[ARVADOS] created: 1.3.0-390-g39a98eef8

Git user git at public.curoverse.com
Wed Feb 27 15:35:57 EST 2019


        at  39a98eef822a5f68a0666c33b97ff1a373bce10c (commit)


commit 39a98eef822a5f68a0666c33b97ff1a373bce10c
Author: Lucas Di Pentima <ldipentima at veritasgenetics.com>
Date:   Wed Feb 27 17:09:10 2019 -0300

    14884: Adds unicode-named file upload test.
    
    Arvados-DCO-1.1-Signed-off-by: Lucas Di Pentima <ldipentima at veritasgenetics.com>

diff --git a/sdk/python/tests/test_arv_put.py b/sdk/python/tests/test_arv_put.py
index a41184d10..01a52a5e6 100644
--- a/sdk/python/tests/test_arv_put.py
+++ b/sdk/python/tests/test_arv_put.py
@@ -1,3 +1,5 @@
+# -*- coding: utf-8 -*-
+
 # Copyright (C) The Arvados Authors. All rights reserved.
 #
 # SPDX-License-Identifier: Apache-2.0
@@ -1284,6 +1286,16 @@ class ArvPutIntegrationTest(run_test_server.TestCaseWithServers,
                          r'^\./%s.*:file2.txt' % os.path.basename(tmpdir))
         self.assertRegex(c['manifest_text'], r'^.*:file3.txt')
 
+    def test_unicode_on_filename(self):
+        tmpdir = self.make_tmpdir()
+        fname = u"i❤arvados.txt"
+        with open(os.path.join(tmpdir, fname), 'w') as f:
+            f.write("This is a unicode named file")
+        col = self.run_and_find_collection("", ['--no-progress', tmpdir])
+        self.assertNotEqual(None, col['uuid'])
+        c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
+        self.assertTrue(fname in c['manifest_text'], u"{} does not include {}".format(c['manifest_text'], fname))
+
     def test_silent_mode_no_errors(self):
         self.authorize_with('active')
         tmpdir = self.make_tmpdir()

commit 23cced39fdeeec170460fca4799a3fa54f8b8aae
Author: Tom Morris <tfmorris at veritasgenetics.com>
Date:   Tue Feb 26 11:58:04 2019 -0500

    14884: Fix some additional formats where Unicode can be used
    
    Including filenames, collection names, and directory path for cache
    
    Arvados-DCO-1.1-Signed-off-by: Tom Morris <tfmorris at veritasgenetics.com>

diff --git a/sdk/python/arvados/collection.py b/sdk/python/arvados/collection.py
index 7ad07cc60..667315d4f 100644
--- a/sdk/python/arvados/collection.py
+++ b/sdk/python/arvados/collection.py
@@ -283,7 +283,7 @@ class CollectionWriter(CollectionBase):
             streampath, filename = split(streampath)
         if self._last_open and not self._last_open.closed:
             raise errors.AssertionError(
-                "can't open '{}' when '{}' is still open".format(
+                u"can't open '{}' when '{}' is still open".format(
                     filename, self._last_open.name))
         if streampath != self.current_stream_name():
             self.start_new_stream(streampath)
@@ -461,22 +461,22 @@ class ResumableCollectionWriter(CollectionWriter):
                 writer._queued_file.seek(pos)
             except IOError as error:
                 raise errors.StaleWriterStateError(
-                    "failed to reopen active file {}: {}".format(path, error))
+                    u"failed to reopen active file {}: {}".format(path, error))
         return writer
 
     def check_dependencies(self):
         for path, orig_stat in listitems(self._dependencies):
             if not S_ISREG(orig_stat[ST_MODE]):
-                raise errors.StaleWriterStateError("{} not file".format(path))
+                raise errors.StaleWriterStateError(u"{} not file".format(path))
             try:
                 now_stat = tuple(os.stat(path))
             except OSError as error:
                 raise errors.StaleWriterStateError(
-                    "failed to stat {}: {}".format(path, error))
+                    u"failed to stat {}: {}".format(path, error))
             if ((not S_ISREG(now_stat[ST_MODE])) or
                 (orig_stat[ST_MTIME] != now_stat[ST_MTIME]) or
                 (orig_stat[ST_SIZE] != now_stat[ST_SIZE])):
-                raise errors.StaleWriterStateError("{} changed".format(path))
+                raise errors.StaleWriterStateError(u"{} changed".format(path))
 
     def dump_state(self, copy_func=lambda x: x):
         state = {attr: copy_func(getattr(self, attr))
@@ -492,7 +492,7 @@ class ResumableCollectionWriter(CollectionWriter):
         try:
             src_path = os.path.realpath(source)
         except Exception:
-            raise errors.AssertionError("{} not a file path".format(source))
+            raise errors.AssertionError(u"{} not a file path".format(source))
         try:
             path_stat = os.stat(src_path)
         except OSError as stat_error:
@@ -505,10 +505,10 @@ class ResumableCollectionWriter(CollectionWriter):
             self._dependencies[source] = tuple(fd_stat)
         elif path_stat is None:
             raise errors.AssertionError(
-                "could not stat {}: {}".format(source, stat_error))
+                u"could not stat {}: {}".format(source, stat_error))
         elif path_stat.st_ino != fd_stat.st_ino:
             raise errors.AssertionError(
-                "{} changed between open and stat calls".format(source))
+                u"{} changed between open and stat calls".format(source))
         else:
             self._dependencies[src_path] = tuple(fd_stat)
 
