I noticed today that doing the following

    write %file read http://foo.com/bar/baz

has the unfortunate result of (apparently) double buffering the entire
contents of baz in memory during the whole process.  Big problem if
you're fetching, as I was, a massive file...

Is there some other mechanism for doing this that would in essence
stream the bits from pipe to disk w/o buffering the whole thing even
once, much less twice?

jb


Reply via email to