Merge pull request #771 from RosettaCommons/kylebarlow/general_dev
Fixing cmake cppdb linking errors in 2013-onward linux g++ versions
Committing fix suggested by Tom Linsky on rosetta-devel@googlegroups.com on Fri, May 8, 2015.
Solution also referenced at: http://stackoverflow.com/questions/20369672/undefined-reference-to-dlsym
This doesn't break the Clang builds on my machine, but I'll be testing that, and the behavior of other g++ versions, on the test server shortly.
Linking error fixed by this:
libcppdb.so: undefined reference to dlsym' libcppdb.so: undefined reference todlclose'
libcppdb.so: undefined reference to `dlopen'
collect2: error: ld returned 1 exit status
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #804 from RosettaCommons/vmullig/peptide_predict
Adds a pilot app to predict structures of backbone-cyclized peptides
This is basically one of my RosettaScripts converted to C++ and enhanced with tunable parameters that can't be adjusted easily in RosettaScripts without rewriting the script.
Tasks completed:
-- Add simple_cycpep_predict.cc in apps/pilot/vmullig.
-- Read in sequence and build peptide.
-- Cyclize.
-- Circular permutation.
-- Add constraints.
-- Read native.
-- Set omega and randomize mainchain torsions.
-- GenKIC closure
-- GenKIC filtering by rama and total hbonds.
-- Added rama filter to GenKIC.
-- Documentation for new rama filter.
-- Integration test for new rama filter.
-- Added scorefxn option to HbondsToResidue filter.
-- Documentation for this option.
-- FastRelax with high hbond strength of each GenKIC solution.
-- FastRelax top GenKIC solution, updating cyclic O and H.
-- Re-filter by total hbonds.
-- De-permute (undo the cyclic permutation).
-- Calculate RMSD to native.
-- Dump pose and filter results.
-- Repeat nstruct times.
-- Add integration test.
-- Add internal documentation.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #820 from RosettaCommons/vmullig/mac_clang_issue
Trying to fix Mac clang unit test build
Removing the "#define private public" and "#define protected public" lines from unit tests. Updating LoopModelerTests and DsspDesignOperationTests to be friend classes to the classes that they're testing.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #823 from RosettaCommons/roccomoretti/fix_dna
Fix for dna_interface_design test
Convert the debug_assert() which is failing to an if(){ utility_exit_with_message() }, with extra debugging output to (possibly) assist users, and increased output levels on dna_interface_design integration test.
Also, fix the underlying issue with the triazolamerN patch, which is related to deleting the H atom when it's used for the internal coordinates of the base residues types. Instead, just make the H virtual, like N. This may or may not be the long-term fix for this patch.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #819 from RosettaCommons/vmullig/tweak_bettercluster
Small tweaks to bettercluster pilot app
Small tweaks to bettercluster pilot app. No test changes expected.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #817 from RosettaCommons/vmullig/mac_clang_issue
Fix the Mac clang 3.7.0 issue
Mac clang 3.7.0 has thread_local as a C++ language keyword, albeit one that throws an error if used in the absence of c++11. This means that our attempts to define thread_local when not using c++11 throw an error. This is an attempt to remedy this. Instead of defining thread_local, we define THREAD_LOCAL as:
-- nothing if we're not using c++11
-- thread_local if we are
Then we replace all instances of thread_local in the code with THREAD_LOCAL.
Note that there's still an issue to resolve in one unit test (ResidueGraphTypesTests), which has the following lines in it:
// Unit Headers
#define private public
#include <core/chemical/ResidueType.hh>
#undef private
This is still preventing the Mac clang unit tests from compiling. I can see why someone would do this to expose the private functions in the class that they're trying to test, but I also think it's important that the compiler flag attempts to redefine language keywords.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #813 from RosettaCommons/vmullig/unfustercluck
Vmullig/unfustercluck
Rolls master back to Sergey Lyskov's commit 1da8a4c, before a whole lot of stuff was broken. @everyday847, @rhiju, I'm going to create a branch and a pull request for reverting this pull request (i.e. restoring all of the broken stuff). IN THAT BRANCH please fix all the broken stuff, test the branch, then re-merge with master.
notify author
notify list [rosetta-logs@googlegroups.com]
Fixed unused variable
I had intended to cut down on having too much warning output per annotated sequence read-in that I left half-implemented. So sorry! This should be it for now... :-(
notify author
notify list [rosetta-logs@googlegroups.com]