summaryrefslogtreecommitdiff
path: root/git/test/db/lib.py
diff options
context:
space:
mode:
Diffstat (limited to 'git/test/db/lib.py')
-rw-r--r--git/test/db/lib.py456
1 files changed, 228 insertions, 228 deletions
diff --git a/git/test/db/lib.py b/git/test/db/lib.py
index 2b3ddde5..df9fec76 100644
--- a/git/test/db/lib.py
+++ b/git/test/db/lib.py
@@ -4,21 +4,21 @@
# the New BSD License: http://www.opensource.org/licenses/bsd-license.php
"""Base classes for object db testing"""
from git.test.lib import (
- with_rw_directory,
- with_packs_rw,
- ZippedStoreShaWriter,
- fixture_path,
- TestBase,
- rorepo_dir,
- )
+ with_rw_directory,
+ with_packs_rw,
+ ZippedStoreShaWriter,
+ fixture_path,
+ TestBase,
+ rorepo_dir,
+ )
from git.stream import Sha1Writer
from git.base import (
- IStream,
- OStream,
- OInfo
- )
-
+ IStream,
+ OStream,
+ OInfo
+ )
+
from git.exc import BadObject
from git.typ import str_blob_type
@@ -28,220 +28,220 @@ from struct import pack
__all__ = ('TestDBBase', 'with_rw_directory', 'with_packs_rw', 'fixture_path')
-
+
class TestDBBase(TestBase):
- """Base Class providing default functionality to all tests such as:
-
- - Utility functions provided by the TestCase base of the unittest method such as::
- self.fail("todo")
- self.failUnlessRaises(...)
-
- - Class level repository which is considered read-only as it is shared among
- all test cases in your type.
- Access it using::
- self.rorepo # 'ro' stands for read-only
-
- The rorepo is in fact your current project's git repo. If you refer to specific
- shas for your objects, be sure you choose some that are part of the immutable portion
- of the project history ( to assure tests don't fail for others ).
-
- Derived types can override the default repository type to create a different
- read-only repo, allowing to test their specific type
- """
-
- # data
- two_lines = "1234\nhello world"
- all_data = (two_lines, )
-
- #{ Configuration
- # The repository type to instantiate. It takes at least a path to operate upon
- # during instantiation.
- RepoCls = None
-
- # if True, a read-only repo will be provided and RepoCls must be set.
- # Otherwise it may remain unset
- needs_ro_repo = True
- #} END configuration
-
- @classmethod
- def setUpAll(cls):
- """
- Dynamically add a read-only repository to our actual type. This way
- each test type has its own repository
- """
- if cls.needs_ro_repo:
- if cls is not TestDBBase:
- assert cls.RepoCls is not None, "RepoCls class member must be set in %s" % cls
- cls.rorepo = cls.RepoCls(rorepo_dir())
- #END handle rorepo
-
- def _assert_object_writing_simple(self, db):
- # write a bunch of objects and query their streams and info
- null_objs = db.size()
- ni = 250
- for i in xrange(ni):
- data = pack(">L", i)
- istream = IStream(str_blob_type, len(data), StringIO(data))
- new_istream = db.store(istream)
- assert new_istream is istream
- assert db.has_object(istream.binsha)
-
- info = db.info(istream.binsha)
- assert isinstance(info, OInfo)
- assert info.type == istream.type and info.size == istream.size
-
- stream = db.stream(istream.binsha)
- assert isinstance(stream, OStream)
- assert stream.binsha == info.binsha and stream.type == info.type
- assert stream.read() == data
- # END for each item
-
- assert db.size() == null_objs + ni
- shas = list(db.sha_iter())
- assert len(shas) == db.size()
- assert len(shas[0]) == 20
-
-
- def _assert_object_writing(self, db):
- """General tests to verify object writing, compatible to ObjectDBW
- :note: requires write access to the database"""
- # start in 'dry-run' mode, using a simple sha1 writer
- ostreams = (ZippedStoreShaWriter, None)
- for ostreamcls in ostreams:
- for data in self.all_data:
- dry_run = ostreamcls is not None
- ostream = None
- if ostreamcls is not None:
- ostream = ostreamcls()
- assert isinstance(ostream, Sha1Writer)
- # END create ostream
-
- prev_ostream = db.set_ostream(ostream)
- assert type(prev_ostream) in ostreams or prev_ostream in ostreams
-
- istream = IStream(str_blob_type, len(data), StringIO(data))
