dumps

How to create the test SVN Dump Files

Many unit tests are based on the files in this directory.

The Makefile calls scripts to create svn dump files. This is useful in order to understand how the files were created. If a script does not exist, then the commands to create that dump should be listed below in this file.

Slowly, we should convert all of the instructions below into scripts.

PREAMBLE

SETUP CODE

Most of these commits use the same SETUP CODE, shown below:

mkdir -p svn-test/repos/testrepo
mkdir -p svn-test/checkout
cd svn-test/repos
svnadmin create testrepo
cd ../checkout
bash -c 'svn checkout "file:///$(dirname $(pwd))/repos/testrepo" testrepo'
cd testrepo

EXPORT CODE

They also use the same EXPORT CODE, shown below:

cd ../../repos
svnadmin dump testrepo > output.dump

Replace output.dump with the name of the dump file you're creating.

add_file.dump

<SETUP CODE>
echo "this is a test file" > README.txt
svn add README.txt
svn commit -m "Committed README.txt"
<EXPORT CODE>

add_file_no_node_properties.dump

Same as add_file.dump but without properties on the node.

different_node_order.dump

This file was created from a bigger dump of the AgreementMaker SVN repo. The main difference here is that the order of the "Node-kind" and "Node-path" are different than add_file.dump ("Node-kind" is before "Node-path").

different_node_order2.dump

This is a hand-hack of different_node_order.dump to add Text-content-sha1 before all of the other node headers. Moral of the story, the order of the headers should not matter.

binary_commit.dump

<SETUP CODE>
dd if=/dev/urandom of=file.bin bs=1K count=12
svn add file.bin
svn commit -m "Adding binary file."
<EXPORT CODE>

svn_rename_no_copy_hashes.dump

Same as svn_rename.dump but without the copy hashes.

svn_add_directory.dump

<SETUP CODE>
svn mkdir testdir
svn commit -m "Added a test directory."
cd testdir
echo "this is a test file" > README.txt
svn add README.txt
svn commit -m "Added a sample file."
cd ..
<EXPORT CODE>

svn_delete_file.dump

<SETUP CODE>
echo "this is a test file" > README.txt
svn add README.txt
svn commit -m "Added a sample file."
svn delete README.txt
echo "this is a test file again" > README.txt
svn add README.txt
svn commit -m "Added it again."
<EXPORT CODE>

svn_delete_with_add.dump

<SETUP CODE>
svn mkdir testdir
svn commit -m "Added a directory."
echo "this is a test file" > README.txt
svn rm testdir
svn add README.txt
svn commit -m "Added and deleted."
<EXPORT CODE>

svn_multi_dir_delete.dump

<SETUP CODE>
svn mkdir testdir1
svn mkdir testdir2
svn mkdir testdir3
svn commit -m "Added 3 dirs."
svn rm testdir1
svn rm testdir2
svn rm testdir3
svn commit -m "Deleted 3 dirs."

inner_dir.dump

<SETUP CODE>
mkdir -p test/innerdir
touch test/file1.txt test/file2.txt test/innerdir/file3.txt
svn add test
svn commit -m "Initial commit."
svn mv test test-renamed
svn commit -m "Changed dir name."
svn rm test-renamed/innerdir/file3.txt
svn commit -m "Go away inner file."
<EXPORT CODE>

add_edit_delete_add.dump

<SETUP CODE>
echo "this is a test file" > README.txt
svn add README.txt
svn commit -m "Initial commit."
echo "this is a test file edited" > README.txt
svn commit -m "Edited file."
svn rm README.txt
svn commit -m "Deleted file."
echo "this is a test file second try" > README.txt
svn add README.txt
svn commit -m "Added again."
<EXPORT CODE>

composite_commit.dump

<SETUP CODE>
mkdir -p d1/d2
echo "test" > d1/d2/readme2.txt
svn add d1
svn commit -m "Initial commit."
mkdir -p d1/d2/d3/d4
echo "test" > d1/d2/d3/d4/readme4.txt
svn add d1/d2/d3
svn commit -m "Additional sub directory."
svn cp d1@1 d1-copy
svn cp d1/d2/d3@2 d1-copy/d2
svn commit -m "Composite commit."
<EXPORT CODE>

undelete.dump

<SETUP CODE>
touch file1.txt
svn add file1.txt
svn commit -m "Initial commit."
svn rm file1.txt
svn commit -m "Deleted file1."
bash -c 'svn cp file:///$(dirname $(dirname $(pwd)))/repos/testrepo/file1.txt@1 file3.txt'
svn commit -m "Brought it back."
<EXPORT CODE>

undelete.invalid

A variation on undelete.dump but instead of undeleting the file from revision 1 we undelete it from revision 2, where it doesn't exist. It's an invalid svn dump, meant to verify that the validator throws an error.

simple_branch_and_merge_renamed.dump

<SETUP CODE>
mkdir -p trunk/renameddir branches
echo "this is a test file" > trunk/renameddir/README.txt
svn add trunk branches
svn commit -m "Initial commit."
svn copy trunk branches/mybranch
svn commit -m "Creating branch."
cd ..
bash -c 'svn checkout "file:///$(dirname $(pwd))/repos/testrepo/branches/mybranch" mybranch'
cd mybranch
echo "branch work" >> renameddir/README.txt
svn commit -m "Branch work."
cd ..
bash -c 'svn checkout "file:///$(dirname $(pwd))/repos/testrepo/trunk" trunk'
cd trunk
bash -c 'svn merge --reintegrate ^/branches/mybranch'
svn commit -m "Merge branch back into trunk."
bash -c  'svn delete ^/branches/mybranch -m "Removing branch."'
<EXPORT CODE>

property_change_on_file.dump

<SETUP CODE>
echo "test file" >> test.txt
svn add test.txt
svn commit -m "Adding test file."
svn ps someproperty value test.txt
svn commit -m "Setting property."
svn rm test.txt
svn commit -m "Delete file."
<EXPORT CODE>

extra_newline_in_log_message.dump

<SETUP CODE>
echo "test file" >> test.txt
svn add test.txt
svn commit -m 'Adding test file.
'
<EXPORT CODE>

property_change_on_root.dump

<SETUP CODE>
svn ps someproperty value .
svn commit -m "Setting property on root."
<EXPORT CODE>