From 24 Nov 2016 to 23 Dec 2016
23 Dec 2016
- 14:08 Refactoring #1744 (Sprint): resolve mutual directory dependence Aggregate <-> Multilayer
- 14:07 Refactoring #1744 (Rejected): resolve mutual directory dependence Aggregate <-> Multilayer
- Recent refactoring (introducing FormFactorCoherentPart) broke the directed-graph structure of Core subdirectories: Ag...
- 14:08 Refactoring #1743 (Sprint): resolve mutual directory dependence Parametrization <-> Scattering
- 14:05 Refactoring #1743 (Resolved): resolve mutual directory dependence Parametrization <-> Scattering
- Recent refactoring (introducing INode) broke the directed-graph structure of Core subdirectories: Parametrization/ an...
20 Dec 2016
- 15:17 Documentation #1742 (Rejected): Website should provide link to GitHub issue tracker. Close forum ...
- The Website should invite users to submit bug reports to the GitHub issue tracker.
The forum entitled »Bug reports...
19 Dec 2016
- 20:11 Feature #1741 (Rejected): Wavelength/radiatio independent way to define material parameters
- It would be helpful and for time of flight modeling necessary to have other possibilities to define the material para...
- 20:09 Bug #1636: Unexpected wavelength depent intensity Roughness vs. ParticleLayout
- I will have to test this again, but I think I found my mistake now. I did not define the materials correctly using SL...
16 Dec 2016
- 11:29 Feature #1740 (Resolved): Implement polarized scattering on magnetic nanoparticles using the magn...
- Following some email conversations with A. Glavic, it has become clear that we need to calculate the polarized scatte...
12 Dec 2016
- 17:23 Refactoring #1732: Optimization in innermost loop in RT computation
- Solved in 2baf9d, pull request 122.
- 15:15 Refactoring #1732: Optimization in innermost loop in RT computation
- Implemented in branch bisectRT+test.
- 08:48 Refactoring #1732 (Rejected): Optimization in innermost loop in RT computation
- SpecularMatrix::calculateUpFromLayer contains an innermost loop.
For thick multilayer system with few scattering lay...
- 17:19 Refactoring #1735 (Resolved): remove include guards from unit tests
- Resolved in 38845bca.
- 13:08 Refactoring #1735: remove include guards from unit tests
- -> pull request 121
- 09:43 Refactoring #1735 (Resolved): remove include guards from unit tests
- There is no reason why a unit test h file should be included twice, therefore there is no reason to have include guar...
- 16:40 Feature #1738 (Rejected): Rename files and classes in Multilayer to resolve ambiguity about "Matrix"
- 13:08 Bug #1734 (Resolved): Investigate numerical stability of multilayer RT computation
- Cannot confirm.
Tests supplied in pull request 121.
- 09:12 Bug #1734 (Resolved): Investigate numerical stability of multilayer RT computation
- Details will follow. I'm investigating.
- 09:15 Feature #1103: extend simulation to transmission geometry (detector images with z<0)
- Probably this requires a renaming of variables or functions at some places: R and T will become Up and Down.
- 09:06 Feature #1706 (Resolved): Profile simulation with multilayers containing large amount of layers
- For proposed solution see #1732 and #1733.
- 09:02 Refactoring #1733 (Resolved): Avoid multiple calculations of same RT coefficients
- Currently, SpecularMatrix::calculateUpFromLayer is called within loops...
- 08:44 Refactoring #1673 (Resolved): Design proposal for all subtasks under #1645
- No comments received, so proposal is accepted.
09 Dec 2016
- 14:05 Feature #31: Implement x-ray polarization factor
- For a multilayer, it's not just one polarization factor.
Rather, polarization must be taken into account at each int...
07 Dec 2016
- 15:50 Feature #1670 (Sprint): Provide interface for adding particle shapes in different layers that are...
- 14:54 Refactoring #1731 (Backlog): Reduce number of TODOs in code
- Currently counting 105.
