Re: [Zope] Help writing test where bobobase_modification_time changes
Thanks for the tips Chris and Dieter, I actually solved it in a very simple way: from OFS.Image import File from time import time class MockFile(File): This file works the same as a normal File expect that in manage_edit() the internal modification time is set immediately. Normally Zope objects get their internal modification set when the transaction manager commits. def manage_edit(self, title, content_type, precondition='', filedata=None, REQUEST=None): # here's the mock hack, we set the _p_mtime NOW instead of letting # the transaction manager do it at the end. self._p_mtime = time() return File.manage_edit(self, title, content_type, precondition=precondition, filedata=filedata, REQUEST=REQUEST) This worked beautifully in the integration test. I'm happy. On 10/04/2008, Chris Withers [EMAIL PROTECTED] wrote: Peter Bengtsson wrote: How can I solve this? Make one big transaction two without having to use transaction.get().commit() which I don't think I can use in unit tests as per Dieter's advice. I'd suggest monkeypatching bobobase_modification_time on the objects you're testing with. You're testing your cache code, not the implementation of those objects, so replacing them with (partially) dummy objects is great for avoiding this kind of problem and is a pattern I've used frequently... cheers, Chris -- Simplistix - Content Management, Zope Python Consulting - http://www.simplistix.co.uk -- Peter Bengtsson, work www.fry-it.com home www.peterbe.com hobby www.issuetrackerproduct.com ___ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )
Re: [Zope] Help writing test where bobobase_modification_time changes
Peter Bengtsson wrote: How can I solve this? Make one big transaction two without having to use transaction.get().commit() which I don't think I can use in unit tests as per Dieter's advice. I'd suggest monkeypatching bobobase_modification_time on the objects you're testing with. You're testing your cache code, not the implementation of those objects, so replacing them with (partially) dummy objects is great for avoiding this kind of problem and is a pattern I've used frequently... cheers, Chris -- Simplistix - Content Management, Zope Python Consulting - http://www.simplistix.co.uk ___ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )
Re: [Zope] Help writing test where bobobase_modification_time changes
Peter Bengtsson wrote at 2008-4-8 19:08 +0100: ... The reason it fails on the last line is that the time difference between the first time and the second time is nil. Adding a time.sleep(1) won't change anything because the bobobase_modification_time() doesn't change. How can I solve this? Not easy. bobobase_modification_time() is in fact a DateTime representation of _p_serial. And _p_serial is the transaction id (also a timestamp) that has committed the object state belonging to this object incarnation. Unfortunately, you must not commit a transaction in a test suite (as this may cause interference with other tests). There is a small chance that a savepoint changed _p_serial -- but the chance is only small... You might be able to write _p_serial yourself and thereby trick bobobase_modification_time(). -- Dieter ___ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )
[Zope] Help writing test where bobobase_modification_time changes
I'll try to explain as briefly as possible.
I've got a working cache pattern that keeps a cached version of a Zope
file's content and only when it changes (by
bobobase_modification_time) the cache is invalidated and recreated.
Here's the gist of my test
class Test(ZopeTestCase.FunctionalTestCase):
def test_recreateCache(self):
bla
self.folder.manage_addFile('foo.js','FOO data')
url = self.folder.mytool.cacheFile('foo.js') # pseudo code this time
path = urlparse(url)[2]
response = self.publish(path)
self.assert... tests on response...
_file = getattr(self.folder, 'foo.js')
_file.manage_edit(_file.title, _file.content_type,
filedata='new data')
newurl = self.folder.mytool.cacheFile('foo.js')
assert newurl != url, url hasn't changed :(
The reason it fails on the last line is that the time difference
between the first time and the second time is nil. Adding a
time.sleep(1) won't change anything because the
bobobase_modification_time() doesn't change.
How can I solve this? Make one big transaction two without having to
use transaction.get().commit() which I don't think I can use in unit
tests as per Dieter's advice.
Grateful for help.
Peter
--
Peter Bengtsson,
work www.fry-it.com
home www.peterbe.com
hobby www.issuetrackerproduct.com
___
Zope maillist - Zope@zope.org
http://mail.zope.org/mailman/listinfo/zope
** No cross posts or HTML encoding! **
(Related lists -
http://mail.zope.org/mailman/listinfo/zope-announce
http://mail.zope.org/mailman/listinfo/zope-dev )
