Summary:
The tree datastores have moved from the svfs to the manifestlog. Let's update
fastmanifest to use them from there.
Test Plan: Ran the tests
Reviewers: #fbhgext, ryanmce
Reviewed By: #fbhgext, ryanmce
Differential Revision: https://phab.mercurial-scm.org/D235
Summary:
This assert is meant to check that at least one source was provided to the
hybridmanifest. The old version was broken though.
The test requires an update because it attempts to construct a hybridmanifest
with no source.
Test Plan: Ran the tests
Reviewers: #fastmanifest, ttung
Reviewed By: ttung
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D3760229
Signature: t1:3760229:1472062113:695085444915cb15a93db7f317580dcffc2f115d
Summary: We don't need this since we loosely lock the cache worker.
Test Plan: pass existing tests.
Reviewers: lcharignon, durham
Reviewed By: durham
Subscribers: mitrandir, mjpieters, quark
Differential Revision: https://phabricator.intern.facebook.com/D3537355
Signature: t1:3537355:1468268248:ebf7de5bf5030b90039823e1111a073d7eacc447
Summary:
This patch adds a test that uncovers logic bugs in the cache. It also
adds fixes for these bugs.
Test Plan:
Add a new test and ran existing tests, observed no change for
exisiting test.
Reviewers: durham, ttung
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D3450430
Summary: This will be useful to refactor the cache
Test Plan:
We change the python test to give a vfs that makes sense, it will
be useful in the next patch when we refactor the cache. Indeed, with the new
cache we use systemawarecachelimit by default instead of no limit
Reviewers: ttung
Differential Revision: https://phabricator.intern.facebook.com/D3427796
Summary:
Before this patch, we were passing None as ui for hybridmanifest's
constructor in test-fastmanifest.py. With this patch, we pass a dummy ui object.
Therefore, it removes the need for None guard in hybridmanifest and simplifies
the code.
Test Plan: Tests pass
Reviewers: ttung
Differential Revision: https://phabricator.intern.facebook.com/D3374069
Summary: We want to wrap the fastmanifest cache priming code in a lock, but we want it to be easily stealable. If the lock is more than X seconds in age, we just assume it's stale and steal it. We remember that we stole it so we don't blow away the lock, but we do update the time of the lock so another process doesn't try to steal the lock.
Test Plan: write test to cover basic functionality (mutual exclusivity). write test to cover stealing, and that only one thief is permitted at a time (subject to race conditions).
Reviewers: durham, lcharignon
Reviewed By: lcharignon
Subscribers: mitrandir, mjpieters, #mercurial
Differential Revision: https://phabricator.intern.facebook.com/D3343478
Tasks: 11385124
Signature: t1:3343478:1464304447:18658ebea60f98bfda0f034414991ffd6c334ca7
Summary: This combines D3351047, D3351086, and D3351137 into one diff. This allows mercurial to better remember where things came from since these are all hg cp commands followed by some edits. However, they were separated into multiple diffs for ease of reviews.
Test Plan: run unit tests
Reviewers: lcharignon
Reviewed By: lcharignon
Subscribers: mitrandir, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D3354405
Signature: t1:3354405:1464291057:da140a02b9c87186b1dd9f03f8236d5996319fff
Summary:
As discussed in the group, we want to shuffle by batches to keep
an approximate ordering and still avoid caching process fighting for the
same entries.
Test Plan: Add a new test
Reviewers: durham, ttung
Differential Revision: https://phabricator.intern.facebook.com/D3344144
Summary:
This adds a new strategy for the cache limit that depends on the free
space availabe in the system. For system with more than 100GB of free space
we allocate 5GB for the cache. With more than 20 GB of free space we allocate
2GB of free space and otherwise 10% of the free space at most.
Test Plan: Add a new test
Reviewers: ttung, durham, rmcelroy
Differential Revision: https://phabricator.intern.facebook.com/D3330894
Summary: since it's a singleton, we can have one instance of the code for retrieving the instance.
Test Plan: run tests/test-fastmanifest.t
Reviewers: lcharignon
Reviewed By: lcharignon
Subscribers: mitrandir, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D3284113
Signature: t1:3284113:1463416610:207c7e9954cc0b830e77f2fecc0a65acfcfd1096
Summary: This adds end to end proof of concept of manifest caching supporting
different manifest type
Test Plan: Added a test
Reviewers: ttung, durham
Differential Revision: https://phabricator.fb.com/D3129530
Summary:
This patch adds a wrapper that allows us to instantiate either flat
or fast manifest when a manifest is read.
This needs a change in core to work. In manifest.manifest in core we are
currently using two abstractions to create manifests: calling manifestdict()
or _newmanifest(). This was a hack added for the developement of treemanifest
and needs to be changed to make things saner.
Test Plan:
I added a new test. I also ran the whole test suite in core with
the extension, all the tests passed except test-treemanifest.t, failing on the
assert in _flatmanifest as expected.
python run-tests.py -j32 -l
--extra-config-opt=extensions.fastmanifest=/Users/lcharignon/facebook-hg-rpms/fb-hgext/fastmanifest.py
Reviewers: rmcelroy, ttung, durham
Subscribers: mjpieters
Differential Revision: https://phabricator.fb.com/D3115446
Tasks: 10589064