- 14:01 Configuration #1496 (Rejected): Switch Windows10 buildslave from virtual box to vagrant box
- Currently, the VM is always on (thanks to Jan). This should resolve many of the problems we experience with the Windo...
- 12:33 Refactoring #1730 (Sprint): Refactor ICompositeSample hierarchy
- Then I start slowly via focused pull requests to have only few changes at a time.
- 10:39 Refactoring #1730 (Rfc): Refactor ICompositeSample hierarchy
- 12:13 Configuration #1729 (Resolved): Fix centos7 CI build on buildbot
- 10:52 Bug #1641 (Resolved): histogram2d test fails under certain configurations
Test was fixed by requiring numpy array to be of type dtype=float64.
Also we've discovered, that there is no 'a...
- 10:35 Bug #1674 (Resolved): Possible bug introduced in Polygon in Python context
06 Dec 2016
- 17:44 Refactoring #1730 (Resolved): Refactor ICompositeSample hierarchy
- To implement #1624 (#1562 is also related), we might have to provide the following
* Turn lots of ISample into ICo...
- 17:21 Feature #1667 (Resolved): Read DESY/dpdak tiff files to BornAgain GUI
- TiffHandler now reads TIFF files with IEEE float sample format.
- 16:54 Configuration #1729 (Resolved): Fix centos7 CI build on buildbot
- 15:04 Configuration #1544: require libcerf instead of providing ThirdParty/Fadeeva
- It might be advantageous to first provide a C++ version of the library (using C++ std::complex instead of C complex t...
- 12:39 Refactoring #1659 (Resolved): rename PixelMap -> Pixel
- 10:49 Refactoring #1407 (Rejected): rotate the sample, not Q
- Not worth a separate refactoring, but will be done when implementing graded layers, as explained in design proposal.
05 Dec 2016
- 17:27 Refactoring #1659 (Sprint): rename PixelMap -> Pixel
- 17:26 Feature #1669 (Resolved): Allow for multiple form factors (dwba or not) in FromFactorWrapper and ...
- 17:25 Bug #1705 (Rejected): Py Fit crashes for zero fit parameters
- Duplicate of #1664, which is resolved.
- 17:15 Bug #1633 (Resolved): Possible memory leakage in LayerStrategyBuilder.
- 12:35 Feature #1680 (Resolved): Provide 1.7.1 hotfix
- 09:33 Feature #1722 (Long Term Idea): self-absorption correction
- Optional correction for extinction along the incoming and outgoing ray, due to scattering.
- 09:30 Feature #1721 (Rejected): Scattered intensity should depend on beam size and sample size
- The 1/sin(alpha_i) dependence of the scattered intensity is only correct if the beam footprint is smaller than the sa...
02 Dec 2016
- 20:38 Refactoring #1693: Make consequential use of default values in initializer lists.
- Particle also has four constructors (but may be more onerous to refactor):...
- 20:35 Refactoring #1720 (Rejected): Vacuum() instead of HomogeneousMaterial("Air", 0.0, 0.0)
- For accuracy, let's not call it air if the refractive index is zero. Change all examples.
For convenience, provide...
- 19:32 Refactoring #1673: Design proposal for all subtasks under #1645
- started writing, gisas/internal_reports/1611_design_gradedLayers.tex
- 15:33 Documentation #1719 (Rejected): Manual: avoid the word "matrix" for embedding, ambient bulk material
- at few places, confusion with mathematical matrix is not entirely impossible
- 15:15 Configuration #1718 (New): Provide Windows installer with debug info in it
Might be an intermediate solution on the way to crash handling.
* We already know how to compile under Windows i...
- 14:57 Documentation #1523 (Resolved): explain numeric value of HomogeneousMagneticMaterial::m_magnetic_...
- Added comment in my current feature branch explaining the numeric value of this prefactor.
