[darcs-users] Re: more feedback on binary file detection

[Mark Stosberg]

On 2004-12-09, Ralph Corderoy <ralph at inputplus.co.uk> wrote:
>
> Well, it could, but this is probably simpler.
>
>     tr -dc '\000\012\032' <file | tr '\000\010\032' '@\010Z' | grep -n .
>
> It prints the line number of any Ctrl-@ (ASCII NUL) or Ctrl-Z (ASCII
> SUB) characters in file.  Perhaps it can find a place in the Wiki?

I verified this works and will add it to the wiki. 

> Although, it would be nice if darcs had -D debug options that could be
> turned on for things like this.

Or '--no-auto-binary', which I sent a bug report about recently. 

> Did my faint memory of `once darcs thinks it's binary, it's binary
> forever' ring any bells?  

My recollection is different: I think darcs is happy to switch a file
back and forth to binary each time. In fact, I think that's what
happened to mine. It was considered binary for a bit, but is text again
now.

    Mark

-- 
http://mark.stosberg.com/