-
- # store returns same istream instance, with new sha set
- my_istream = db.store(istream)
- sha = istream.binsha
- assert my_istream is istream
- assert db.has_object(sha) != dry_run
- assert len(sha) == 20
-
- # verify data - the slow way, we want to run code
- if not dry_run:
- info = db.info(sha)
- assert str_blob_type == info.type
- assert info.size == len(data)
-
- ostream = db.stream(sha)
- assert ostream.read() == data
- assert ostream.type == str_blob_type
- assert ostream.size == len(data)
- else:
- self.failUnlessRaises(BadObject, db.info, sha)
- self.failUnlessRaises(BadObject, db.stream, sha)
-
- # DIRECT STREAM COPY
- # our data hase been written in object format to the StringIO
- # we pasesd as output stream. No physical database representation
- # was created.
- # Test direct stream copy of object streams, the result must be
- # identical to what we fed in
- ostream.seek(0)
- istream.stream = ostream
- assert istream.binsha is not None
- prev_sha = istream.binsha
-
- db.set_ostream(ZippedStoreShaWriter())
- db.store(istream)
- assert istream.binsha == prev_sha
- new_ostream = db.ostream()
-
- # note: only works as long our store write uses the same compression
- # level, which is zip_best
- assert ostream.getvalue() == new_ostream.getvalue()
- # END for each data set
- # END for each dry_run mode
-
- def _assert_object_writing_async(self, db):
- """Test generic object writing using asynchronous access"""
- ni = 5000
- def istream_generator(offset=0, ni=ni):
- for data_src in xrange(ni):
- data = str(data_src + offset)
- yield IStream(str_blob_type, len(data), StringIO(data))
- # END for each item
- # END generator utility
-
- # for now, we are very trusty here as we expect it to work if it worked
- # in the single-stream case
-
- # write objects
- reader = IteratorReader(istream_generator())
- istream_reader = db.store_async(reader)
- istreams = istream_reader.read() # read all
- assert istream_reader.task().error() is None
- assert len(istreams) == ni
-
- for stream in istreams:
- assert stream.error is None
- assert len(stream.binsha) == 20
- assert isinstance(stream, IStream)
- # END assert each stream
-
- # test has-object-async - we must have all previously added ones
- reader = IteratorReader( istream.binsha for istream in istreams )
- hasobject_reader = db.has_object_async(reader)
- count = 0
- for sha, has_object in hasobject_reader:
- assert has_object
- count += 1
- # END for each sha
- assert count == ni
-
- # read the objects we have just written
- reader = IteratorReader( istream.binsha for istream in istreams )
- ostream_reader = db.stream_async(reader)
-
- # read items individually to prevent hitting possible sys-limits
- count = 0
- for ostream in ostream_reader:
- assert isinstance(ostream, OStream)
- count += 1
- # END for each ostream
- assert ostream_reader.task().error() is None
- assert count == ni
-
- # get info about our items
- reader = IteratorReader( istream.binsha for istream in istreams )
- info_reader = db.info_async(reader)
-
- count = 0
- for oinfo in info_reader:
- assert isinstance(oinfo, OInfo)
- count += 1
- # END for each oinfo instance
- assert count == ni
-
-
- # combined read-write using a converter
- # add 2500 items, and obtain their output streams
- nni = 2500
- reader = IteratorReader(istream_generator(offset=ni, ni=nni))
- istream_to_sha = lambda istreams: [ istream.binsha for istream in istreams ]
-
- istream_reader = db.store_async(reader)
- istream_reader.set_post_cb(istream_to_sha)
-
- ostream_reader = db.stream_async(istream_reader)
-
- count = 0
- # read it individually, otherwise we might run into the ulimit
- for ostream in ostream_reader:
- assert isinstance(ostream, OStream)
- count += 1
- # END for each ostream
- assert count == nni
-
-
+ """Base Class providing default functionality to all tests such as:
+
+ - Utility functions provided by the TestCase base of the unittest method such as::
+ self.fail("todo")
+ self.failUnlessRaises(...)
