fix: index filenames with utf-8 characters on Windows

Fixes #184

src/components/exiftool/stream.js

@@ -1,4 +1,5 @@
 const childProcess = require('child_process')
+const trace = require('debug')('thumbsup:trace')
 const debug = require('debug')('thumbsup:debug')
 const error = require('debug')('thumbsup:error')
 const es = require('event-stream')
@@ -18,6 +19,8 @@ exports.parse = (rootFolder, filePaths) => {
     '%+.6f', // lat/long = float values
     '-struct', // preserve XMP structure
     '-json', // JSON output
+    '-charset', // allow UTF8 filenames
+    'filename=utf8', // allow UTF8 filenames
     '-@', // specify more arguments separately
     '-' // read arguments from standard in
   ]
@@ -25,7 +28,7 @@ exports.parse = (rootFolder, filePaths) => {
   // create a new <exiftool> child process
   const child = childProcess.spawn('exiftool', args, {
     cwd: rootFolder,
-    stdio: [ 'pipe', 'pipe', 'ignore' ]
+    stdio: [ 'pipe', 'pipe', 'pipe' ]
   })
   child.on('error', (err) => {
     error(`Error: please verify that <exiftool> is installed on your system`)
@@ -35,6 +38,10 @@ exports.parse = (rootFolder, filePaths) => {
     debug(`Exiftool exited with code ${code}`)
   })
 
+  child.stderr.on('data', chunk => {
+    trace('Exiftool output:', chunk.toString())
+  })
+
   // write all files to <stdin>
   // exiftool will only start processing after <stdin> is closed
   const allFiles = filePaths.join('\n')

test/components/exiftool/integration.spec.js

@@ -28,6 +28,22 @@ describe('exiftool', function () {
     })
   })
 
+  it('can process files with UTF8 names', (done) => {
+    // generate some photos in a temp folder
+    const structure = {
+      'photoà.jpg': fixtures.fromDisk('photo.jpg')
+    }
+    const tmpdir = fixtures.createTempStructure(structure)
+    const processed = []
+    const stream = exiftool.parse(tmpdir, Object.keys(structure))
+    stream.on('data', entry => {
+      processed.push(entry.SourceFile)
+    }).on('end', () => {
+      should(processed).eql(['photoà.jpg'])
+      done()
+    })
+  })
+
   it('can process badly encoded fields', (done) => {
     // here we test with an XMP file because it's easier to see what's wrong
     // but the problem will more likely be with a badly encoded XMP section inside a JPG file