changegroup: avoid large copies - handle chunk headers separately rather than prepending them to (potentially large) chunks - break large chunks into 1M pieces for compression - don't prepend file metadata onto (potentially large) file data

#!/bin/sh

hg init
python -c 'print "confuse str.splitlines\nembedded\rnewline"' > a
hg ci -Ama -d '1 0'
echo clean diff >> a
hg ci -mb -d '2 0'
hg diff -r0 -r1