bug#19393: 25.0.50; Emacs cannot determine coding system of ISO-8859 encoded files

[Eli Zaretskii]

> From: Wolfgang Jenkner <wjenkner@inode.at>
> Cc: 19393@debbugs.gnu.org
> Date: Tue, 13 Jan 2015 18:12:54 +0100
> 
> > Is the problem with compressed files due to the fact that the size is
> > unknown in advance?
> 
> I only know that loading the compressed file from disk with the same
> coding system conversion as above takes just a few seconds, i.e., doing
> something like
> 
> C-x RET c l a t i n - 1 <return> C-x C-f m o v i e s . l i s t . g z 
> 
> is fast (enough).

Then it's probably not what I had in mind.

> > If so, perhaps enlarging by more than was
> > requested (e.g., twice as large) will alleviate the problem?
> 
> IIUC, this is your previous suggestion about improving
> insert-file-contents itself?

According to what you see, it sounds like determining the encoding is
what takes the time here, for some reason triggering massive
memmove's.