[petsc-dev] git : list branches started off maint - but not yet merged back

Satish Balay balay at mcs.anl.gov
Fri Aug 30 00:06:50 CDT 2013

On Thu, 29 Aug 2013, Jed Brown wrote:

> Sorry, here's an update.


> > Candidate for maint: refs/remotes/origin/jed/fix-gamg-coarse Wed May 29 22:29:42 2013 -0500
> > 248bfaf PCBJacobi: call KSPSetFromOptions(subksp) when PCSetUp was called first
> > 5b42dca PCGAMG: set bjacobi->same_local_solves=TRUE to clean coarse grid viewing
> I think this can merge.
> > Candidate for maint: refs/remotes/origin/jed/fix-viewer-ascii-tab-whitespace Wed Jun 19 09:19:55 2013 +0200
> > 17c4b54 Viewer ASCII: print tabs to petsc_history, not fd
> Should merge.  (Bitbucket is down so I'm not doing it now.  You're
> welcome to.)
> > Candidate for maint: refs/remotes/origin/jed/is-view-from-options Sun Aug 25 01:10:01 2013 -0500
> > 817ea41 AO: fix erroneous processing of -ao_view and factor into AOViewFromOptions
> > dba0e79 doc: fix named argument in {Vec,Mat,DM}ViewFromOptions
> Should merge.
> > Candidate for maint: refs/remotes/origin/jed/mat-factor-shift Sat Jun 1 12:23:49 2013 -0500
> > 5e9742b Mat PCFactor: include MatFactorShiftType in PCView_Factor output
> I don't care whether this merges.  It changes output which is
> unattractive, but it's harmless and the output is more useful.
> > Candidate for maint: refs/remotes/origin/jed/pcmg-residual-underscore Wed Jun 26 16:22:39 2013 -0500
> > 21b6e91 PCMGResidualDefault: fix fortran def [error on windows build]
> > 54b2cd4 PCMGResidualDefault: deprecate old semi-private name and make public
> NOT a candidate.  The first of three commits went to 'maint', but the
> others should not.
> > Candidate for maint: refs/remotes/origin/toby/ksp-dgmres-view Tue Jun 4 06:20:21 2013 -0500
> > e788a01 dgmres: doc for -ksp_dgmres_view_deflation_vecs
> > 9d49465 dgmres: added option to view deflation vecs after solve.
> > 62aa035 dgmres: fix real typed as int
> Okay to merge.
> > Candidate for maint: refs/remotes/origin/toby/pcfactorgetlevels Fri Jun 7 14:20:33 2013 -0500
> > 2591b31 PCFactor: add PCFactorGetLevels
> Okay to merge, but PCDestroy_XXX should also be _clearing_ out its
> composed functions.  None of them are cleared here, so it's okay that
> this commit doesn't clear its.
> > Candidate for maint: refs/remotes/origin/toby/pcgamg-rebuild-coarse Wed Jul 24 16:12:18 2013 -0500
> > 0a97e77 PCGAMG: Always rebuild coarse Mat the first time
> I haven't heard about any bugs related to this, so I think it should be
> merged.

thanks, Merged the 'ok'd  ones.

Here are the two outstanding ones [removing the NAK one - and the ones starting at or prior to v3.4]



$ candidate-branches master maint
Candidate for maint: refs/remotes/origin/barry/petscdatatypefromstring Fri Aug 23 21:12:13 2013 -0500
8e4b2d1 Sys: add PetscDataTypeFromString() and test code

Candidate for maint: refs/remotes/origin/knepley/fix-configure-package Wed Jul 17 07:53:11 2013 -0500
506c33d Configure: Fixed logic when using --with-package-lib when the package has no includes


Candidate for maint: refs/remotes/origin/jed/pcmg-residual-underscore Wed Jun 26 16:22:39 2013 -0500
21b6e91 PCMGResidualDefault: fix fortran def [error on windows build]
54b2cd4 PCMGResidualDefault: deprecate old semi-private name and make public

