2015-12-31 22:59  xavier

	* [r14939] Missing object destructions.

2015-12-31 23:55  xavier

	* [r14940] The Hartree exchange operator works with batches.

2016-01-01 00:09  xavier

	* [r14941] The exchange operator subroutine receives batch
	  arguments.

2016-01-01 00:23  xavier

	* [r14942] * The RDMFT operator receives batch arguments.
	  * Added a profiling region for the exchange operators.

2016-01-01 00:43  xavier

	* [r14943] The hamiltonian can now be applied in packed mode for
	  more cases.

2016-01-01 19:46  xavier

	* [r14944] More code is compiled without scalapack.

2016-01-01 19:53  xavier

	* [r14945] matrix_diagonalize_hermitian now can use scalapack to do
	  the
	  diagonalization in parallel.

2016-01-01 21:39  xavier

	* [r14946] * Defined the matrix_gather subroutine to collect a
	  matrix distributed
	  from different processors.
	  * Added push_subs and pop_subs.

2016-01-01 22:52  xavier

	* [r14947] Implemented parallel diagonalization for complex
	  matrices.

2016-01-01 23:15  xavier

	* [r14948] Non-MPI compilation fix.

2016-01-02 00:06  xavier

	* [r14949] Added checks for scalapack function return status.

2016-01-02 02:01  xavier

	* [r14950] Fixed compilation with BerkeleyGW.

2016-01-02 17:02  xavier

	* [r14951] Removed the complex-scaling pointers from states.

2016-01-03 00:38  xavier

	* [r14952] Replaced pointers by allocatables.

2016-01-03 01:03  xavier

	* [r14953] Bugfix: The BerkeleyGW interface was setting hm%hf_st
	  and leaving it
	  like that. This caused a double deallocation of the states. Now
	  the
	  pointer is nullified.

2016-01-03 01:35  xavier

	* [r14954] Octopus now prints the call stack (if running in debug
	  mode) after a segmentation
	  fault.

2016-01-03 02:50  xavier

	* [r14955] Added argument names to lapack calls.

2016-01-03 16:06  xavier

	* [r14956] The recipe calculation mode is now handled at the
	  beginning of the
	  execution. This fixes a test failing in the valgrind buildbot.

2016-01-03 16:17  xavier

	* [r14957] An array was not initialized in the convert utility.
	  This fixes a
	  failing test in the valgrind buildbot.

2016-01-03 16:22  xavier

	* [r14958] Fixed in style, indentation and removed unused
	  variables.

2016-01-03 16:44  xavier

	* [r14959] Removed unused global variable that was causing valgrind
	  errors.

2016-01-03 16:49  xavier

	* [r14960] Removed unused global variable that was causing valgrind
	  errors.
	  (Missing file from my previous commit.)

2016-01-03 16:57  xavier

	* [r14961] Reading and writing a string at the same time is not a
	  good idea.

2016-01-03 17:12  xavier

	* [r14962] Missing object destructions.

2016-01-03 17:24  xavier

	* [r14963] Updated the results for the convert test, that changed
	  with my bugfix.

2016-01-03 19:14  xavier

	* [r14964] Reverted oct-convert to run in parallel in the tests.

2016-01-03 23:55  xavier

	* [r14965] Die properly after a segmentation fault.

2016-01-04 00:08  xavier

	* [r14966] Removed a debugging sleep.

2016-01-04 00:54  xavier

	* [r14967] Bugfix: An argument had the wrong intent, this was
	  causing some test failures.

2016-01-04 04:10  xavier

	* [r14968] Removed an assertion, since this routine is used inside
	  assertions.

2016-01-04 04:10  xavier

	* [r14969] Bugfix: The sparskit and arpack macros tried to link the
	  libraries
	  given as argument even if the detection failed.

2016-01-04 17:49  xavier

	* [r14970] Wrong name in POP_SUB.

2016-01-04 18:18  xavier

	* [r14971] Fixed multiple errors in PUSH_SUB/POP_SUB from the
	  'frozen' code.

2016-01-05 02:57  xavier

	* [r14972] Replaced the variable 'DebugLevel' by 'Debug'. Now it is
	  possible to
	  select the debug information that octopus provides.
	  
	  For example 'Debug = info' prints the extra info to the screen.
	  'Debug
	  = trace' generates the call stack and prints it when something
	  goes
	  wrong. 'Debug = trace_file' and 'Debug = trace_term' print to
	  files or
	  the terminal. Of course you can mix, for example 'Debug = info +
	  trace_term'.

