From 1a43ef5b83d7378faa653451ebc38a54725a7481 Mon Sep 17 00:00:00 2001
From: mvdbeek <m.vandenbeek@gmail.com>
Date: Thu, 18 Jan 2018 17:48:32 +0100
Subject: [PATCH 1/3] Work around (temporarily) wrong getsize() output

I'm not sure how/why this happens, but this can be reproduced by copying a file to a location in one process, while checking the size in another process:

Run this to copy a file indefinitely:

```
import os
import shutil

open('test').write("foo")
while True:
    shutil.copy('test', 'test.file')
    os.remove('test.file')
```

In the same directory run this

```
import os

count = 0
wrong_size = False
while True:
    try:
        size = os.path.getsize('test.file')
        if size == 0:
            wrong_size = True
            wrong_count = count
        elif wrong_size and os.path.getsize('test.file') != 0:
            # We have seen the wrong size being reported, but now it's correct again
            print("Got wrong file size at %dnd try, was fixed in %dnd try" % (wrong_count, count))
            break
    except Exception:
        pass
    count += 1
```

It'll (probably) fail like this:

```
Got wrong file size at 0nd try, was fixed in 1nd try
```
or more often:
```
Got wrong file size at 31nd try, was fixed in 32nd try
```

This seems to be fixed when running os.path.getsize() another time,
so we simply do that here.
---
 lib/galaxy/objectstore/__init__.py | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/lib/galaxy/objectstore/__init__.py b/lib/galaxy/objectstore/__init__.py
index dc0c25668b6..bb9504f08f0 100644
--- a/lib/galaxy/objectstore/__init__.py
+++ b/lib/galaxy/objectstore/__init__.py
@@ -10,6 +10,7 @@ import os
 import random
 import shutil
 import threading
+import time
 from xml.etree import ElementTree
 
 try:
@@ -367,7 +368,12 @@ class DiskObjectStore(ObjectStore):
         """
         if self.exists(obj, **kwargs):
             try:
-                return os.path.getsize(self.get_filename(obj, **kwargs))
+                size = os.path.getsize(self.get_filename(obj, **kwargs))
+                if size == 0:
+                    # May be legitimately 0, or there may be an issue with the FS / kernel, so we try again
+                    time.sleep(0.01)
+                    size = os.path.getsize(self.get_filename(obj, **kwargs))
+                return size
             except OSError:
                 return 0
         else:

From 16344a62d3a8f30f4c468b552ce49ca989a1f0b7 Mon Sep 17 00:00:00 2001
From: mvdbeek <m.vandenbeek@gmail.com>
Date: Thu, 18 Jan 2018 15:03:16 +0100
Subject: [PATCH 2/3] Alternative fix for missing metadata on datasets that
 report output size 0

---
 lib/galaxy/jobs/__init__.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/galaxy/jobs/__init__.py b/lib/galaxy/jobs/__init__.py
index de1c32538bc..1bb036495de 100644
--- a/lib/galaxy/jobs/__init__.py
+++ b/lib/galaxy/jobs/__init__.py
@@ -1251,7 +1251,7 @@ class JobWrapper(object, HasResourceParameters):
                 if job.states.ERROR == final_job_state:
                     dataset.blurb = "error"
                     dataset.mark_unhidden()
-                elif not purged and dataset.has_data():
+                elif not purged:
                     # If the tool was expected to set the extension, attempt to retrieve it
                     if dataset.ext == 'auto':
                         dataset.extension = context.get('ext', 'data')
@@ -1299,7 +1299,7 @@ class JobWrapper(object, HasResourceParameters):
                     except Exception:
                         dataset.set_peek()
                 else:
-                    # Handle an empty dataset.
+                    # Handle purged datasets.
                     dataset.blurb = "empty"
                     if dataset.ext == 'auto':
                         dataset.extension = context.get('ext', 'txt')

From 161467513b3f911cde15b890c404ed373f4571f4 Mon Sep 17 00:00:00 2001
From: mvdbeek <m.vandenbeek@gmail.com>
Date: Thu, 18 Jan 2018 20:22:07 +0100
Subject: [PATCH 3/3] Use a for loop when checking size in object store

and adjust `empty()` to use the new code. Thanks @nsoranzo.
---
 lib/galaxy/objectstore/__init__.py | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/lib/galaxy/objectstore/__init__.py b/lib/galaxy/objectstore/__init__.py
index bb9504f08f0..8144f8167df 100644
--- a/lib/galaxy/objectstore/__init__.py
+++ b/lib/galaxy/objectstore/__init__.py
@@ -359,7 +359,7 @@ class DiskObjectStore(ObjectStore):
 
     def empty(self, obj, **kwargs):
         """Override `ObjectStore`'s stub by checking file size on disk."""
-        return os.path.getsize(self.get_filename(obj, **kwargs)) == 0
+        return self.size(obj, **kwargs) == 0
 
     def size(self, obj, **kwargs):
         """Override `ObjectStore`'s stub by return file size on disk.
@@ -368,11 +368,13 @@ class DiskObjectStore(ObjectStore):
         """
         if self.exists(obj, **kwargs):
             try:
-                size = os.path.getsize(self.get_filename(obj, **kwargs))
-                if size == 0:
+                filepath = self.get_filename(obj, **kwargs)
+                for _ in range(0, 2):
+                    size = os.path.getsize(filepath)
+                    if size != 0:
+                        break
                     # May be legitimately 0, or there may be an issue with the FS / kernel, so we try again
                     time.sleep(0.01)
-                    size = os.path.getsize(self.get_filename(obj, **kwargs))
                 return size
             except OSError:
                 return 0