diff --git a/sdk/python/arvados/commands/put.py b/sdk/python/arvados/commands/put.py
index 61258632b..54fa356d3 100644
--- a/sdk/python/arvados/commands/put.py
+++ b/sdk/python/arvados/commands/put.py
@@ -360,7 +360,7 @@ class ResumeCache(object):
         try:
             fcntl.flock(fileobj, fcntl.LOCK_EX | fcntl.LOCK_NB)
         except IOError:
-            raise ResumeCacheConflict("{} locked".format(fileobj.name))
+            raise ResumeCacheConflict(u"{} locked".format(fileobj.name))
 
     def load(self):
         self.cache_file.seek(0)
@@ -502,7 +502,7 @@ class ArvPutUploadJob(object):
                     raise ArvPutUploadIsPending()
                 self._write_stdin(self.filename or 'stdin')
             elif not os.path.exists(path):
-                 raise PathDoesNotExistError("file or directory '{}' does not exist.".format(path))
+                 raise PathDoesNotExistError(u"file or directory '{}' does not exist.".format(path))
             elif os.path.isdir(path):
                 # Use absolute paths on cache index so CWD doesn't interfere
                 # with the caching logic.
@@ -742,7 +742,7 @@ class ArvPutUploadJob(object):
             elif file_in_local_collection.permission_expired():
                 # Permission token expired, re-upload file. This will change whenever
                 # we have a API for refreshing tokens.
-                self.logger.warning("Uploaded file '{}' access token expired, will re-upload it from scratch".format(filename))
+                self.logger.warning(u"Uploaded file '{}' access token expired, will re-upload it from scratch".format(filename))
                 should_upload = True
                 self._local_collection.remove(filename)
             elif cached_file_data['size'] == file_in_local_collection.size():
@@ -757,7 +757,7 @@ class ArvPutUploadJob(object):
                 # Inconsistent cache, re-upload the file
                 should_upload = True
                 self._local_collection.remove(filename)
-                self.logger.warning("Uploaded version of file '{}' is bigger than local version, will re-upload it from scratch.".format(source))
+                self.logger.warning(u"Uploaded version of file '{}' is bigger than local version, will re-upload it from scratch.".format(source))
         # Local file differs from cached data, re-upload it.
         else:
             if file_in_local_collection:
@@ -834,11 +834,11 @@ class ArvPutUploadJob(object):
         if self.use_cache:
             cache_filepath = self._get_cache_filepath()
             if self.resume and os.path.exists(cache_filepath):
-                self.logger.info("Resuming upload from cache file {}".format(cache_filepath))
+                self.logger.info(u"Resuming upload from cache file {}".format(cache_filepath))
                 self._cache_file = open(cache_filepath, 'a+')
             else:
                 # --no-resume means start with a empty cache file.
-                self.logger.info("Creating new cache file at {}".format(cache_filepath))
+                self.logger.info(u"Creating new cache file at {}".format(cache_filepath))
                 self._cache_file = open(cache_filepath, 'w+')
             self._cache_filename = self._cache_file.name
             self._lock_file(self._cache_file)
@@ -924,7 +924,7 @@ class ArvPutUploadJob(object):
         try:
             fcntl.flock(fileobj, fcntl.LOCK_EX | fcntl.LOCK_NB)
         except IOError:
-            raise ResumeCacheConflict("{} locked".format(fileobj.name))
+            raise ResumeCacheConflict(u"{} locked".format(fileobj.name))
 
     def _save_state(self):
         """
@@ -1234,9 +1234,9 @@ def main(arguments=None, stdout=sys.stdout, stderr=sys.stderr,
     else:
         try:
             if args.update_collection:
-                logger.info("Collection updated: '{}'".format(writer.collection_name()))
+                logger.info(u"Collection updated: '{}'".format(writer.collection_name()))
             else:
-                logger.info("Collection saved as '{}'".format(writer.collection_name()))
+                logger.info(u"Collection saved as '{}'".format(writer.collection_name()))
             if args.portable_data_hash:
                 output = writer.portable_data_hash()
             else:

commit 11103ee6c45acceb1ebdd02c4a190bbabbf8621c
Author: Tom Morris <tfmorris at veritasgenetics.com>
Date:   Mon Feb 25 16:29:39 2019 -0500

    14884: Allow non-ASCII filenames in manifests.
    
        Arvados-DCO-1.1-Signed-off-by: Tom Morris <tfmorris at veritasgenetics.com>

diff --git a/sdk/python/arvados/_normalize_stream.py b/sdk/python/arvados/_normalize_stream.py
index b579d41ed..485c757e7 100644
--- a/sdk/python/arvados/_normalize_stream.py
+++ b/sdk/python/arvados/_normalize_stream.py
@@ -55,13 +55,13 @@ def normalize_stream(stream_name, stream):
                 if streamoffset == current_span[1]:
                     current_span[1] += segment.segment_size
                 else:
-                    stream_tokens.append("{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
+                    stream_tokens.append(u"{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
                     current_span = [streamoffset, streamoffset + segment.segment_size]
 
         if current_span is not None:
-            stream_tokens.append("{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
+            stream_tokens.append(u"{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
 
         if not stream[streamfile]:
-            stream_tokens.append("0:0:{0}".format(fout))
+            stream_tokens.append(u"0:0:{0}".format(fout))
 
     return stream_tokens

-----------------------------------------------------------------------


hooks/post-receive
-- 




More information about the arvados-commits mailing list