2016-01-05 05:41  xavier

	* [r14973] Missing file from my previous commit.

2016-01-05 05:48  xavier

	* [r14974] * Fixed compilation dependencies.
	  * PUSH_SUB and POP_SUB macros use the new debug values.
	  * debug is exported through messages_m.

2016-01-05 14:20  xavier

	* [r14975] The debug variable is now defined in messages_m.

2016-01-05 14:27  xavier

	* [r14976] Removed the in_debug_mode variable from global_m, it is
	  now replaced by
	  debug%info from messages_m.

2016-01-05 14:39  xavier

	* [r14977] Added a new debug option 'Debug = extra_checks' that can
	  be used to
	  enable consistency checks that might be too costly for regular
	  executions.

2016-01-05 14:48  xavier

	* [r14978] Corrected POP_SUB.

2016-01-05 15:04  xavier

	* [r14979] Fixed indentation.

2016-01-05 15:08  xavier

	* [r14980] * Fixed PUSH_SUB/POP_SUB.
	  * Removed unnecessary or buggy 'return' statements.

2016-01-05 15:34  xavier

	* [r14981] Style fixes to the pcm_m module.

2016-01-05 17:45  jrfsousa

	* [r14983] Frozen:
	  
	  * Add missing dependencies for Makefile addressing issue #136
	  
	  * Fix some more PUSH_SUB/POP_SUB

2016-01-06 17:46  jrfsousa

	* [r14985] Frozen:
	  
	  base_density.F90, base_functional.F90, base_hamiltonian.F90,
	  storage.F90:
	  
	  * Minor changes.
	  
	  fio_config.F90, fio_density.F90:
	  
	  * Add support to spin inversion.
	  
	  fio_density.F90, fio_external.F90:
	  
	  * Reflect "intrpl_inc.F90" template removal.
	  
	  fio_states.F90, fio_system.F90, frozen_density.F90,
	  frozen_external.F90:
	  
	  * Reflect changes.
	  
	  interface_xc.F90:
	  
	  * Bug fix.

2016-01-06 18:34  dstrubbe

	* [r14986] Set svn:ignore properties for new Makefiles.

2016-01-06 19:22  xavier

	* [r14987] I accidentally made two integer values into floats.

2016-01-06 19:51  xavier

	* [r14988] The frozen/ tests filenames now follow the convention as
	  the other
	  tests.

2016-01-06 20:03  xavier

	* [r14989] The pseudopotential localization threshold must be read
	  before the
	  initialization. This fixes a valgrind test failure.

2016-01-06 20:25  jrfsousa

	* [r14990] Build:
	  
	  mk_kfunctionals_list.pl:
	  
	  * Documentation improvement.
	  
	  Frozen:
	  
	  ssys_config.F90:
	  
	  * Documentation improvement.
	  
	  * Add support for non polarized calculation of the kinetic non
	  additional term.
	  
	  System:
	  
	  output_h_inc.F90:
	  
	  * Reflect changes.
	  
	  v_ks.F90:
	  
	  * Bug fix.

2016-01-06 21:07  jrfsousa

	* [r14991] * Bug fix, section::subsection order in documentation.

2016-01-06 23:22  xavier

	* [r14995] Bugfix: When copying a states_group_t, do not allocate
	  objects unless
	  the original is initialized. This solves some valgrind errors.

2016-01-06 23:33  xavier

	* [r14996] Missing block parser destruction. Fixes a valgrind
	  error.

2016-01-06 23:55  xavier

	* [r14997] Bugfix: messages_experimental must be called after
	  io_init. Fixes a bug in the g95 builder.

2016-01-07 00:04  xavier

	* [r14998] Converted an allocatable member to pointer. Workaround
	  for a g95 problem.

2016-01-07 00:19  xavier

	* [r14999] Bugfix: the call stack was not printed unless debug info
	  was enabled.

2016-01-07 00:38  xavier

	* [r15000] * Print the debug trace with more signals, including INT
	  (CTRL + C) ,
	  KILL (CTRL + \) and TSTP (CTRL + z). This last one is quite
	  useful,
	  because one can know where the execution is without killing
	  Octopus.
	  
	  * Improved the debug message.

2016-01-07 06:56  xavier

	* [r15001] Bugfix: There is a problem with the calculation of the
	  mGGA terms with packed
	  batches in parallel.