+
+ - Class level repository which is considered read-only as it is shared among
+ all test cases in your type.
+ Access it using::
+ self.rorepo # 'ro' stands for read-only
+
+ The rorepo is in fact your current project's git repo. If you refer to specific
+ shas for your objects, be sure you choose some that are part of the immutable portion
+ of the project history ( to assure tests don't fail for others ).
+
+ Derived types can override the default repository type to create a different
+ read-only repo, allowing to test their specific type
+ """
+
+ # data
+ two_lines = "1234\nhello world"
+ all_data = (two_lines, )
+
+ #{ Configuration
+ # The repository type to instantiate. It takes at least a path to operate upon
+ # during instantiation.
+ RepoCls = None
+
+ # if True, a read-only repo will be provided and RepoCls must be set.
+ # Otherwise it may remain unset
+ needs_ro_repo = True
+ #} END configuration
+
+ @classmethod
+ def setUpAll(cls):
+ """
+ Dynamically add a read-only repository to our actual type. This way
+ each test type has its own repository
+ """
+ if cls.needs_ro_repo:
+ if cls is not TestDBBase:
+ assert cls.RepoCls is not None, "RepoCls class member must be set in %s" % cls
+ cls.rorepo = cls.RepoCls(rorepo_dir())
+ #END handle rorepo
+
+ def _assert_object_writing_simple(self, db):
+ # write a bunch of objects and query their streams and info
+ null_objs = db.size()
+ ni = 250
+ for i in xrange(ni):
+ data = pack(">L", i)
+ istream = IStream(str_blob_type, len(data), StringIO(data))
+ new_istream = db.store(istream)
+ assert new_istream is istream
+ assert db.has_object(istream.binsha)
+
+ info = db.info(istream.binsha)
+ assert isinstance(info, OInfo)
+ assert info.type == istream.type and info.size == istream.size
+
+ stream = db.stream(istream.binsha)
+ assert isinstance(stream, OStream)
+ assert stream.binsha == info.binsha and stream.type == info.type
+ assert stream.read() == data
+ # END for each item
+
+ assert db.size() == null_objs + ni
+ shas = list(db.sha_iter())
+ assert len(shas) == db.size()
+ assert len(shas[0]) == 20
+
+
+ def _assert_object_writing(self, db):
+ """General tests to verify object writing, compatible to ObjectDBW
+ :note: requires write access to the database"""
+ # start in 'dry-run' mode, using a simple sha1 writer
+ ostreams = (ZippedStoreShaWriter, None)
+ for ostreamcls in ostreams:
+ for data in self.all_data:
+ dry_run = ostreamcls is not None
+ ostream = None
+ if ostreamcls is not None:
+ ostream = ostreamcls()
+ assert isinstance(ostream, Sha1Writer)
+ # END create ostream
+
+ prev_ostream = db.set_ostream(ostream)
+ assert type(prev_ostream) in ostreams or prev_ostream in ostreams
+
+ istream = IStream(str_blob_type, len(data), StringIO(data))
+
+ # store returns same istream instance, with new sha set
+ my_istream = db.store(istream)
+ sha = istream.binsha
+ assert my_istream is istream
+ assert db.has_object(sha) != dry_run
+ assert len(sha) == 20
+
+ # verify data - the slow way, we want to run code
+ if not dry_run:
+ info = db.info(sha)
+ assert str_blob_type == info.type
+ assert info.size == len(data)
+
+ ostream = db.stream(sha)
+ assert ostream.read() == data
+ assert ostream.type == str_blob_type
+ assert ostream.size == len(data)
+ else:
+ self.failUnlessRaises(BadObject, db.info, sha)
+ self.failUnlessRaises(BadObject, db.stream, sha)
+
+ # DIRECT STREAM COPY
+ # our data hase been written in object format to the StringIO
+ # we pasesd as output stream. No physical database representation
+ # was created.
