replay: fetch full revision at most once per run (issue252) Before this change, hgsubversion was fetching full revisions from the first revision the project was created to the first revision containing converted data. Unfortunately, some projects exhibits such spans longer than 500 revisions, during which hgsubversion was uselessly scanning the whole tree. The fix is not technically perfect, we could record somewhere that while no data was converted we scanned the project already, instead of scanning once at every hgsubversion run until a revision is converted. But it should be good enough unless someone runs hgsubversion once for every target revision. One repository exhibiting this behaviour: svn://svn.zankasoftware.com/zanka

#!/bin/sh
#
# Generate unorderedbranch.svndump
#

mkdir temp
cd temp

mkdir project-orig
cd project-orig
mkdir trunk
mkdir branches
cd ..

svnadmin create testrepo
svnurl=file://`pwd`/testrepo
svn import project-orig $svnurl -m "init project"

svn co $svnurl project
cd project/trunk
echo a > a
svn add a
svn ci -m "add a in trunk"
echo b > b
echo z > z
svn add b z
svn ci -m "add b and z in trunk"
svn up
cd ../branches
# Copy from trunk past revision. The converted used to take the last
# trunk revision as branch parent instead of the specified one.
svn cp -r 2 ../trunk branch
svn cp ../trunk/z branch
echo c > branch/c
svn add branch/c
svn ci -m 'branch and add c'
cd ../..

svnadmin dump testrepo > ../unorderedbranch.svndump