message #25 received at 17663 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Nicolas Petton <nicolas <at> petton.fr>
Cc: 17663 <at> debbugs.gnu.org
Subject: Re: bug#17663: 24.4.50; arc-mode.el has unrar-free hard-coded
Date: Wed, 30 Sep 2015 05:51:32 -0400

> I just installed a patch in master to use unar and lsar.

Thanks.

> My patch removes the use of the "unrar-free" binary completely as 1) it
> is no longer available in Fedora, Ubuntu and Debian and 2)

According to https://packages.qa.debian.org/u/unrar-free.html,
it's available (again) in Debian.

> it has been reported not to work properly, but I guess that's up
> for discussion.

AFAIK it was working OK, except for the fact that it did not handle all
compression methods supported by the nonfree "unrar".

But since unar works arguably better, it's OK to drop
unrar-free, indeed.


        Stefan


PS: Btw, if you're up for some maintenance work on arc-mode.el, there's
a real problem to fix in it: opening a large .zip file (for example) can
take a *very* long time.  This time is spent in set-buffer-multibyte
(turning the buffer from unibyte to multibyte) which ends up doing
O(N^2) work.  We have a workaround for this problem in auto-coding-alist
where we arrange for those binary files to be loaded directly into
a multibyte buffer (thanks to the no-conversion-multibyte
coding-system), thus avoiding the nasty O(N^2) problem.  But a better
solution would be to do what we do in tar-mode.el: rather than keep the
archive's binary data and its user-displayed listing both in the same
buffer, we use two different buffers: a unibyte one containing the
binary data, and a multibyte one displaying the contents.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.