+ # Test direct stream copy of object streams, the result must be
+ # identical to what we fed in
+ ostream.seek(0)
+ istream.stream = ostream
+ assert istream.binsha is not None
+ prev_sha = istream.binsha
+
+ db.set_ostream(ZippedStoreShaWriter())
+ db.store(istream)
+ assert istream.binsha == prev_sha
+ new_ostream = db.ostream()
+
+ # note: only works as long our store write uses the same compression
+ # level, which is zip_best
+ assert ostream.getvalue() == new_ostream.getvalue()
+ # END for each data set
+ # END for each dry_run mode
+
+ def _assert_object_writing_async(self, db):
+ """Test generic object writing using asynchronous access"""
+ ni = 5000
+ def istream_generator(offset=0, ni=ni):
+ for data_src in xrange(ni):
+ data = str(data_src + offset)
+ yield IStream(str_blob_type, len(data), StringIO(data))
+ # END for each item
+ # END generator utility
+
+ # for now, we are very trusty here as we expect it to work if it worked
+ # in the single-stream case
+
+ # write objects
+ reader = IteratorReader(istream_generator())
+ istream_reader = db.store_async(reader)
+ istreams = istream_reader.read() # read all
+ assert istream_reader.task().error() is None
+ assert len(istreams) == ni
+
+ for stream in istreams:
+ assert stream.error is None
+ assert len(stream.binsha) == 20
+ assert isinstance(stream, IStream)
+ # END assert each stream
+
+ # test has-object-async - we must have all previously added ones
+ reader = IteratorReader( istream.binsha for istream in istreams )
+ hasobject_reader = db.has_object_async(reader)
+ count = 0
+ for sha, has_object in hasobject_reader:
+ assert has_object
+ count += 1
+ # END for each sha
+ assert count == ni
+
+ # read the objects we have just written
+ reader = IteratorReader( istream.binsha for istream in istreams )
+ ostream_reader = db.stream_async(reader)
+
+ # read items individually to prevent hitting possible sys-limits
+ count = 0
+ for ostream in ostream_reader:
+ assert isinstance(ostream, OStream)
+ count += 1
+ # END for each ostream
+ assert ostream_reader.task().error() is None
+ assert count == ni
+
+ # get info about our items
+ reader = IteratorReader( istream.binsha for istream in istreams )
+ info_reader = db.info_async(reader)
+
+ count = 0
+ for oinfo in info_reader:
+ assert isinstance(oinfo, OInfo)
+ count += 1
+ # END for each oinfo instance
+ assert count == ni
+
+
+ # combined read-write using a converter
+ # add 2500 items, and obtain their output streams
+ nni = 2500
+ reader = IteratorReader(istream_generator(offset=ni, ni=nni))
+ istream_to_sha = lambda istreams: [ istream.binsha for istream in istreams ]
+
+ istream_reader = db.store_async(reader)
+ istream_reader.set_post_cb(istream_to_sha)
+
+ ostream_reader = db.stream_async(istream_reader)
+
+ count = 0
+ # read it individually, otherwise we might run into the ulimit
+ for ostream in ostream_reader:
+ assert isinstance(ostream, OStream)
+ count += 1
+ # END for each ostream
+ assert count == nni
+
+