svnwrap: use custom StringIO class in get_file() The wrappers were calling ra.get_file() with a cStringIO object. Empirically, svn 1.7.5 is writing 16kB blocks to the stream object, and cStringIO reallocates its internal buffer and doubles its size whenever it is filled. With large committed files this requires two large memory blocks at the same time. SimpleStringIO implements the mimimum StringIO interface used by ra.get_file() but instead stores all the blocks and "join" them at the end. It means more fragmentation but requires only one large block, without overallocation. Also, 16kB blocks should be friendly to most allocators. In practice, this simple change let me convert a revision containing multiple moderately large files, the largest being around 450MB, with a 32-bits Windows setup, python 2.7, swig svn 1.7.5, in stupid mode, while it was previously aborting with "not enough memory". The same revision still fails in replay mode.

#!/bin/sh
#
# Generate unrelatedbranch.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"
cd ../branches
# Create an unrelated branch with another file. It used to lead the converter
# to think branch1 was a copy of trunk, even without copy information.
mkdir branch1
echo b > branch1/b
svn add branch1
svn ci -m "add b in branch1"
# Make a real branch too for comparison
svn cp ../trunk branch2
echo b > branch2/b
svn add branch2/b
svn ci -m "add b to branch2"
# Add a file in the branch root for fun
echo c > c
svn add c
svn ci -m "add c in branches/"
# Even update it
echo c >> c
svn ci -m "change c"
cd ../..

svnadmin dump testrepo > ../unrelatedbranch.svndump