builder: use the same tar extract logic for all artifacts

Before only the staging area had symbolic links handled, now
strata and systems will have the same logic.

4 files changed, 86 insertions, 95 deletions

diff --git a/morphlib/bins.py b/morphlib/bins.py
index 8fd01182..3947e28d 100644
--- a/morphlib/bins.py
+++ b/morphlib/bins.py
@@ -24,6 +24,7 @@ Binaries are chunks, strata, and system images.
 import logging
 import os
 import re
+import stat
 import tarfile
 
 
@@ -109,17 +110,87 @@ def unpack_binary_from_file(f, dirname): # pragma: no cover
     '''
 
     tf = tarfile.open(fileobj=f)
-    tf.extractall(path=dirname)
-    tf.close()
-
-
-def unpack_binary(filename, dirname, ex):
-    '''Unpack a binary into a directory.
+        
+    # This is evil, but necessary. For some reason Python's system
+    # call wrappers (os.mknod and such) do not (always?) set the
+    # filename attribute of the OSError exception they raise. We
+    # fix that by monkey patching the tf instance with wrappers
+    # for the relevant methods to add things. The wrapper further
+    # ignores EEXIST errors, since we do not (currently!) care about
+    # overwriting files.
     
-    The directory must exist already.
+    def follow_symlink(path): # pragma: no cover
+        try:
+            return os.stat(path)
+        except OSError:
+            return None
     
-    '''
+    def prepare_extract(tarinfo, targetpath): # pragma: no cover
+        '''Prepare to extract a tar file member onto targetpath?
+        
+        If the target already exist, and we can live with it or
+        remove it, we do so. Otherwise, raise an error.
+        
+        It's OK to extract if:
+
+        * the target does not exist
+        * the member is a directory a directory and the 
+          target is a directory or a symlink to a directory
+          (just extract, no need to remove)
+        * the member is not a directory, and the target is not a directory
+          or a symlink to a directory (remove target, then extract)
+        
+        '''
+
+        try:
+            existing = os.lstat(targetpath)
+        except OSError:
+            return True # target does not exist
+
+        if tarinfo.isdir():
+            if stat.S_ISDIR(existing.st_mode):
+                return True
+            elif stat.S_ISLNK(existing.st_mode):
+                st = follow_symlink(targetpath)
+                return st and stat.S_ISDIR(st.st_mode)
+        else:
+            if stat.S_ISDIR(existing.st_mode):
+                return False
+            elif stat.S_ISLNK(existing.st_mode):
+                st = follow_symlink(targetpath)
+                if st and not stat.S_ISDIR(st.st_mode):
+                    os.remove(targetpath)
+                    return True
+            else:
+                os.remove(targetpath)
+                return True
+        return False
+
+    def monkey_patcher(real):
+        def make_something(tarinfo, targetpath): # pragma: no cover
+            prepare_extract(tarinfo, targetpath)
+            try:
+                return real(tarinfo, targetpath)
+            except OSError, e:
+                if e.errno != errno.EEXIST:
+                    if e.filename is None:
+                        e.filename = targetpath
+                        raise e
+                    else:
+                        raise
+        return make_something
+
+    tf.makedir = monkey_patcher(tf.makedir)
+    tf.makefile = monkey_patcher(tf.makefile)
+    tf.makeunknown = monkey_patcher(tf.makeunknown)
+    tf.makefifo = monkey_patcher(tf.makefifo)
+    tf.makedev = monkey_patcher(tf.makedev)
+    tf.makelink = monkey_patcher(tf.makelink)
 
-    logging.debug('Unpacking %s into %s' % (filename, dirname))
-    ex.runv(['tar', '-C', dirname, '-xvhf', filename])
+    tf.extractall(path=dirname)
+    tf.close
 
+def unpack_binary(filename, dirname):
+    f = open(filename, "rb")
+    unpack_binary_from_file(f, dirname)
+    f.close()
diff --git a/morphlib/bins_tests.py b/morphlib/bins_tests.py
index 82ceb246..dd2fb886 100644
--- a/morphlib/bins_tests.py
+++ b/morphlib/bins_tests.py
@@ -113,7 +113,7 @@ class ChunkTests(BinsTest):
 
     def unpack_chunk(self):
         os.mkdir(self.unpacked)
-        morphlib.bins.unpack_binary(self.chunk_file, self.unpacked, self.ex)
+        morphlib.bins.unpack_binary(self.chunk_file, self.unpacked)
         
