Activity
From 18 Apr 2016 to 17 May 2016
17 May 2016
- 14:13 Refactoring #1428 (Backlog): Compactify code that handles parameters (use abstract mechanism inst...
- agreed in meeting today
- 14:13 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- By convention, all lengths are in nm, all angles are internally in rad. Therefore, we only need the physical dimensio...
- 14:10 Refactoring #1430 (Resolved): Consider moving FormFactor unit tests to a single functional test
- The main concern will be resolved per issue #1433.
Renaming of subdirectories will follow later, when Core/ is res... - 13:59 Refactoring #1433 (Resolved): move massive numeric tests under 'make check' / 'ctest'
- to be done when the polyhedral FF is stable (#1432)
- 13:56 Refactoring #1432 (Rejected): final polish of polyhedral form factor
- * specific tests for frusta that are true pyramids
* exact treatment of vertices that coincide
* -improve edge sum ... - 11:42 Documentation #1431 (Resolved): simplify cross-referencing using package cleveref
13 May 2016
- 17:10 Feature #1238: nicer layout for newsletter@bornagainproject.org (reopened: v1.6.0 round mail stil...
- I really do not like the black topbar.
Plain ASCII would be nicer than this HTML. - 13:53 Refactoring #1424: reduce verbosity of googletest output
- Feature request submitted at https://github.com/google/googletest/issues/771.
12 May 2016
- 13:21 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- So I withdraw default units from the current proposal. Are units acceptable for you?
- 13:12 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- Maybe this issue should also be split: default values vs units
- 13:11 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- > > It's not their responsibility.
>
> We are free to assign responsibilities to our creatures as we deem appropri... - 13:01 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- So yes, you want users to get default parametrized items. OK, I accept.
> Core form factors should not have any de... - 12:38 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- In GUI default values should be provided to reduce amount of clicks/typing. This allows to start a simulation with a ...
- 11:59 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- The very first question then: Is it actually appropriate for the GUI to assign default parameters to newly created it...
- 11:43 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- The {FF}Items need default values because the GUI needs to create them without user initialization. There is no need ...
- 11:39 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- interesting consequence. why not?
- 10:57 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
> Each parameter should have a default value
What does it mean? "new FormFactorSphere()" becomes valid and becom...- 13:16 Refactoring #1430: Consider moving FormFactor unit tests to a single functional test
- The number 200k comes from arbitrary choices and can be changed if time is a concern.
The tests are concerned with... - 13:00 Refactoring #1430: Consider moving FormFactor unit tests to a single functional test
- I also agree that these tests conceptually do not belong to unit tests. It also obscures unit test coverage (having 2...
- 12:54 Refactoring #1430: Consider moving FormFactor unit tests to a single functional test
- My previous reply disappeared since you commented same record I was commenting.
> Why do we actually have a functi... - 12:42 Refactoring #1430: Consider moving FormFactor unit tests to a single functional test
- > [ RUN] OK is better than just [——]
Agree. Implemented in 91d8e8f5b0. - 12:36 Refactoring #1430: Consider moving FormFactor unit tests to a single functional test
- Nothing in googletest says it is limited to unit tests, let alone to unit tests in the most narrow sense. Why do we a...
- 12:31 Refactoring #1430: Consider moving FormFactor unit tests to a single functional test
- I put two issues in one, that's true. The major issue is to move Unit to Functional.
- 12:09 Refactoring #1430: Consider moving FormFactor unit tests to a single functional test
- These are two different issues:
* Move extensive ff tests from Unit to Functional?
* Don't split Core/Unit tests in... - 10:43 Refactoring #1430 (Resolved): Consider moving FormFactor unit tests to a single functional test
- I think that tests of form factor behaviour near singularity and along symmetry axes belongs to the functional tests,...
11 May 2016
- 18:28 Bug #1371: presence of some boost components not checked by cmake
- Btw: testing under Debian is somewhat difficult because of the non-standard packaging of Boost. Reported at https://b...
- 17:39 Bug #1371 (Resolved): presence of some boost components not checked by cmake
- Resolved in ac0d39e, using a patched FindBoost.cmake that is to be submitted to CMake: https://github.com/rleigh-dund...
- 18:10 Refactoring #1366: Revise boost libraries usage
- Migration of regex from Boost to regex is suspended because compilers are not ready. => commit d14b964d17bb
- 17:34 Refactoring #1366: Revise boost libraries usage
- Only now enacted the deletion of boost components regex and thread in cmake/modules/SearchInstalledSoftware.cmake.
- 14:36 Refactoring #1424 (Resolved): reduce verbosity of googletest output
- 14:36 Refactoring #1424: reduce verbosity of googletest output
- Solved in rather unexpected way in 24204e5, namely by brutally patching gtest. Inheriting outside gtest.cc from the d...