- 14:30 Documentation #1523 (Sprint): explain numeric value of HomogeneousMagneticMaterial::m_magnetic_pr...
- 14:46 Bug #1717 (Rejected): GISASSimulation: setBeamParameters and setDetectorParameters have opposite ...
- For each choice I can guess a reason. Nonetheless it's confusing and potentially misleading.
One solution could be...
- 14:41 Feature #1716 (New): GISASSimulation::setBeamParameter: default value 0 for third parameter phi_i
- and simplify examples accordingly
- 14:02 Feature #1715 (Rejected): start sample construction with SlabSample() or FilmSample()
- The semi-infinite bottom and top layers are special. This should be reflected in the API.
- 13:40 Feature #1714 (Rejected): == Enhance functionality ==
- 13:39 Refactoring #1581 (Backlog): get... -> create... if new pointer is returned
- 13:37 Refactoring #1713 (Rejected): light refactoring proposals
- While reading code. Please ignore this for the moment, as I will add more.
* "override" -> "final" in FormFactorDW...
- 13:36 Feature #1712 (Rejected): == Code maintenance ==
- Refactorings that leave functionality invariant. To be done preferably at the beginning of a sprint.
01 Dec 2016
- 17:41 Documentation #1710 (New): Manual: explain particle density
- 17:03 Bug #1633 (Sprint): Possible memory leakage in LayerStrategyBuilder.
- 16:48 Refactoring #1707 (Resolved): refactor LayerStrategyBuilder, DecoratedLayerComputation -> LayoutS...
- These two classes are concerned with a single Layout, not with all layouts in a layer. Therefore:
* Rename them
- 16:43 Feature #1706 (Resolved): Profile simulation with multilayers containing large amount of layers
- Simulation of multilayer structures with more then 30 layers is very slow. It is would be very helpful to find a poss...
- 16:05 Refactoring #1549 (Rejected): rename class FitSuite
- will come up automatically with #1469, no need for action till then
- 15:09 Refactoring #1469 (Long Term Idea): Package Multifit-Suite as separate library
- 14:41 Feature #1653 (Resolved): Outcome of BornAgain user meetings in November
- Resolved into issues #1694 - #1705. Transcribed everything except for
* IPython matplotlib problem (if somebodey can...
- 14:40 Bug #1705 (Sprint): Py Fit crashes for zero fit parameters
- Assigned to sprint since it sounds easy and important.
- 14:40 Bug #1705 (Rejected): Py Fit crashes for zero fit parameters
- If no fit parameters are defined, Python fit crashes. On Windows, it does this without any clear message. So we need ...
- 14:38 Feature #1704 (New): Revise chi squared module
- Provide examples/functional tests to create custom chi2 module from python. Provide log10, sqrt amplitude function in...
- 14:37 Feature #1703 (Rejected): GUI: provide convenient way to set min,max for relative difference map.
- And probably all other properties of plot on FitView
- 14:35 Feature #1702 (Resolved): Fit: Provide extended print output for fit parameters during iterations.
- Artur pointed out, that we do not have a way to see how many sample parameters are steered by given fit parameter.
- 14:34 Feature #1701 (Rejected): import real-space sample configuration, e.g. from molecular dynamics (m...
- 14:32 Refactoring #1700 (Rejected): Implement FitSuite::printParameters (or ::parametersToString.)
- Now GISASSimulation::printParameters() reports parameter names which are different from what FitSuite has....
- 14:31 Feature #1699 (Rejected): functional tests: Improve MinimizerCatalogue::toString
- While preparing advanced fitting tutorial realized that the existing MinimizerCatalogue::toString provide too little ...
- 14:14 Bug #1681 (Resolved): Fitting GUI: free parameter error
- 13:55 Bug #1681 (Sprint): Fitting GUI: free parameter error
- 14:07 Feature #1698 (New): GUI: save/load detector mask
- 14:06 Feature #1697 (New): Save/load detector mask
- Define data format, and implement save and load functions for a detector mask.