     def test_empties_everything(self):
         self.create_chunk(['.'])
@@ -158,7 +158,7 @@ class StratumTests(BinsTest):
         morphlib.bins.create_stratum(self.instdir, self.stratum_f, self.ex)
         self.stratum_f.flush()
         os.mkdir(self.unpacked)
-        morphlib.bins.unpack_binary(self.stratum_file, self.unpacked, self.ex)
+        morphlib.bins.unpack_binary(self.stratum_file, self.unpacked)
         self.assertEqual(self.recursive_lstat(self.instdir),
                          self.recursive_lstat(self.unpacked))
 
diff --git a/morphlib/builder2.py b/morphlib/builder2.py
index 2a7ca2d2..62e484f4 100644
--- a/morphlib/builder2.py
+++ b/morphlib/builder2.py
@@ -81,7 +81,7 @@ def check_overlap(artifact, constituents, lac): #pragma: no cover
         logging.warning('Overlaps in artifact %s detected' % artifact.name)
         for overlapping, files in sorted(overlaps.iteritems()):
             logging.warning('  Artifacts %s overlap with files:' %
-                ', '.join(a.name for a in sorted(overlapping))
+                ', '.join(sorted(a.name for a in overlapping))
             )
             for filename in sorted(files):
                 logging.warning('    %s' % filename)
diff --git a/morphlib/stagingarea.py b/morphlib/stagingarea.py
index 8c84ac36..4f6d76bf 100644
--- a/morphlib/stagingarea.py
+++ b/morphlib/stagingarea.py
@@ -18,8 +18,6 @@ import errno
 import logging
 import os
 import shutil
-import stat
-import tarfile
 
 import morphlib
 
@@ -91,86 +89,8 @@ class StagingArea(object):
         '''
 
         logging.debug('Installing artifact %s' % 
-                        getattr(handle, 'name', 'unknown name'))        
-        tf = tarfile.open(fileobj=handle)
-        
-        # This is evil, but necessary. For some reason Python's system
-        # call wrappers (os.mknod and such) do not (always?) set the
-        # filename attribute of the OSError exception they raise. We
-        # fix that by monkey patching the tf instance with wrappers
-        # for the relevant methods to add things. The wrapper further
-        # ignores EEXIST errors, since we do not (currently!) care about
-        # overwriting files.
-        
-        def follow_symlink(path): # pragma: no cover
-            try:
-                return os.stat(path)
-            except OSError:
-                return None
-        
-        def prepare_extract(tarinfo, targetpath): # pragma: no cover
-            '''Prepare to extract a tar file member onto targetpath?
-            
-            If the target already exist, and we can live with it or
-            remove it, we do so. Otherwise, raise an error.
-            
-            It's OK to extract if:
-
-            * the target does not exist
-            * the member is a directory a directory and the 
-              target is a directory or a symlink to a directory
-              (just extract, no need to remove)
-            * the member is not a directory, and the target is not a directory
-              or a symlink to a directory (remove target, then extract)
-            
-            '''
-
-            try:
-                existing = os.lstat(targetpath)
-            except OSError:
-                return True # target does not exist
-
-            if tarinfo.isdir():
-                if stat.S_ISDIR(existing.st_mode):
-                    return True
-                elif stat.S_ISLNK(existing.st_mode):
-                    st = follow_symlink(targetpath)
-                    return st and stat.S_ISDIR(st.st_mode)
-            else:
-                if stat.S_ISDIR(existing.st_mode):
-                    return False
-                elif stat.S_ISLNK(existing.st_mode):
-                    st = follow_symlink(targetpath)
-                    if st and not stat.S_ISDIR(st.st_mode):
-                        os.remove(targetpath)
-                        return True
-                else:
-                    os.remove(targetpath)
-                    return True
-            return False
-
-        def monkey_patcher(real):
-            def make_something(tarinfo, targetpath): # pragma: no cover
-                prepare_extract(tarinfo, targetpath)
-                try:
-                    return real(tarinfo, targetpath)
-                except OSError, e:
-                    if e.errno != errno.EEXIST:
-                        if e.filename is None:
-                            e.filename = targetpath
-                            raise e
-                        else:
-                            raise
-            return make_something
-
-        tf.makedir = monkey_patcher(tf.makedir)
-        tf.makefile = monkey_patcher(tf.makefile)
-        tf.makeunknown = monkey_patcher(tf.makeunknown)
-        tf.makefifo = monkey_patcher(tf.makefifo)
-        tf.makedev = monkey_patcher(tf.makedev)
-        tf.makelink = monkey_patcher(tf.makelink)
-
-        tf.extractall(path=self.dirname)
+                        getattr(handle, 'name', 'unknown name'))
+        morphlib.bins.unpack_binary_from_file(handle, self.dirname)
 
     def remove(self):
         '''Remove the entire staging area.