- 13:07 Bug #1429 (Resolved): Unit tests files are not shown in Qt Creator IDE
- Solved in 03d6c5d, using 'add_custom_target' as suggested in http://stackoverflow.com/questions/28384935/qtcreator-cm...
- 08:48 Bug #1429 (Archived): Unit tests files are not shown in Qt Creator IDE
- Files from UnitTests/TestCore directory (and similar) are not shown in "Project Structure" by Qt creator IDE. It is n...
- 12:39 Refactoring #1428: Compactify code that handles parameters (use abstract mechanism instead of tre...
- To automatically generate Python constructors,
* either use named arguments
* or make sure that parameters are orde...
10 May 2016
- 14:08 Refactoring #1428 (Archived): Compactify code that handles parameters (use abstract mechanism ins...
- Each parameter should have -a default value and- a -unit- physical dimension, assigned upon registration.
This wil... - 12:47 Documentation #1279 (Archived): Inkscape tasks
- 12:46 Documentation #1280 (Resolved): draw tetrahedron as seen from -x
- Completed in 4003c83.
04 May 2016
- 19:16 Refactoring #1424 (Sprint): reduce verbosity of googletest output
- with increasing number of tests, this is becoming urgent
- 14:19 Feature #1427: Implement new roughness calculation that is more stable for large roughness
- Resolved in #d212f6feba85d4a7f41db0b15ae90a697e1acd01
- 14:19 Feature #1427 (Resolved): Implement new roughness calculation that is more stable for large rough...
- 14:18 Feature #1427 (Archived): Implement new roughness calculation that is more stable for large rough...
29 Apr 2016
- 17:54 Refactoring #1366 (Resolved): Revise boost libraries usage
- Good progress has been made, thanks Jonathan. There is no need to get rid of _all_ of Boost.
- 17:54 Refactoring #1366 (Sprint): Revise boost libraries usage
- 17:48 Bug #1416 (Waiting): Core: FF: form factors with numeric integration fail for small and for large q
- waiting for student to work this out
- 17:17 Refactoring #1418 (Resolved): provide cotangent, and replace division by tan(alpha) in pyramids
- Done in 8c53f8f.
The remaining division by tan(alpha) in FormFactorCone shall be removed later, when this integratio... - 16:00 Bug #1420 (Resolved): Symmetry S2 not usable
- A factor i had been accidentally deleted. Now it's back working. Note however that computations with and without sym_...
28 Apr 2016
- 15:50 Refactoring #1424 (Archived): reduce verbosity of googletest output
- http://stackoverflow.com/questions/36847322/
- 12:00 Bug #1423 (Resolved): == eliminate warnings ==
- We should aim for zero warnings from the standard build process.
Current warnings
* From third-party C++ code => ... - 11:29 Documentation #875 (Archived): Provide content for http://gisaxs.com
- 11:28 Documentation #875 (Resolved): Provide content for http://gisaxs.com
- 11:28 Documentation #875 (Sprint): Provide content for http://gisaxs.com
- 10:23 Documentation #987 (Rejected): Tidy up the gisas/theory directory
- not a task to be done at once, but work in progress
- 10:15 Documentation #1262 (Resolved): Fix IsGISAXS references in form factor section
- Done in 0871f28. Hopefully, it's now clear where our computation deviates from IsGISAXS.
27 Apr 2016
- 13:41 Bug #1417 (Resolved): numeric inaccuracies in form factors
- Anisopyramid reimplemented using polyhedral ff; sym_S2 in Pyramid disabled until resolution of #1420
- 13:37 Refactoring #1422 (Resolved): provide clue when test fails
- * when reference test fails, indicate location of old and new image
* when python test fails, provide a clue what to... - 08:46 Bug #1420: Symmetry S2 not usable
- wrong results, therefore disabled in Prism6 and Pyramid
26 Apr 2016
- 17:41 Feature #975 (Backlog): man page should be built at release time
- 17:40 Documentation #1262 (Sprint): Fix IsGISAXS references in form factor section
- 17:39 Refactoring #1065 (Sprint): Revise particle form factors: change names, add few form factors, cha...
- agreed since long that this should be part of the next API-breaking sprint
- 17:20 Feature #1415 (Backlog): Parameterization: onChange() should invalidate cache, and not immediatel...
- 16:42 Bug #1421 (Resolved): FullSphere F(q)=0 for small q
- Done in c3568ef.
- 12:50 Bug #1421 (Resolved): FullSphere F(q)=0 for small q
- runff 1 0 11 1 0 0 0 0 0 1e-7 => (4.15786325694,0)
runff 1 0 11 1 0 0 0 0 0 1e-8 => (0,0) - 12:48 Bug #1420 (Resolved): Symmetry S2 not usable
- therefore the symmetry is not used until this is resolved
- 12:46 Bug #1294 (Resolved): provide substantial unit tests for factor computations
- Done in 6b09bc4.