2016-01-07 11:34  jrfsousa

	* [r15002] States:
	  
	  states.F90:
	  
	  * Add substates_set_charge procedure to set live subsystem
	  charge.
	  
	  testsuite/frozen:
	  
	  * Add some more tests.

2016-01-07 15:28  jrfsousa

	* [r15003] Frozen:
	  
	  base_density.F90
	  
	  base_hamiltonian.F90, base_potential.F90, base_states.F90,
	  storage.F90:
	  
	  * Uniformization of allocation verification.
	  
	  testsuite/frozen:
	  
	  * Some more tests.

2016-01-07 19:09  xavier

	* [r15004] Bugfix: The alternative LCAO was using the wrong
	  communicator when
	  running parallel in states without Scalapack.

2016-01-07 23:26  xavier

	* [r15005] Bugfix: oct-local_multipoles was trying to use an
	  uninitialized object.

2016-01-07 23:40  xavier

	* [r15006] Add the OpenCL library to the path to test for OpenCL
	  libraries.

2016-01-08 00:09  xavier

	* [r15007] Only release the opencl buffer from mesh_cube_map_t if
	  the array is
	  also allocated. This should fix the problem with frozen and the
	  opencl
	  buildbot.

2016-01-08 01:13  dstrubbe

	* [r15008] Bugfix: the attempt to determine a denominator for the
	  k-point weights is not well-conditioned numerically; some more
	  tolerance is required to work properly on the new Cori machine at
	  NERSC, or else all cases are rejected as having non-rational
	  numbers.

2016-01-08 09:32  umberto

	* [r15009] Restructuring the oct-photoelectron-spectrum utility.
	  
	  The changelog includes:
	  * Support for zero-weight path in reciprocal space (only along kx
	  and ky).
	  * Support for spin and state resolved spectra.
	  * The quantity to be calculated is now specified in the input
	  file with the PhotoelectronSpectrumOutput option. This
	  substitutes the '-m' command line option of the old version.

2016-01-08 11:58  jjornet

	* [r15010] - Bugfix: Reverted revision 14957. Now the variable is
	  properly initialized.
	  - Reverting changes of revision 14963 for the test values.

2016-01-08 15:37  xavier

	* [r15011] OpenCL: Try the cl_khr_fp64 extension first, then
	  cl_amd_fp64.

2016-01-08 19:56  xavier

	* [r15013] clblas interface: Replaced the deprecated clAmdBlas
	  prefix to clblas.

2016-01-08 21:16  xavier

	* [r15014] * Renamed clblas files and preprocessor macro.

2016-01-08 21:24  xavier

	* [r15015] Renamed files and macros from clamdfft to clfft,
	  following the change
	  in library name.

2016-01-08 21:44  xavier

	* [r15016] Renamed clamdfft functions to clfft.

2016-01-08 23:29  xavier

	* [r15017] Inteface for the dot product clblas functions.

2016-01-08 23:36  xavier

	* [r15018] The vector dot product subroutine now uses clblas.

2016-01-09 00:48  xavier

	* [r15019] * Interface to the nrm2 function of clblas.
	  * mesh_batch_nrm2 now uses clblas.

2016-01-09 00:56  xavier

	* [r15020] Removed unused variables.

2016-01-09 00:58  xavier

	* [r15021] The nrm2_vector OpenCL kernel is no longer required.

2016-01-09 04:19  xavier

	* [r15022] The name of the testfile is now included in the test
	  temporary file.

2016-01-10 19:33  xavier

	* [r15036] Added a profiling region for output.

2016-01-11 17:59  irina

	* [r15037] *Corrected total energy of periodic systems related to G
	  = 0 contribution of pseudopotentials. Currently added to the
	  ion-ion interaction energy. *Corresponding changes to the
	  testsuite for periodic systems are introduced (total energy and
	  ion-ion energy).

2016-01-11 18:20  dstrubbe

	* [r15038] Bugfix (trivial): remove duplicated message about libxc
	  version 3 from r14515.

2016-01-11 19:42  xavier

	* [r15040] Fixed the default value in the documentation for
	  EwaldAlpha.

2016-01-12 09:34  philipp

	* [r15046] Adding some factors in pes_flux and pes_spm.

2016-01-12 09:58  philipp

	* [r15047] Changing in tests due to previous commit.