- 14:04 Feature #1696 (Sprint): GUI: Improve drag-and-drop construction of a MultiLayer.
- I suspect there is a deeper reason for the observed problem, namely the mixed nature of the current sample constructi...
- 13:53 Feature #1696 (Resolved): GUI: Improve drag-and-drop construction of a MultiLayer.
- At the Hamburg and Garching tutorials, many users have been experiencing problems while dropping Layer on top of a Mu...
- 13:52 Bug #1661 (Resolved): FitSuite doesn't know about TestMinimizer
- Fixed, but without providing functional test. Later will be created as separate issue
- 13:51 Feature #1082: GUI: provide unddo/redo functionality
- This can be done first for sample construction. I think that we have to implement the full fledged "Command" pattern....
- 13:44 Feature #1695 (Rejected): GUI: Implement crash handler, and restoration of state before crash
- During Hamburg and Garching tutorials, there were quite some number of GUI's crashes (about 10-20 per 8 hours of tuto...
- 13:37 Feature #1694 (Rejected): GUI: Provide a way to save fitted image into the file.
- In FitView there is no any button (or right mouse click action) to save fitted image into the file.
- 13:16 Configuration #1635 (Resolved): update to gtest-1.8
- resolved in 9de8f297de
- 13:10 Bug #1674: Possible bug introduced in Polygon in Python context
- Fixed in hotfix. Unit tests still have to be created.
- 13:09 Bug #1662 (Resolved): CMake: require Qt version 5.4
- 13:09 Bug #1692 (Resolved): GUI and Py: simulation crash when sample contains undefined material
- 10:43 Bug #1692: GUI and Py: simulation crash when sample contains undefined material
- Python problem is different one (which I was not aware of, thanks). Should be solved in different way, than in GUI. B...
- 10:33 Bug #1692: GUI and Py: simulation crash when sample contains undefined material
- It's not only a GUI problem.
Take any python example, and change the line...
- 09:33 Bug #1692 (Resolved): GUI and Py: simulation crash when sample contains undefined material
How to reproduce:
* Drag and drop sample from the list of standard samples
* Go to material editor ('m'), remov...
- 13:08 Bug #1658 (Resolved): add qt5-svg as a dependency
- 12:58 Bug #1639 (Resolved): GUI: crash if fitting parameter removed
- 11:46 Refactoring #1693: Make consequential use of default values in initializer lists.
- I think such sequence of constructors with growing list of parameters was because of boost-python bindings. Probably ...
- 11:13 Refactoring #1693 (Backlog): Make consequential use of default values in initializer lists.
- For example, in ParticleLayout.h we currently have three constructors, and four versions of @addParticle@:...
30 Nov 2016
- 13:04 Bug #1689 (Resolved): memory leak in ParameterSample
- 09:43 Bug #1689: memory leak in ParameterSample
- I believe that due to API changes @wl_distr.generateSamples(nsamples, xmin, xmax)@ should instead be something like @...
- 11:07 Feature #1691 (Resolved): Decrease number of warnings produced by Windows build
Currently number of Windows warnings is tremendous, makes it difficult to analyze logs.
* In Release 1.8 Appveyo...
29 Nov 2016
- 16:44 Bug #1689: memory leak in ParameterSample
- It appears that the template @std::vector< ParameterSample > >@ is not being exported by swig. So it should be a simp...
- 16:28 Bug #1689 (Sprint): memory leak in ParameterSample
- 16:14 Bug #1689 (Resolved): memory leak in ParameterSample
- The following simple script...
- 15:03 Documentation #1141: Manual: describe multi layers
- will come back to it later
- 11:38 Bug #1658: add qt5-svg as a dependency
- Resolved in f1de40e, pull request #93.f1de40e
Can be transfered to hotfix branch using git cherry-pick f1de40e.
