Minor change to Splice. Deleted old commented lines
Corrected PSSM2BfactorMover . It can now handle poses with more than one chain.
Added braces to if statement in OopCreatorMover . Before it caused compliation error in gcc compiler
notify author
notify list [rosetta-logs@googlegroups.com]
Convert FragmentPicker from boost::thread to std::thread for easier use.
Now, to pick fragments with multiple cpus, you must compile with extras=cxx11thread.
Also, fixing an issue in ChiralMover that was causing a compile error on my machine.
Specifically, there is no need to add tempaltes to std::make_pair. No test changes
expected.
notify author
notify list [rosetta-logs@googlegroups.com]
WARNINGS WAR Z: This should finally fix the linux.clang build.
I tested all of these on the testing server, which has a newer version of Clang.
----------------------------------------
Unit test status: Pass
notify author
notify list [rosetta-logs@googlegroups.com]
Warning Removal: turning one error back to warning
waiting on Fang to fix 3 remaining enum-compare warnings
This should fix the build for those using versions of GCC
more recent than the test server's.
notify author
notify list [rosetta-logs@googlegroups.com]
02/27/14 Changing Enzdes to be more campitable with talaris; default scorefxn is now talaris2013_cst, which results in ~2% increase in seq recovery for enzdes benchmarks.
NEWS: In EnzdesBaseProtocol.cc, the default scorefxn is now talaris2013_cst. To be compatible with talaris2013, the exclude_protein_protein_fa_elec_ bool option in the constructor is now set to false by default (it used to be true, which was necessary when using the enzdes scorefxn so not to double count fa_elec of interface sidechains). Using talaris with talaris2013_cst.wts gives ~2% increase in sequence recovery on the enzdes benchmark.
----------------------------------------
Unit test status: Failed when I ran it by not because of my changes ("unit_tests - Broken by Kale Kundert:56400.")
Integration test status: 3 changed (as expected)
enzdes: outputs different results/scores due to using new scorefxn and weights (talaris2013_cst)
inverse_rotamer_remodel: changed the flags file to use talaris_2013_cst.wts; outputs different results/score due to different scorefxn
ligand_motif_design: I did not touch this; this test uses a custom weights file that is not in the database, so I left it as is. Let me know if this is a problem for anyone (sboyken@gmail.com).
notify author
notify list [rosetta-logs@googlegroups.com]
Warnings Removal: removed many uninitialized variable warnings
I am still trying to get the test server to build in linux.clang, but the
version it is using is really new and higher than what most are using, I think.
notify author
notify list [rosetta-logs@googlegroups.com]