From 198eefcf39aa3f07e2f676bdf35ccd64d90e66c9 Mon Sep 17 00:00:00 2001
From: shreyas-goenka <88374338+shreyas-goenka@users.noreply.github.com>
Date: Thu, 22 Dec 2022 11:35:32 +0100
Subject: [PATCH] Fix folder syncing on windows (#149)

Tested by running the unit and integration tests locally

Tested manually on windows

Screenshot from windows sync logs indicating that the correct slashed
for paths were used:
<img width="623" alt="Screenshot 2022-12-21 at 9 09 13 PM"
src="https://user-images.githubusercontent.com/88374338/208943937-146670b2-1afd-4e0b-8f4e-6091c8c7e17a.png">

@pietern with this the state machine for syncing becomes slightly more
complicated, indicating a stronger need for a tree based approach herre

Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
---
 cmd/sync/snapshot.go      | 15 +++++++++++----
 cmd/sync/snapshot_test.go | 34 ++++++++++++++++++++++++++++++++++
 2 files changed, 45 insertions(+), 4 deletions(-)

diff --git a/cmd/sync/snapshot.go b/cmd/sync/snapshot.go
index 9ab3f445..54a7e29a 100644
--- a/cmd/sync/snapshot.go
+++ b/cmd/sync/snapshot.go
@@ -226,17 +226,24 @@ func (s *Snapshot) diff(all []git.File) (change diff, err error) {
 		lastSeenModified, seen := lastModifiedTimes[f.Relative]
 
 		if !seen || modified.After(lastSeenModified) {
-			change.put = append(change.put, f.Relative)
 			lastModifiedTimes[f.Relative] = modified
 
-			// Keep track of remote names of local files so we can delete them later
+			// change separators to '/' for file paths in remote store
+			unixFileName := filepath.ToSlash(f.Relative)
+
+			// put file in databricks workspace
+			change.put = append(change.put, unixFileName)
+
+			// get file metadata about whether it's a notebook
 			isNotebook, typeOfNotebook, err := getNotebookDetails(f.Absolute)
 			if err != nil {
 				return change, err
 			}
-			remoteName := f.Relative
+
+			// strip `.py` for python notebooks
+			remoteName := unixFileName
 			if isNotebook && typeOfNotebook == "PYTHON" {
-				remoteName = strings.TrimSuffix(f.Relative, `.py`)
+				remoteName = strings.TrimSuffix(remoteName, `.py`)
 			}
 
 			// If the remote handle of a file changes, we want to delete the old
diff --git a/cmd/sync/snapshot_test.go b/cmd/sync/snapshot_test.go
index 54da750b..1fedb9fe 100644
--- a/cmd/sync/snapshot_test.go
+++ b/cmd/sync/snapshot_test.go
@@ -136,6 +136,40 @@ func TestDiff(t *testing.T) {
 	assert.Equal(t, map[string]string{"world.txt": "world.txt"}, state.RemoteToLocalNames)
 }
 
+func TestFolderDiff(t *testing.T) {
+	// Create temp project dir
+	projectDir := t.TempDir()
+	fileSet := git.NewFileSet(projectDir)
+	state := Snapshot{
+		LastUpdatedTimes:   make(map[string]time.Time),
+		LocalToRemoteNames: make(map[string]string),
+		RemoteToLocalNames: make(map[string]string),
+	}
+
+	err := os.Mkdir(filepath.Join(projectDir, "foo"), os.ModePerm)
+	assert.NoError(t, err)
+	f1 := createFile(t, filepath.Join(projectDir, "foo", "bar.py"))
+	defer f1.close(t)
+	f1.overwrite(t, "# Databricks notebook source\nprint(\"abc\")")
+
+	files, err := fileSet.All()
+	assert.NoError(t, err)
+	change, err := state.diff(files)
+	assert.NoError(t, err)
+	assert.Len(t, change.delete, 0)
+	assert.Len(t, change.put, 1)
+	assert.Contains(t, change.put, "foo/bar.py")
+
+	f1.remove(t)
+	files, err = fileSet.All()
+	assert.NoError(t, err)
+	change, err = state.diff(files)
+	assert.NoError(t, err)
+	assert.Len(t, change.delete, 1)
+	assert.Len(t, change.put, 0)
+	assert.Contains(t, change.delete, "foo/bar")
+}
+
 func TestPythonNotebookDiff(t *testing.T) {
 	// Create temp project dir
 	projectDir := t.TempDir()