Add support for smudge filters

If defined in .gitattributes call smudge filter during checkout.

To support checkout where current HEAD,index do not contain attributes
we need to also consider attributes from the tree we checkout. Therefore
CanonicalTreeParser has to learn how to provide attributes.

Change-Id: I168fdb81a8e1a9f991587b3e95a36550ea845f0a
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>

2 files changed, 126 insertions, 5 deletions

diff --git a/org.eclipse.jgit.test/tst/org/eclipse/jgit/api/CheckoutCommandTest.java b/org.eclipse.jgit.test/tst/org/eclipse/jgit/api/CheckoutCommandTest.java
index 82d509f6f3..4bfb128cbc 100644
--- a/org.eclipse.jgit.test/tst/org/eclipse/jgit/api/CheckoutCommandTest.java
+++ b/org.eclipse.jgit.test/tst/org/eclipse/jgit/api/CheckoutCommandTest.java
@@ -72,12 +72,14 @@ import org.eclipse.jgit.api.errors.RefNotFoundException;
 import org.eclipse.jgit.api.errors.TransportException;
 import org.eclipse.jgit.dircache.DirCache;
 import org.eclipse.jgit.dircache.DirCacheEntry;
+import org.eclipse.jgit.junit.JGitTestUtil;
 import org.eclipse.jgit.junit.RepositoryTestCase;
 import org.eclipse.jgit.lib.ConfigConstants;
 import org.eclipse.jgit.lib.Constants;
 import org.eclipse.jgit.lib.Ref;
 import org.eclipse.jgit.lib.RefUpdate;
 import org.eclipse.jgit.lib.Repository;
+import org.eclipse.jgit.lib.Sets;
 import org.eclipse.jgit.lib.StoredConfig;
 import org.eclipse.jgit.revwalk.RevCommit;
 import org.eclipse.jgit.storage.file.FileBasedConfig;
@@ -86,6 +88,7 @@ import org.eclipse.jgit.transport.RemoteConfig;
 import org.eclipse.jgit.transport.URIish;
 import org.eclipse.jgit.util.FileUtils;
 import org.junit.Before;
+import org.junit.Ignore;
 import org.junit.Test;
 
