[sirikata/sirikata] fc26b2: Make coalescing first set of results optional, dis...

[GitHub]

Branch: refs/heads/master
Home: https://github.com/sirikata/sirikata
Commit: fc26b2648fa62dade3d2c99b1e74090d06787aac
https://github.com/sirikata/sirikata/commit/fc26b2648fa62dade3d2c99b1e74090d06787aac
Author: Ewen Cheslack-Postava <ewe...@cs.stanford.edu>
Date: 2013-04-05 (Fri, 05 Apr 2013)

Changed paths:
M externals/prox
M libspace/plugins/prox/LibproxProximity.cpp
M libspace/plugins/prox/Options.hpp
M libspace/plugins/prox/PluginInterface.cpp
M libtwitter/include/sirikata/twitter/TermRegionQueryHandler.hpp

Log Message:
-----------
Make coalescing first set of results optional, disable by default.

Add an option to the prox plugin to make coalescing the first set of
declarative (non-manual) query results optional. Coalescing them
sounds nice because it gets rid of any intermediate steps that need to
be reported and dumps all the data to the client at once. However, it
means we need to wait for the initial query evaluation to finish,
which could take a bit when the tree is large. Disable coalescing by
default.


Commit: 70ff140b66b7c03020639c2d531e25e2d8046037
https://github.com/sirikata/sirikata/commit/70ff140b66b7c03020639c2d531e25e2d8046037
Author: Ewen Cheslack-Postava <ewe...@cs.stanford.edu>
Date: 2013-04-05 (Fri, 05 Apr 2013)

Changed paths:
M externals/prox

Log Message:
-----------
Update for ZernikeDescriptor bug fix in libprox.


Commit: df60d8b66e8fed43c783145582d9747aa16060c2
https://github.com/sirikata/sirikata/commit/df60d8b66e8fed43c783145582d9747aa16060c2
Author: Ewen Cheslack-Postava <ewe...@cs.stanford.edu>
Date: 2013-04-05 (Fri, 05 Apr 2013)

Changed paths:
M libcore/include/sirikata/core/network/SSTImpl.hpp
M test/unit/libcore/SSTTest.hpp

Log Message:
-----------
Fix bug in tracking available SST segments.

Add a bunch of tests for SST::ReceivedSegmentList, including adding in
various orders, merging in different directions, and repeatedly
inserting parts of segments. This tracked down a bug in the
implementation of ReceivedSegmentList where the wrong segment was
being modified (it should have been next_it, not it).


Commit: 847763cfd07a859a0ee4f98e206a66364d5cc5e9
https://github.com/sirikata/sirikata/commit/847763cfd07a859a0ee4f98e206a66364d5cc5e9
Author: Ewen Cheslack-Postava <ewe...@cs.stanford.edu>
Date: 2013-04-05 (Fri, 05 Apr 2013)

Changed paths:
M test/unit/libcore/SSTTest.hpp

Log Message:
-----------
Add a few more simple SST segment tests.

Check that merging segments and repeated insertion of partial
segments that don't touch the boundaries of the existing segment
(i.e. repeat of a middle packet where 3 packets are already available)
work properly.


Compare: https://github.com/sirikata/sirikata/compare/83a143a6a8ca...847763cfd07a