We are now testing:
* whether F(0)=V
* continuity (comparing F(q) for small q with F(0))
* sym...
25 Apr 2016
- 16:06 Refactoring #1419 (Archived): move test code out of core
- qGenerator.h has been provisionally packed in Core/StandardSamples.
If I see correctly, this directory contains much... - 15:03 Bug #1294: provide substantial unit tests for factor computations
- Specialization tests implemented.
More tests needed. - 15:02 Refactoring #1418 (Resolved): provide cotangent, and replace division by tan(alpha) in pyramids
- somehow 1/tan(pi/2) works, but we should not rely on it
- 10:51 Bug #1417 (Resolved): numeric inaccuracies in form factors
- * According to FormFactorSpecializationTest:
** Pyramid and AnisoPyramid agree only for moderate values of |q|.
* -... - 09:09 Bug #1416 (Backlog): Core: FF: form factors with numeric integration fail for small and for large q
- * For small q: inaccurate results, as demonstrated by TruncatedSphere vs FullSphere in FormFactorSpecializationTests....
22 Apr 2016
- 15:17 Bug #1371 (Waiting): presence of some boost components not checked by cmake
- 14:21 Feature #1415 (Sprint): Parameterization: onChange() should invalidate cache, and not immediately...
- on advise by Walter, we go for option 1:
* each thread has it's own copy of sample parts, so threading will not be a... - 11:27 Feature #1415 (Rfc): Parameterization: onChange() should invalidate cache, and not immediately ex...
- 09:17 Feature #1415 (Backlog): Parameterization: onChange() should invalidate cache, and not immediatel...
- This is not only for elegance, but improves functionality: Currently, an exception is thrown whenever one parameter i...
- 13:08 Documentation #1102 (Backlog): Core: instrument simulation: review normalization, especially in v...
- 11:36 Refactoring #1410 (Resolved): rename GCCXML_SKIP_THIS
- Resolved in d163e2e.
- 11:28 Refactoring #1410: rename GCCXML_SKIP_THIS
- no, only the name is obsolete
21 Apr 2016
- 18:28 Documentation #1400 (Sprint): blender: paint dodecahedron and icosahedron
- I put it in the sprint, and when later assign to myself when in the 'blender' mood.
- 18:27 Bug #1411 (Rejected): FindGSL version test fails under cmake-3.0 => BornAgain does not build on D...
- Should be fixed by now.
- 18:26 Bug #1404 (Resolved): compilation broken for libgsl2
- I have updated minimizers and removed check for GSL version. It compiles with 1.16 and 2.1 without any additional cma...
- 17:59 Bug #1294 (Sprint): provide substantial unit tests for factor computations
- first, repair broken tests related to parameter pool
- 17:56 Support #852: Make GUI available at JCNS servers
- Not worth the fight - wait unless we have user contact
- 17:55 Bug #1413 (Resolved): polyhedra won't react to ex-post parameter variation
- Solved in 5f3e7a72: things are now precomputed upon parameter change. Broken tests will be restored under issue #1294.
- 16:57 Documentation #1412: avoid horizontal scrolling in code examples
- two things to do:
* decrease font size by one; determine maximum number of characters per line
* rework examples to... - 12:52 Bug #1370 (Resolved): Fix numerous "features" introduced by latest major GUI refactoring
- With this the GUI "ParameterizedItem" refactoring is over. Old functionality is restored.
* It's not possible to o...
19 Apr 2016
- 12:20 Support #852: Make GUI available at JCNS servers
- Answer from Josef Heinen 29mar16:
* BornAgain is installed on the Macs
* This seems not to include the Mac notebo... - 11:34 Documentation #1414 (Rejected): regenerate form factor images in manual
- - in k coordinates
- choose q along more and less directions
18 Apr 2016
- 20:05 Bug #1413 (Archived): polyhedra won't react to ex-post parameter variation
- needs to be corrected, and test mus be written
- 16:57 Documentation #1412 (Archived): avoid horizontal scrolling in code examples
- Some Python examples have some really long lines.
To read them, one needs to scroll horizontally.
But some examples... - 10:24 Bug #1411: FindGSL version test fails under cmake-3.0 => BornAgain does not build on Debian Jessie
- Let's not even attempt to repair the version testing. First resolve 1404, then remove the version testing from FindGSL.
- 09:48 Bug #1411 (Rejected): FindGSL version test fails under cmake-3.0 => BornAgain does not build on D...
- The failure seems to be related to the recent changes in the file cmake/modules/FindGSL.cmake
The error message:
... - 08:28 Feature #1399 (Resolved): Implement form factors for dodecahedron and icosahedron
Also available in: Atom