- 11:22 Bug #1662: CMake: require Qt version 5.4
- In hotfix branch, apply git cherry-pick b4f148
- 11:20 Bug #1662: CMake: require Qt version 5.4
- Solved in b4f14 (pull request #93).
- 10:40 Configuration #1615: in each third-party source directory, provide information about used downloa...
- The information file shall be called 00_PROVENANCE.
Work on this issue started with gtest-1.8.0.
- 10:39 Configuration #1635: update to gtest-1.8
- Done in 2ff998fba, pull request #92.
- 09:37 Refactoring #1660 (Resolved): remove global functions
- Learned from Walter that not all global functions are bad. Unnecessary globals are mostly removed with pull request #88.
- 09:27 Refactoring #1479 (Rejected): presence of some boost components not checked by cmake
- Rejected for two reasons:
* CMake and Boost are changing so quickly that there is no way for us to provide a lasting...
- 09:00 Feature #1688 (Rejected): == Roughness ==
28 Nov 2016
- 17:06 Bug #1687 (Resolved): LMA is incorrectly normalized
- Fixed in hotfix-1.7.1 branch
- 15:45 Bug #1687 (Resolved): LMA is incorrectly normalized
- When two layouts are attached to one layer, relative abundances are scaled with respect to the sum of all particles, ...
- 17:05 Refactoring #1456 (Rejected): Automatize update of Py Examples
- automatizing is not an aim in itself
- 14:51 Feature #1686 (Rejected): == Release procedures and pre-release actions ==
- 14:32 Configuration #1665 (Rejected): Take swig-generated Python wrappers out of git repository
- Agreed. Restriction of the Python API now requested by #1685.
- 14:07 Configuration #1665: Take swig-generated Python wrappers out of git repository
- Another solution could be to severely limit the public API we export to Python. This would result in regeneration of ...
- 14:31 Refactoring #1685 (New): Restrict Python API to what the user needs.
- The user constructs models, runs simulation, and retrieves results. Nothing else needs to be supported by the Python ...
- 14:25 Feature #1684 (Resolved): nicer layout for firstname.lastname@example.org
- #1238 was archived too early; the announcement of 1.7 still came with no left margin and a fat black top bar.
- 14:16 Feature #1683 (Resolved): GUI: provide persistence of plot settings (interpolation, color scale, ...
- See the request on BornAgain forum:
- 14:13 Refactoring #1604 (Resolved): move expected inaccuracy (m_variability) out of class OutputData
- Resolved in 5baab.
For standard simulations, the expected inaccuracy is hard-coded in StandardSimulationsRegistry.cp...
- 09:22 Feature #1667: Read DESY/dpdak tiff files to BornAgain GUI
- I've got a user permission to upload the data file to redmine.
25 Nov 2016
- 13:32 Bug #1661: FitSuite doesn't know about TestMinimizer
- yes, it was requested and it is easy to fix
- 13:30 Bug #1661: FitSuite doesn't know about TestMinimizer
- hope I understand correctly that this is to go into the hotfix?
- 13:31 Bug #1682 (Resolved): hotfix: rm local version of FindBoost.cmake
- 13:24 Bug #1682 (Resolved): hotfix: rm local version of FindBoost.cmake
- cmake/generic/modules/FindBoost.cmake was created because cmake-3.6 did not recognize boost >= 1.62.
This problem is...
- 13:31 Bug #1679 (Resolved): Remove leading U+FEFF from FindYamlCpp05.cmake => needs hotfix
- 12:10 Bug #1679 (Resolved): Remove leading U+FEFF from FindYamlCpp05.cmake => needs hotfix
- Somehow, a leading unicode mark has slipped into FindYamlCpp05.cmake.
For at least two users, this caused CMake to a...
- 12:32 Bug #1681 (Resolved): Fitting GUI: free parameter error
- Setting fitting parameter as 'free' causes an error (see also the attached screenshot):...
- 12:19 Feature #1680 (Resolved): Provide 1.7.1 hotfix
- See subtasks.
