Merge branch 'liw/hardlink-staging-area-pre-rebase'

This is based on, and incorporates, Joe Burmeister's changes to fill
staging areas using hardlinking, for speed reasons. I have reformatted
the code a bit, and fixed it to pass the test suite, and fixed a bug
in how commands are run in the staging area chroot (preserving the
environment).

3 files changed, 89 insertions, 33 deletions

diff --git a/morphlib/builder2.py b/morphlib/builder2.py
index c7d25e1a..28a5e0ba 100644
--- a/morphlib/builder2.py
+++ b/morphlib/builder2.py
@@ -259,53 +259,31 @@ class ChunkBuilder(BuilderBase):
 
     def build_and_cache(self):  # pragma: no cover
         with self.build_watch('overall-build'):
-            mounted = self.do_mounts()
+            
+            builddir, destdir = \
+                self.staging_area.chroot_open(self.artifact.source)
             log_name = None
             try:
-                builddir = self.staging_area.builddir(self.artifact.source)
                 self.get_sources(builddir)
-                destdir = self.staging_area.destdir(self.artifact.source)
                 with self.local_artifact_cache.put_source_metadata(
                         self.artifact.source, self.artifact.cache_key,
                         'build-log') as log:
                     log_name = log.real_filename
                     self.run_commands(builddir, destdir, log)
             except:
-                self.do_unmounts(mounted)
+                self.staging_area.chroot_close()
                 if log_name:
                     with open(log_name) as f:
                         for line in f:
                             logging.error('OUTPUT FROM FAILED BUILD: %s' %
                                           line.rstrip('\n'))
                 raise
-            self.do_unmounts(mounted)
+            self.staging_area.chroot_close()
             built_artifacts = self.assemble_chunk_artifacts(destdir)
 
         self.save_build_times()
         return built_artifacts
 
-    to_mount = (
-        ('proc',    'proc',  'none'),
-        ('dev/shm', 'tmpfs', 'none'),
-    )
-
-    def do_mounts(self):  # pragma: no cover
-        mounted = []
-        if not self.setup_mounts:
-            return mounted
-        for mount_point, mount_type, source in ChunkBuilder.to_mount:
-            logging.debug('Mounting %s in staging area' % mount_point)
-            path = os.path.join(self.staging_area.dirname, mount_point)
-            if not os.path.exists(path):
-                os.makedirs(path)
-            self.app.runcmd(['mount', '-t', mount_type, source, path])
-            mounted.append(path)
-        return mounted
-
-    def do_unmounts(self, mounted):  # pragma: no cover
-        for path in mounted:
-            logging.debug('Unmounting %s in staging area' % path)
-            morphlib.fsutils.unmount(self.app.runcmd, path)
 
     def get_sources(self, srcdir):  # pragma: no cover
         '''Get sources from git to a source directory, for building.'''
diff --git a/morphlib/stagingarea.py b/morphlib/stagingarea.py
index a87b45c3..77eeab7d 100644
--- a/morphlib/stagingarea.py
+++ b/morphlib/stagingarea.py
@@ -40,6 +40,8 @@ class StagingArea(object):
         self._app = app
         self.dirname = dirname
         self.tempdir = tempdir
+        self.builddirname = None
+        self.destdirname = None        
 
     # Wrapper to be overridden by unit tests.
     def _mkdir(self, dirname):  # pragma: no cover
@@ -90,7 +92,21 @@ class StagingArea(object):
 
         logging.debug('Installing artifact %s' %
                       getattr(handle, 'name', 'unknown name'))
-        morphlib.bins.unpack_binary_from_file(handle, self.dirname)
+
+        unpacked_artifact = os.path.join(
+            self._app.settings['cachedir'],
+            'artifacts',
+            os.path.basename(handle.name) + '.d')
+        if not os.path.exists(unpacked_artifact):
+            self._mkdir(unpacked_artifact)
+            morphlib.bins.unpack_binary_from_file(
+                handle, unpacked_artifact + '/')
+            
+        if not os.path.exists(self.dirname):
+            self._mkdir(self.dirname)
+
+        self._app.runcmd(
+            ['cp', '-al', unpacked_artifact + '/.', self.dirname + '/.'])
 
