2016-01-11 19:45  dstrubbe

	* .: Tagging version 5.0.1.

2016-01-11 18:45  dstrubbe

	* Backport of 15038 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r15038 | dstrubbe | 2016-01-11 13:20:57 -0500 (Mon, 11 Jan 2016)
	  | 2 lines
	  
	  Bugfix (trivial): remove duplicated message about libxc version 3
	  from r14515.
	  
	  ------------------------------------------------------------------------

2016-01-09 20:31  xavier

	* Backport of 14874 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14874 | jrfsousa | 2015-12-07 04:35:16 -0800 (Mon, 07 Dec 2015)
	  | 7 lines
	  
	  
	  Basic, frozen:
	  
	  * Bugfix, missing initialization and recursive declaration.
	  
	  
	  
	  ------------------------------------------------------------------------

2016-01-09 20:27  xavier

	* Backport of 14878 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14878 | jrfsousa | 2015-12-17 10:21:32 -0800 (Thu, 17 Dec 2015)
	  | 9 lines
	  
	  
	  Basic:
	  
	  json.F90:
	  
	  * Workaround PGI compiler bug.
	  
	  
	  
	  ------------------------------------------------------------------------

2016-01-09 20:25  xavier

	* Backport of 14879 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14879 | xavier | 2015-12-20 10:56:57 -0800 (Sun, 20 Dec 2015) |
	  5 lines
	  
	  Bugfix: some 2 body integrals were missing in the output (fix by
	  Michael Schüler).
	  
	  
	  
	  ------------------------------------------------------------------------

2016-01-09 20:21  xavier

	* Backport of 14931 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14931 | xavier | 2015-12-30 18:37:35 -0800 (Wed, 30 Dec 2015) |
	  5 lines
	  
	  Simplified the initialization of the mesh communicator. An MPI
	  group
	  object is now always required. This fixes an issue where an
	  unconsistent mesh could be generated.
	  
	  
	  ------------------------------------------------------------------------

2016-01-09 20:09  xavier

	* Backport of 14930 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14930 | xavier | 2015-12-30 17:34:07 -0800 (Wed, 30 Dec 2015) |
	  4 lines
	  
	  Bugfix: the communicator for the fine mesh was initialized to
	  MPI_COMM_WORLD when domain parallelization was not used.
	  
	  
	  ------------------------------------------------------------------------

2016-01-09 19:53  xavier

	* Backport of 14953 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14953 | xavier | 2016-01-02 17:03:41 -0800 (Sat, 02 Jan 2016) |
	  5 lines
	  
	  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.
	  
	  
	  ------------------------------------------------------------------------

2016-01-09 19:46  xavier

	* Backport of 14967 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14967 | xavier | 2016-01-03 16:54:38 -0800 (Sun, 03 Jan 2016) |
	  3 lines
	  
	  Bugfix: An argument had the wrong intent, this was causing some
	  test failures.
	  
	  
	  ------------------------------------------------------------------------

2016-01-09 19:43  xavier

	* Backport of 14969 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14969 | xavier | 2016-01-03 20:10:46 -0800 (Sun, 03 Jan 2016) |
	  5 lines
	  
	  Bugfix: The sparskit and arpack macros tried to link the
	  libraries
	  given as argument even if the detection failed.
	  
	  
	  
	  ------------------------------------------------------------------------

2016-01-09 17:32  xavier

	* Replaced // with + in lines that are preprocessed. With this
	  5.0.x
	  compiles with gcc 5.

2016-01-09 17:22  xavier

	* Backport of r14515 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14515 | xavier | 2015-08-26 11:53:07 -0700 (Wed, 26 Aug 2015) |
	  3 lines
	  
	  Support for the trunk version of libxc.
	  
	  
	  ------------------------------------------------------------------------

2016-01-09 16:46  xavier

	* Backport of 14378 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14378 | dstrubbe | 2015-06-25 12:54:18 -0700 (Thu, 25 Jun 2015)
	  | 2
	  lines
	  
	  Use our new '+' string concatenation operator for the PUSH_SUB
	  and POP_SUB macros.
	  