Also it is absolutely necessary to run some Windows simulations next days and try to find the reason...
- 11:33 Feature #1677 (Backlog): Independent roughness cross-correlation for disjoint sets of interfaces
- To support Henrich, it is required to simulate resonator on top of the lipid bilayer system. Both, resonator and set ...
24 Nov 2016
- 18:02 Feature #1562: Provide ISample with possibility to export parameter tree in Python dictionary
- * std output generated by IParameterized::printParameters doesn't appear in ipython notebook
* all std::cout usage i...
- 18:00 Feature #1562 (Sprint): Provide ISample with possibility to export parameter tree in Python dicti...
- 17:57 Feature #1675 (Rejected): Get rid of ctypes.addressof in PySampleBuilder
As Jonathan suggested, we can get rid from ctype's dependency in Python by replacing it with its numpy alternative....
- 17:47 Refactoring #1623 (Sprint): Parameter name translation (GUI -> domain)
- 17:46 Feature #1624 (Sprint): Make all reasonable parameter values into fittable parameter
- 17:36 Refactoring #1673 (Sprint): Design proposal for all subtasks under #1645
- well, let's see whether and when we write up how the API will change ...
- 11:40 Refactoring #1673 (Resolved): Design proposal for all subtasks under #1645
- 17:33 Feature #1631 (Rejected): No impact of abundance optional argument in ParticleLayout addParticle ...
- Seems we agreed with Artur that it is a feature, not bug. Rejecting
- 17:32 Feature #1667 (Sprint): Read DESY/dpdak tiff files to BornAgain GUI
- 10:06 Feature #1667 (Resolved): Read DESY/dpdak tiff files to BornAgain GUI
- Some DESY tiff files are not loaded correctly in BornAgain GUI. File is loaded, but instead of the GISAXS pattern, us...
- 17:32 Refactoring #1604 (Sprint): move expected inaccuracy (m_variability) out of class OutputData
- 17:31 Feature #1617 (Sprint): Implement rotation of 2D OutputData on 90 deg
- 14:35 Bug #1658 (Sprint): add qt5-svg as a dependency
- 14:34 Bug #1674 (Resolved): Possible bug introduced in Polygon in Python context
Earlier version of Polygon had copying of input arrays instead of const reference.
This was because of crash if on...
- 14:21 Bug #1641 (Sprint): histogram2d test fails under certain configurations
- From the discussion with Jonathan and my wage remembering, it will be enough to use...
- 14:09 Testing #1609: Clarify relation between PyPersistence tests and tutorial examples
- May I also add that this refactoring could also be used to get rid of massive extra data in our repository and tarbal...
- 14:04 Testing #1609: Clarify relation between PyPersistence tests and tutorial examples
- Then I will summarize my proposal here.
* Move persistence check from "make check" to "make stresscheck"
- 11:31 Feature #1670 (Rfc): Provide interface for adding particle shapes in different layers that are to...
- 11:31 Feature #1670: Provide interface for adding particle shapes in different layers that are to be tr...
- Add one particle to several layers? Or add particles at Sample=Multilayer level, and let the internal machinery figur...
- 11:28 Feature #1670 (Resolved): Provide interface for adding particle shapes in different layers that a...
- 11:30 Feature #1672 (Resolved): Implement subdivision of a layer into multiple layers for graded interf...
- 11:29 Feature #1671 (Resolved): Provide automatic splitting of particle shapes when they cross a layer ...
- 11:27 Feature #1669 (Resolved): Allow for multiple form factors (dwba or not) in FromFactorWrapper and ...
- Since the dwba form factors contain the respective layer Fresnel coefficients, this would allow for a coherent treatm...
- 10:24 Bug #1664 (Resolved): FitSuite.runFit() crashes with unclear error message
- 09:57 Feature #1666 (New): Constructor of Histogram2D should accept also numpy arrays with dtype=int
- For the moment, constructor ...
Also available in: Atom