2016-01-12 10:01  philipp

	* [r15048] Correction of some volume elements in pes_spm &
	  pes_flux.

2016-01-12 15:53  dstrubbe

	* [r15050] Add new release to PACKAGING file in trunk; yaml is
	  internal/optional.

2016-01-13 01:26  xavier

	* [r15051] Proper implementation of the application of the time
	  dependent fields to time independent calculations:
	  
	  * Now it is done optionally, enabled by the experimental TimeZero
	  variable. It is disabled by default since this can be quite
	  confusing for users.
	  
	  * The implemetation is handled inside Hamiltonian update, so it
	  is always done. Before it was done in v_ks, leading to
	  inconsistent behavior.
	  
	  * This behavior is described in the documentation for
	  TDExternalFields.
	  
	  * Updated a test that uses this feature.

2016-01-13 15:31  xavier

	* [r15052] Corrected the energy values for the ETSF I/O test.

2016-01-13 17:29  dstrubbe

	* [r15053] Bugfix: the current checks on whether metis is double
	  precision are only effective for some versions; others have
	  different things in metis.h. Now another possibility is caught
	  too.

2016-01-14 02:10  dstrubbe

	* [r15054] Rewrite stitch routines for complex-scaling to avoid
	  ifort warnings about creation of array temporaries. They are
	  quite small here, but this helps to be able to run with "-check
	  all" enabled and focus on any serious issues that arise.

2016-01-14 06:56  xavier

	* [r15055] Bugfix: when writing the restart, with state
	  parallelization and real
	  states, some of the files were missing.

2016-01-14 12:19  jrfsousa

	* [r15056] testsuite/frozen:
	  
	  * Tighten convergence parameters.

2016-01-14 15:11  jrfsousa

	* [r15057] testsuite/frozen:
	  
	  * Kinetic functional tests.

2016-01-14 16:50  xavier

	* [r15058] The default values defined in the variable documentation
	  are now
	  accessible from the code. For variable "var" the default value is
	  a
	  macro called DEFAULT__VAR. For the moment this is
	  straight-forward to
	  use with integer values. I think I can fix the code to handle
	  other
	  cases (real values, booleans and options).

2016-01-14 17:25  xavier

	* [r15060] Fixed an error in my previous commit, the options.h file
	  was not
	  generated properly.

2016-01-14 17:30  xavier

	* [r15061] Apparently some compiler requires the types of the
	  submembers of a
	  type to be public.

2016-01-14 20:02  jrfsousa

	* [r15062] testsuite/frozen:
	  
	  * Kinetic functional with spin tests.

2016-01-14 23:34  xavier

	* [r15063] Bugfix: Corrected the eigensolver progress bar when
	  state or k-point
	  parallelization is used.

2016-01-15 00:43  dstrubbe

	* [r15064] svn:ignore for defaults.h.

2016-01-15 00:45  dstrubbe

	* [r15065] Improvements to variable reference.

2016-01-15 10:40  jjornet

	* [r15066] - Removed sr_dist variable to avoid buffer overflow. Now
	  the distances between tesserae and grid points are computed on
	  the fly in pcm_pot_rs subroutine.

2016-01-15 11:06  jjornet

	* [r15067] - Added default van der Waals radius values from Alvarez
	  S., Dalton Trans., 2013, 42, 8617-8636.
	  (this values will be used in PCM calculations)

2016-01-15 13:14  jjornet

	* [r15068] - vdW radius parameter added in the species_t.

2016-01-15 13:23  jjornet

	* [r15069] - vdW radius are now obtained from
	  /share/pseudopotential/element file or from %Species block.
	  - pcm inputs modified to agree with the testsuite results.

2016-01-15 16:30  xavier

	* [r15070] Centering of a value for the KS inversion test.

2016-01-15 21:41  jrfsousa

	* [r15071] Build:
	  
	  mk_kfunctionals_list.pl:
	  
	  * Update script to add version number.
	  
	  Frozen:
	  
	  kfunctionals_list.F90:
	  
	  * Regenerate with older libxc.

2016-01-15 21:42  jrfsousa

	* [r15072] testsuite/frozen:
	  
	  * Adress test failures.

2016-01-15 21:43  jrfsousa

	* [r15073] testsuite/frozen:
	  
	  * Remove tests with missing functionals.

2016-01-16 15:49  xavier

	* [r15074] Updated the ion-ion and total energy results for the
	  BerkeleyGW test.