 public class CheckoutCommandTest extends RepositoryTestCase {
@@ -556,4 +559,125 @@ public class CheckoutCommandTest extends RepositoryTestCase {
 		}
 		org.junit.Assume.assumeTrue(foundUnsmudged);
 	}
+
+	@Test
+	public void testSmudgeFilter_modifyExisting() throws IOException, GitAPIException {
+		File script = writeTempFile("sed s/o/e/g");
+		StoredConfig config = git.getRepository().getConfig();
+		config.setString("filter", "tstFilter", "smudge",
+				"sh " + slashify(script.getPath()));
+		config.save();
+
+		writeTrashFile(".gitattributes", "*.txt filter=tstFilter");
+		git.add().addFilepattern(".gitattributes").call();
+		git.commit().setMessage("add filter").call();
+
+		writeTrashFile("src/a.tmp", "x");
+		// Caution: we need a trailing '\n' since sed on mac always appends
+		// linefeeds if missing
+		writeTrashFile("src/a.txt", "x\n");
+		git.add().addFilepattern("src/a.tmp").addFilepattern("src/a.txt")
+				.call();
+		RevCommit content1 = git.commit().setMessage("add content").call();
+
+		writeTrashFile("src/a.tmp", "foo");
+		writeTrashFile("src/a.txt", "foo\n");
+		git.add().addFilepattern("src/a.tmp").addFilepattern("src/a.txt")
+				.call();
+		RevCommit content2 = git.commit().setMessage("changed content").call();
+
+		git.checkout().setName(content1.getName()).call();
+		git.checkout().setName(content2.getName()).call();
+
+		assertEquals(
+				"[.gitattributes, mode:100644, content:*.txt filter=tstFilter][Test.txt, mode:100644, content:Some change][src/a.tmp, mode:100644, content:foo][src/a.txt, mode:100644, content:foo\n]",
+				indexState(CONTENT));
+		assertEquals(Sets.of("src/a.txt"), git.status().call().getModified());
+		assertEquals("foo", read("src/a.tmp"));
+		assertEquals("fee\n", read("src/a.txt"));
+	}
+
+	@Test
+	public void testSmudgeFilter_createNew()
+			throws IOException, GitAPIException {
+		File script = writeTempFile("sed s/o/e/g");
+		StoredConfig config = git.getRepository().getConfig();
+		config.setString("filter", "tstFilter", "smudge",
+				"sh " + slashify(script.getPath()));
+		config.save();
+
+		writeTrashFile("foo", "foo");
+		git.add().addFilepattern("foo").call();
+		RevCommit initial = git.commit().setMessage("initial").call();
+
+		writeTrashFile(".gitattributes", "*.txt filter=tstFilter");
+		git.add().addFilepattern(".gitattributes").call();
+		git.commit().setMessage("add filter").call();
+
+		writeTrashFile("src/a.tmp", "foo");
+		// Caution: we need a trailing '\n' since sed on mac always appends
+		// linefeeds if missing
+		writeTrashFile("src/a.txt", "foo\n");
+		git.add().addFilepattern("src/a.tmp").addFilepattern("src/a.txt")
+				.call();
+		RevCommit content = git.commit().setMessage("added content").call();
+
+		git.checkout().setName(initial.getName()).call();
+		git.checkout().setName(content.getName()).call();
+
+		assertEquals(
+				"[.gitattributes, mode:100644, content:*.txt filter=tstFilter][Test.txt, mode:100644, content:Some change][foo, mode:100644, content:foo][src/a.tmp, mode:100644, content:foo][src/a.txt, mode:100644, content:foo\n]",
+				indexState(CONTENT));
+		assertEquals("foo", read("src/a.tmp"));
+		assertEquals("fee\n", read("src/a.txt"));
+	}
+
+	@Test
+	@Ignore
+	public void testSmudgeAndClean() throws IOException, GitAPIException {
+		// @TODO: fix this test
+		File clean_filter = writeTempFile("sed s/V1/@version/g -");
+		File smudge_filter = writeTempFile("sed s/@version/V1/g -");
+
+		Git git = new Git(db);
+		StoredConfig config = git.getRepository().getConfig();
+		config.setString("filter", "tstFilter", "smudge",
+				"sh " + slashify(smudge_filter.getPath()));
+		config.setString("filter", "tstFilter", "clean",
+				"sh " + slashify(clean_filter.getPath()));
+		config.save();
+		writeTrashFile(".gitattributes", "*.txt filter=tstFilter");
+		git.add().addFilepattern(".gitattributes").call();
+		git.commit().setMessage("add attributes").call();
+
+		writeTrashFile("filterTest.txt", "hello world, V1");
+		git.add().addFilepattern("filterTest.txt").call();
+		git.commit().setMessage("add filterText.txt").call();
+		assertEquals(
+				"[.gitattributes, mode:100644, content:*.txt filter=tstFilter][Test.txt, mode:100644, content:Some other change][filterTest.txt, mode:100644, content:hello world, @version]",
+				indexState(CONTENT));
+
+		git.checkout().setCreateBranch(true).setName("test2").call();
+		writeTrashFile("filterTest.txt", "bon giorno world, V1");
+		git.add().addFilepattern("filterTest.txt").call();
+		git.commit().setMessage("modified filterText.txt").call();
+
+		assertTrue(git.status().call().isClean());
+		assertEquals(
+				"[.gitattributes, mode:100644, content:*.txt filter=tstFilter][Test.txt, mode:100644, content:Some other change][filterTest.txt, mode:100644, content:bon giorno world, @version]",
+				indexState(CONTENT));
+
+		git.checkout().setName("refs/heads/test").call();
+		assertTrue(git.status().call().isClean());
+		assertEquals(
+				"[.gitattributes, mode:100644, content:*.txt filter=tstFilter][Test.txt, mode:100644, content:Some other change][filterTest.txt, mode:100644, content:hello world, @version]",
+				indexState(CONTENT));
+		assertEquals("hello world, V1", read("filterTest.txt"));
+	}
+
+	private File writeTempFile(String body) throws IOException {
+		File f = File.createTempFile("AddCommandTest_", "");
+		JGitTestUtil.write(f, body);
+		return f;
+	}
 }
diff --git a/org.eclipse.jgit.test/tst/org/eclipse/jgit/lib/DirCacheCheckoutTest.java b/org.eclipse.jgit.test/tst/org/eclipse/jgit/lib/DirCacheCheckoutTest.java
index 51a6b5a8f5..d768e0fa0b 100644
--- a/org.eclipse.jgit.test/tst/org/eclipse/jgit/lib/DirCacheCheckoutTest.java
+++ b/org.eclipse.jgit.test/tst/org/eclipse/jgit/lib/DirCacheCheckoutTest.java
@@ -113,7 +113,7 @@ public class DirCacheCheckoutTest extends RepositoryTestCase {
 		return dco.getRemoved();
 	}
 
-	private Map<String, ObjectId> getUpdated() {
+	private Map<String, String> getUpdated() {
 		return dco.getUpdated();
 	}
 
@@ -268,8 +268,6 @@ public class DirCacheCheckoutTest extends RepositoryTestCase {
 	@Test
 	public void testRules1thru3_NoIndexEntry() throws IOException {
 		ObjectId head = buildTree(mk("foo"));
-		TreeWalk tw = TreeWalk.forPath(db, "foo", head);
-		ObjectId objectId = tw.getObjectId(0);
 		ObjectId merge = db.newObjectInserter().insert(Constants.OBJ_TREE,
 				new byte[0]);
 
@@ -279,10 +277,9 @@ public class DirCacheCheckoutTest extends RepositoryTestCase {
 
 		prescanTwoTrees(merge, head);
 
-		assertEquals(objectId, getUpdated().get("foo"));
+		assertTrue(getUpdated().containsKey("foo"));
 
 		merge = buildTree(mkmap("foo", "a"));
-		tw = TreeWalk.forPath(db, "foo", merge);
 
 		prescanTwoTrees(head, merge);