** NACKed **

Candidate for maint: refs/remotes/origin/karlrupp/feature-viennacl Mon Jun 10 16:44:57 2013 -0500
4cf1874 ViennaCL: Added synchronization to bindings just as for CUSP in order to allow for correct results with log_summary
7121bf1 Added custom FORTRAN stubs for MatCreateAIJViennaCL and MatCreateSeqAIJViennaCL
023073b Corrected comments in ViennaCL bindings.
2f39c30 Improved performance of VecMAXPY_SeqViennaCL and VecAXPBYPCZ_SeqViennaCL for gamma=1.
65e3cb3 Adjusted parsing of ViennaCL device type options.
5e5cd3c Adjusted example config file for ViennaCL
50de6e9 Fixed type warning when passing PetscScalar (*)[2] to MPI_Allreduce().
56f911c Fixed detection of OpenCL on Mac OS.
824c3e1 Fixed memory leak in VecGetArray()/VecGetArrayRead() for ViennaCL bindings.
4f97bae ViennaCL: fix LOCDIR
ab6435e ViennaCL: Updated code for new interface to PetscObjectComposeFunction() and XXXRegister()
8ff5867 ViennaCL: Updated to new AMS viewer model
586241c ViennaCL: Fixed includedir in viennacl.py package handler
0d73d53 ViennaCL: Applied const-correctness to functions.
19fddfa ViennaCL: Eliminated whitespace violations, removed/fixed erroneous comments.
b4986ff ViennaCL: Device type can now be specified via command line.
4076e18 ViennaCL: Now catching the correct type of exceptions (std::exception)
0a9f7e6 ViennaCL: Eliminated dead/unused code
67c87b7 ViennaCL: Fixed issues if size 0 is passed.
3db98f8 ViennaCL: Setting VIENNACL_WITH_OPENCL flag to enable OpenCL backend as default.
dcc04cf ViennaCL: Added package detector for OpenCL, added OpenCL-dependency for ViennaCL.
8f86e40 ViennaCL: Added MPI-implementation for AIJ matrices.
9f10968 ViennaCL: Removed seqviennacl/aijassemble.cxx (would trigger error, now uses existing fallback)
e4a0ef1 ViennaCL: Renamed files to fulfill conventions (.cpp -> .cxx, all lowercase), removed restriction of C++ everywhere.
d67ff14 ViennaCL: Added sequential AIJ matrix.
32d0529 ViennaCL: Added MPI vectors.
ca5856b ViennaCL: Removed unnecessary call to resize() at vector construction.
3f559ed ViennaCL: Download of package now working.
b17c682 ViennaCL: First draft, including all operations for sequential vectors.

*** Development from 3.4 *******

Candidate for maint: refs/remotes/origin/knepley/fix-mat-chop Wed May 29 06:46:26 2013 -0500
5a4c130 Mat: fix PetscObjectComm call in last commit
e22710e Mat: Remove PETSC_COMM_WORLD from MatChop()

** Starts from 3.3 branch so old master [and not maint] - and an ABI change.

Candidate for maint: refs/remotes/origin/knepley/reordering Tue Aug 27 22:39:10 2013 -0500
aa932e7 Mat ex170: Comments
0e8f585 Mat ex170: Test for MatMult() using max instead of plus - Finds the number of connected components in parallel - Can still optimize better in parallel
b434eb9 Mat: Added stuff to let me do (max, mult) algebra things for reordering - Added MatMultMax_SeqAIJ() and MatMultAddMax_SeqAIJ() - Added PetscSparseDenseMaxDot()
849b11c Vec: Added VecUniqueEntries() - This is purely diagnostic, but I think its useful for tests
745b41b Sys: Added PetscSortRemoveDupsReal()

** Starts from 3.3 branch so old master [and not maint]? - looks like development changes.

Candidate for maint: refs/remotes/origin/prbrune/pcgamg-classical Thu Jun 27 12:30:19 2013 -0500
9d2699c PCGAMG Classical: no need to explicitly zero formkktprol
65b3d5b Fixed types for PCGAMGProlongator_Classical and PCGAMGGraph_Classical, strict C fix.
6774142 Complex fixes for classical
30454c4 Added test examples for classical AMG
e5a0faa Fixed a number of bugs in Classical GAMG
b375234 Fix to the threshold calculation
f82387e Fixed PCGAMG_Classical so that thresholding decreases fill as bugs are fixed
3c9ab2c Fixed one-processor memory corruption bug for classical GAMG.
bfde193 Namespacing in gamg/classical.c
8e6d0c3 Added classical AMG to GAMG

** started from 3.3 - and looks like development **

Candidate for maint: refs/remotes/origin/prbrune/snes-npcjacobianlagging Tue May 14 09:50:15 2013 -0500
e514fbc Tests for persistent lagging of the Jacobian or Preconditioner
9d7e2de Suggested documentation fixes for SNESSetLagJacobian/PreconditionerPersists()
971e163 SNESSetCountersReset() lets SNES to keep nfuncs and linear_its for many solves
902f982 Control over initial SNES func/norm in one place; inner ASPIN fix.
25a90ac Fixed the behavior of left-pc SNESes preconditioning right-pc SNESes.
37ec4e1 Added SNESGet/SetLagJacobian/PreconditionerPersists()

** started from 3.3 - and looks like development **

More information about the petsc-dev mailing list