2016-01-18 17:12  jjornet

	* [r15076] - mpi routines implemented to avoid simultanious
	  writting to the same file when running in parallel.
	  - PCMGamessBenchmark variable controls to output the gamess
	  formatted pcm_mat.

2016-01-20 22:29  adelgado

	* [r15077] - pcm.F90: In subroutine 'pcm_pot_rs' the selection
	  between the regularized and standard expresion to calculate
	  solvent reaction field in real space is pulled out of the loops
	  through the grid and cavity representative points. This should
	  speed up PCM calculations for large systems.

2016-01-21 09:46  philipp

	* [r15078] Reading sample points from block for different
	  dimensions.

2016-01-21 10:34  philipp

	* [r15079] Changing floats to integers for counters.

2016-01-21 12:07  philipp

	* [r15080] Choosing surface points according to maximum angular
	  momentum.

2016-01-26 12:51  philipp

	* [r15082] Changing default scheme for sample point method.

2016-01-27 15:07  philipp

	* [r15083] * Renaming some variables in pes_spm.
	  * Changing output of momentum distribution.
	  * Fixing a small bug in pes_flux and changing values of test.

2016-01-29 09:11  dstrubbe

	* [r15084] Small correction to XSF output, as pointed out by Ask in
	  ASE interface. The files were considered legal by XCrySDen still.

2016-02-02 14:27  jjornet

	* [r15085] - Convert utility now write some input information when
	  doing a FT of the mesh functions.
	  - Modified default value of ConvertReadSize.

2016-02-02 18:22  jjornet

	* [r15086] - Rewriting some variables and loops to make it more
	  comprensible.

2016-02-03 14:58  jjornet

	* [r15087] - Corrected the way to subtract reference mesh function.
	  - Now the convert can perform a fourier transform of real mesh
	  function in parallel in domains.

2016-02-04 12:10  jrfsousa

	* [r15088] Frozen:
	  
	  interface_xc.F90:
	  
	  * Some optimizations.
	  
	  testsuite/frozen:
	  
	  * Comments, turn on gga_k_apbe test.

2016-02-05 09:23  umberto

	* [r15089] Fix: Properly fill spin values with states and/or
	  kpoints parallelization

2016-02-08 13:41  jjornet

	* [r15091] - Bugfix : input information is just written by
	  mpi_world%rank == 0

2016-02-08 16:59  jjornet

	* [r15092] - If statement missplaced in the last commit.
	  - Corrected the starting and ending energy labels.

2016-02-10 14:11  jjornet

	* [r15093] - change in some variable range to be consistent with
	  the results obtained using the oct-propagation_spectrum utility

2016-02-10 16:03  jjornet

	* [r15094] - Reset the array range of out_fft variable.

2016-02-12 17:58  jjornet

	* [r15095] - Implementing parallelization in domains in PCM
	  calculation as an experimental feature.

2016-02-12 18:41  adelgado

	* [r15096] Fixing calling statements of the previous (15095)
	  commit.

2016-02-15 18:36  acastro

	* [r15097] Added one more linear solver: bicgstab(2). It is
	  supposed to be better than
	  bicgstab. Right now only for real operators.

2016-02-16 17:08  umberto

	* [r15098] Fix: NFFT grid should be consistent with the FFT one for
	  periodic dimensions

2016-02-23 17:00  umberto

	* [r15099] Allow for parallel pcm tests

2016-02-23 17:05  umberto

	* [r15100] * update pcm potential every given iteration number
	  * comment out unused trilinear interpolation variables

2016-02-23 17:58  jjornet

	* [r15101] - The pcm_info.out is printed by the root.

2016-02-24 18:45  acastro

	* [r15102] Bug fix in the Runge-Kutta propagator, introduced in
	  r14905. Unfortunately
	  there was not test checking this.

2016-02-24 23:20  umberto

	* [r15103] Calculate the pcm_potential as a solution of the Poisson
	  equation for a gaussian smeared polarizaion charge density. This
	  is much faster than the previously implemente direct sum but the
	  results are slightly different. These discrepancies need to be
	  carefully checked.

2016-02-25 14:27  jjornet

	* [r15105] - PCM iteration information is now printed by the root
	  in the pcm module.