     def remove(self):
         '''Remove the entire staging area.
@@ -103,14 +119,61 @@ class StagingArea(object):
 
         shutil.rmtree(self.dirname)
 
+    def chroot_open(self, source): # pragma: no cover
+        '''Setup staging area for use as a chroot.'''
+
+        assert self.builddirname == None and self.destdirname == None
+
+        builddir = self.builddir(source)
+        destdir = self.destdir(source)
+        self.builddirname = self.relative(builddir).lstrip('/')
+        self.destdirname = self.relative(destdir).lstrip('/')
+
+        for mount_point in ['proc','dev/shm']:
+            path = os.path.join(self.dirname, mount_point)
+            if not os.path.exists(path):
+                os.makedirs(path)
+
+        return builddir, destdir
+
+    def chroot_close(self): # pragma: no cover
+        '''Undo changes by chroot_open.
+        
+        This should be called after the staging area is no longer needed.
+        
+        '''
+
     def runcmd(self, argv, **kwargs):  # pragma: no cover
-        '''Run a command in a chroot in the staging area.'''
+        '''Run a command in the staging area.'''
+
         cwd = kwargs.get('cwd') or '/'
         if 'cwd' in kwargs:
             cwd = kwargs['cwd']
             del kwargs['cwd']
         else:
             cwd = '/'
-        real_argv = ['chroot', self.dirname, 'sh', '-c',
-                     'cd "$1" && shift && exec "$@"', '--', cwd] + argv
+
+        if self._app.settings['staging-chroot']:
+            entries = os.listdir(self.dirname)
+
+            friends = [self.builddirname, self.destdirname,
+                       'dev', 'proc', 'tmp']
+            for friend in friends:
+                if friend in friends:
+                    entries.remove(friend)
+
+            real_argv = ['linux-user-chroot']
+
+            for entry in entries:
+                real_argv += ['--mount-readonly',"/"+entry]
+            
+            real_argv += ['--mount-proc','/proc']
+            real_argv += ['--mount-bind','/dev/shm','/dev/shm']
+            real_argv += [self.dirname]
+        else:
+            real_argv = ['chroot', '/']
+
+        real_argv += ['sh', '-c', 'cd "$1" && shift && exec "$@"', '--', cwd]
+        real_argv += argv
+
         return self._app.runcmd(real_argv, **kwargs)
diff --git a/morphlib/stagingarea_tests.py b/morphlib/stagingarea_tests.py
index 3230b9e3..15131272 100644
--- a/morphlib/stagingarea_tests.py
+++ b/morphlib/stagingarea_tests.py
@@ -14,6 +14,7 @@
 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 
 
+import cliapp
 import os
 import shutil
 import tarfile
@@ -31,14 +32,28 @@ class FakeSource(object):
         }
 
 
+class FakeApplication(object):
+
+    def __init__(self, cachedir):
+        self.settings = {
+            'cachedir': cachedir,
+        }
+
+    def runcmd(self, *args, **kwargs):
+        cliapp.runcmd(*args, **kwargs)
+
+
 class StagingAreaTests(unittest.TestCase):
 
     def setUp(self):
         self.tempdir = tempfile.mkdtemp()
+        self.cachedir = os.path.join(self.tempdir, 'cachedir')
+        os.mkdir(self.cachedir)
+        os.mkdir(os.path.join(self.cachedir, 'artifacts'))
         self.staging = os.path.join(self.tempdir, 'staging')
         self.created_dirs = []
-        self.sa = morphlib.stagingarea.StagingArea(object(), self.staging,
-                                                   self.staging)
+        self.sa = morphlib.stagingarea.StagingArea(
+            FakeApplication(self.cachedir), self.staging, self.staging)
 
     def tearDown(self):
         shutil.rmtree(self.tempdir)