Re: Towards a large file mode, slowness

[Kevin Rodgers <kevin.d.rodgers@gmail.com>]

Andreas Röhler wrote:
> Hi,
> 
> dealing with large files, exceeding emacs capabilities,
> idea is to use `sed' to deliver subsequent chunks.
> 
> This seems possible if the files structur is line-separated, 
> if no single line exceeds max volume.
> 
> For example that way:
> 
> sed -n '111,130p' ein.txt
> 
> In the real program below the range is given by var
> `ltm-start-line' and
> 
> (+ ltm-start-line ltm-line-step)
> 
> If I call `sed' from within an Emacs-Shell, its quite
> fast regardless of the files volume.
> 
> Using `shell-command-to-string' or `call-process' as
> the program below, it's slowed down
> considerably.
> 
> Any ideas how to avoid this?

Read the thread re: "opening large files (few hundred meg)" from
January - February for some other ideas on how to implement this.

-- 
Kevin Rodgers
Denver, Colorado, USA