2016-02-25 18:44  dstrubbe

	* [r15106] Fix comment syntax which is misunderstood by doxygen as
	  being special information for its use, resulting in several
	  warnings in the doxygen log: "warning: inconsistency between
	  intent attribute and documenation for variable: iter"

2016-02-25 19:30  dstrubbe

	* [r15108] Bugfix to misleading configure --help output: GSL is not
	  optional.

2016-02-25 20:47  dstrubbe

	* [r15112] Add bash script for buildbot to use in checking config.h
	  against reference.

2016-02-25 22:51  xavier

	* [r15116] Bugfix: A vectorial expression with implicit bounds was
	  causing an out
	  of bounds access.

2016-02-25 22:51  dstrubbe

	* [r15117] Clearer output from checking config.h.

2016-02-25 23:20  dstrubbe

	* [r15121] For consistency with the other configure options, the
	  gsl ones should say DIR instead of PFX.

2016-02-25 23:32  xavier

	* [r15122] Fixed another out of bounds access.

2016-02-25 23:36  xavier

	* [r15123] Avoid using the messages array as a temporary.

2016-02-25 23:42  dstrubbe

	* [r15124] Clarification on --enable-mpi(=PATH).

2016-02-26 03:19  dstrubbe

	* [r15126] We will now let a diff in config.h cause the configure
	  step to be considered failing on the buildbot.

2016-02-26 16:07  jjornet

	* [r15127] - An option added to use the optimized vdw radii for pcm
	  calculation. By default octopus will use these values.
	  - Restored the pcm input file for the testsuite.
	  - Error fixed on species file.

2016-02-26 16:37  umberto

	* [r15128] The nearest neighbor gaussian smearing used to generate
	  the polarization charge density can be enlarged to an arbitrary
	  number of neighboring points with PCMChargeSmearNN. This allow to
	  find the optimal number to converge the results.

2016-02-26 18:16  dstrubbe

	* [r15131] Cleaning up documentation on PCM variables. Quim, please
	  note that the 'Default' field is only for the information of the
	  user and not manipulated directly by the code, so it does not
	  offer any information to write a variable name there like
	  'default_value'.

2016-02-26 18:55  dstrubbe

	* [r15132] The GSL configure macro should report problems with
	  AC_MSG_ERROR not just echo. Clearer message when an inadequate
	  version is found. Simplifying macro by getting rid of arguments.

2016-02-26 19:52  xavier

	* [r15133] Recentering of a test.

2016-02-26 19:55  xavier

	* [r15135] Fixed wrong constant declaration.

2016-02-27 13:01  xavier

	* [r15139] The OpenSCAD file output test fails for OpenCL. It seems
	  that small
	  numerical differences cause the file to be quite different. For
	  the
	  moment I will disable OpenCL for this test.

2016-02-29 14:52  xavier

	* [r15140] Bugfix: The imaginary part of the potential was not
	  considered with
	  OpenCL.

2016-02-29 17:01  xavier

	* [r15143] Updated again the results for this test. One buildbot
	  was slightly of
	  outside the range.

2016-02-29 19:57  dstrubbe

	* [r15145] * The FEAST m4 should set LIBS_FEAST and that should be
	  used in the link line.
	  * Fix indentation for FEAST m4. Remove amusing but irrelevant
	  commented part.
	  * Move LIBS_MPI to a more appropriate location in link line.
	  * Comment on libyaml.

2016-02-29 20:36  dstrubbe

	* [r15147] Update buildbot_query.pl given new buildbot step to
	  check config.h: the test is now shell #4.

2016-02-29 20:57  dstrubbe

	* [r15148] Removing useless semicolon.

2016-02-29 20:57  dstrubbe

	* [r15149] Removing unused variables.

2016-02-29 21:53  dstrubbe

	* [r15150] Removed papi support. It relies on a long-obsolete
	  version of the papi library.

2016-03-01 01:14  xavier

	* [r15152] Removed some highly experimental methods to calculate
	  the current density.

2016-03-01 14:58  xavier

	* [r15153] Bugfix: 'TDOutput = eigenvalues' was printing only
	  eigenvalues for one
	  k-point. (Thanks to Nicolas Tancogne-Dejean for reporting this.)

2016-03-01 16:54  dstrubbe

	* [r15154] Being more explicit about the meaning of Monkhorst-Pack
	  k-point shifts.

2016-03-01 22:01  umberto

	* [r15157] Contingency plans in case some os nearest neighbors of
	  each tessera falls outside the mesh.
	  The code tries to reduce the number of points until it fits and
	  warn the user. It also should warn the user in case one forces a
	  large number anyways.