	  ------------------------------------------------------------------------

2016-01-09 16:45  xavier

	* This is the first of several changes to get 5.0.x to compile with
	  the
	  preprocessor of gcc 5. The changes are larger that I would like,
	  but I
	  this code has been used for months in trunk without problems.
	  
	  Backport of 14361 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14361 | xavier | 2015-06-24 09:26:28 -0700 (Wed, 24 Jun 2015) |
	  4 lines
	  
	  Proof of concept to overload the + operator to concatenate
	  strings.
	  This works
	  with gfortran 4.9.2. Let's see what other compilers think.
	  
	  Backport of 14364 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14364 | dstrubbe | 2015-06-24 12:27:34 -0700 (Wed, 24 Jun 2015)
	  | 2
	  lines
	  
	  Fix to r14361, which did not compile with g95 and path. A version
	  of
	  the string concatenation function that uses more normal syntax
	  and is
	  definitely accepted by g95. Also, this one doesn't dynamically
	  allocate memory, so it shouldn't create memory leaks.

2016-01-08 18:49  dstrubbe

	* Backport of 15008 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r15008 | dstrubbe | 2016-01-07 17:13:41 -0800 (Thu, 07 Jan 2016)
	  | 2 lines
	  
	  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.
	  
	  ------------------------------------------------------------------------

2016-01-06 15:44  xavier

	* Backport of the trunk preprocessor macro to 5.0.x. This allows
	  Octopus
	  to be compiled with gcc 5.

2015-11-08 17:48  dstrubbe

	* Backport of 14752 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14752 | dstrubbe | 2015-11-07 23:04:12 -0500 (Sat, 07 Nov 2015)
	  | 2 lines
	  
	  Bugfix: configure prefix for BerkeleyGW now works with an
	  installed version (e.g. from a package) as well as from the
	  source location.
	  
	  ------------------------------------------------------------------------

2015-11-07 20:39  dstrubbe

	* Bugfix, needed only for 5.0.x: an extra blank line is needed in
	  output for skipping a test, to be sure the reason for it appears
	  in the output from 'make check'.

2015-11-06 15:48  fnog

	* Backport of r14740: Fixing autotools logical tests in order to
	  accept gsl 2.0.

2015-11-01 16:29  dstrubbe

	* Backport of 14713 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14713 | dstrubbe | 2015-10-29 21:58:23 -0400 (Thu, 29 Oct 2015)
	  | 3 lines
	  
	  * Bugfix to r14659: basically, undoing most of that change; what
	  was really needed was just to set 'acx_external_metis=no'. My
	  change mistakenly had prevented enabling external METIS at all.
	  Now the three cases properly work: --with-metis-prefix=no, a
	  prefix, or unset.
	  * Bugfix: libmetis.a was not removed by 'make clean'.
	  
	  ------------------------------------------------------------------------

2015-10-09 20:10  dstrubbe

	* Backport of 14659 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14659 | dstrubbe | 2015-10-09 10:17:20 -0400 (Fri, 09 Oct 2015)
	  | 2 lines
	  
	  Bugfix: "./configure --with-metis-prefix=no" would lead to
	  neither the internal nor an external METIS being used, yet the
	  octopus code supporting metis being used anyway, and then failing
	  to link. Some confusion between "disabled" and "no" is now sorted
	  out.
	  
	  ------------------------------------------------------------------------

2015-10-09 14:34  dstrubbe

	* Bugfix to documentation in 5.0.x: the help string for the
	  configure option --disable-utils said the opposite of what was
	  meant.

2015-10-07 21:03  dstrubbe

	* Recentering test value for 5.0.x which failed on Stampede with
	  discrepancy 0.0215 or so.

2015-10-07 19:55  dstrubbe

	* Backport of 14653 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14653 | dstrubbe | 2015-10-07 14:39:01 -0500 (Wed, 07 Oct 2015)
	  | 2 lines
	  
	  Restore proper date, accidentally changed in r14640.
	  
	  ------------------------------------------------------------------------

2015-10-07 10:13  micael

	* Backport of 14651 to branches/5.0.x:
	  ------------------------------------------------------------------------
	  r14651 | micael | 2015-10-07 12:13:23 +0200 (Wed, 07 Oct 2015) |
	  2 lines
	  
	  Added 5.0.0 release.
	  
	  ------------------------------------------------------------------------