2016-03-02 15:26  xavier

	* [r15158] Created the 6.x branch.

2016-03-02 15:28  xavier

	* [r15159] Set the version number 6.x.

2016-03-02 22:11  xavier

	* [r15171] Backport of 15170 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15170 | xavier | 2016-03-02 14:10:35 -0800 (Wed, 02 Mar 2016) |
	  3 lines
	  
	  Changed the default installation location to /usr/local.
	  
	  
	  ------------------------------------------------------------------------

2016-03-02 22:23  xavier

	* [r15173] Backport of 15172 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15172 | xavier | 2016-03-02 14:22:52 -0800 (Wed, 02 Mar 2016) |
	  3 lines
	  
	  Bugfix: wrong build dependency.
	  
	  
	  ------------------------------------------------------------------------

2016-03-04 22:01  dstrubbe

	* [r15182] Backport of 15176 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15176 | dstrubbe | 2016-03-04 14:07:03 -0500 (Fri, 04 Mar 2016)
	  | 2 lines
	  
	  Bugfix: the nfft m4 script unwisely appends its CFLAGS and LIBS
	  to the global ones (which should not be done regardless), even if
	  it found that they did not work.
	  
	  ------------------------------------------------------------------------
	  Backport of 15178 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15178 | dstrubbe | 2016-03-04 15:34:48 -0500 (Fri, 04 Mar 2016)
	  | 2 lines
	  
	  Bugfix to r15176: while nothing should be appended to CFLAGS and
	  LIBS in the nff4 m4 script, they should be restored to the saved
	  values.
	  
	  ------------------------------------------------------------------------

2016-03-27 16:08  xavier

	* [r15240] Backport of 15231 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15231 | xavier | 2016-03-23 12:40:28 -0700 (Wed, 23 Mar 2016) |
	  3 lines
	  
	  Bugfix: the clfft macro should add the OpenCL flags to the tests.
	  
	  
	  ------------------------------------------------------------------------

2016-03-30 05:53  xavier

	* [r15245] Backport of 15244 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15244 | dstrubbe | 2016-03-29 13:08:03 -0700 (Tue, 29 Mar 2016)
	  | 2 lines
	  
	  Minor bugfix: since we know the name of the variable to be looked
	  up, with an obsolete variable, we should recommend 'oct-help -p'.
	  
	  ------------------------------------------------------------------------

2016-06-14 15:44  xavier

	* [r15411] Backport of 15234 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15234 | xavier | 2016-03-25 06:30:30 -0700 (Fri, 25 Mar 2016) |
	  3 lines
	  
	  Bugfix: A variable was not initialized.
	  
	  
	  ------------------------------------------------------------------------

2016-06-14 15:49  xavier

	* [r15412] Backport of 15330 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15330 | xavier | 2016-05-02 09:55:11 -0700 (Mon, 02 May 2016) |
	  4 lines
	  
	  Bugfix: The ConvForce criterion did not convert the units of the
	  default value.
	  
	  
	  ------------------------------------------------------------------------

2016-06-14 15:49  xavier

	* [r15413] Backport of 15331 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15331 | xavier | 2016-05-02 10:08:19 -0700 (Mon, 02 May 2016) |
	  3 lines
	  
	  Bugfix: ConvRelEv units were not converted.
	  
	  
	  ------------------------------------------------------------------------

2016-08-25 06:57  nicolastd

	* [r15592] Disabling the non-orthogonal cells.

2016-09-01 12:57  micael

	* [r15596] Increasing tolerance of LOBPCG test.

2016-09-01 15:01  nicolastd

	* [r15597] Correct default value for the total current calculation.

2016-09-02 08:16  micael

	* [r15598] Updating results of a test that depends on the default
	  method to calculate the current.

2016-09-02 10:52  micael

	* [r15600] Backport of 15599 to branches/6.x:
	  ------------------------------------------------------------------------
	  r15599 | micael | 2016-09-02 11:44:49 +0200 (Fri, 02 Sep 2016) |
	  2 lines
	  
	  Removing tests for functionals that were removed in Libxc 3.0.
	  
	  ------------------------------------------------------------------------

2016-09-02 13:39  micael

	* [r15602] Updated tolerance in linear_response test.

2016-09-05 07:39  micael

	* [r15603] Increasing tolerance of test.

