Merging develop branch into master #33
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The include guard was referring to the envelope geometry rather than the union geometry. This resulted in the union geometry being undefined when the envelope header was also included. (Error reported by Randy Taylor Dec. 2014)
Change rejection criterion for fat photons arising from DBS from iphatsrc=1 to iphatsrc>1. Now, with directional source biasing implemented in BEAMnrc, fat particles have 1 < iphatsrc <= NBRSPL, while thin particles have iphatsrc=1. (Bug and solution by Marc-Andre Renaud).
Macro intended to enforce a minimum step of 1e-5 cm provided this is < USTEP. Previous version allowed particles to 1) take vanishingly small steps near leaf boundaries (infinite loops) and 2) take a step > USTEP. Definition of macro now matches that in MLCE, which was corrected in 2009. Propagation of this bug must be due to the fact that SYNCMLCE was developed from an older version of MLCE. (Bug found by user Michael Martyn at National University of Ireland, Galway).
Log files for EGSnrc configuration (configure script) and user configuration (finalize_egs_foruser script) are now tagged with the configuration name and the user name, so that previous configuration logs are not overwritten.
Scan $EGS_HOME environment variable. Allow to reconfigure with a different configuration name inside an existing EGS_HOME directory, but in that case do not copy the user codes to prevent overwriting changes by the user. Constrain the user code compilation list to user codes actually found in EGS_HOME in case user has removed some of them.
…vacuum. If E or B field exists, then ustep in vacuum is limited to smaxir(irl), allowing first-order effects to be applied to charged particle's before they're allowed to step right through the vacuum region. Not the exact solution we're working towards but okay for now.
And fix an infinite-rerender regression that was created in the initial port. Several files were renamed due to the UI porting. I eliminated the XML declarations in the converted UI files, because older versions of Qt (4.2) try to interpret them as version declarations.
On a ~1300 region file of mine, sometimes the region list in the display would overflow. It appeared that the "draw regions" while loop was continuing, even though reg > maxreg. Changing form to a for loop fixed the bug. The logic appeared sound, so I expect something was trampling on memory. In the process I corrected a `delete` to `delete[]` to comply with Address Sanitizer.
Port egs_view to Qt4
- Made clipping plane table fit within the widget view
- Improved visual layout of the viewcontrol widget
The interactive scene render parameters have changed slightly. For the scaling algorithm for each frame, we now use, not the past render time, but the past render time per pixel, to determine how much to scale the window down by. We also ignore the fixed time overhead due to the track rendering. Region picking has been moved into the ImageWindow class, and the region list rendering has been optimized slightly to avoid repeating the same work/checking number sizes too often.
Removed dead code and unneeeded variables in GeometryViewControl. The transformation code has been moved to the ImageWindow, since that reduces the message passing between that and the ViewControl.
In other words, give transformations priority over full-detail renders, so that when an interaction is requested, the full-detail render is aborted so that the faster, interactive rendering can start.
Note the removal of the XML declarations on the first lines of the UI files -- old Qt uic does not handle them properly.
Threaded egs_view
The change covered the else case of a branch in the dynamic render scaling section. This branch is taken when it takes less that 30ms to render a frame at full resolution; before this patch, when that happened, the requested view scaling was equal to the view size so that only the top-leftmost pixel was shown on screen. Included are changes that prevent the ImageWindow from rendering a scene before `render(...)` has been called at least once with an actual Geometry.
Fix egs_view thread porting regression where simple scenes black out
Take into account midpoint of cylinder when sampling within the shape. Remove unnecessary 'this'
Fix sampling of point within cylinder shape
The egslst file from BEAMnrc does not list the Bremsstrahlung cross section enhancement (BCSE) factors or if BCSE is even used. Modify the code slightly to output BCSE info immediately after the info about Directional Bremsstrahlung Splitting (DBS). (Submitted by Dave Rogers)
Some file systems are case-insensitive. For example OS X relies on HFS by default, which is case-insensitive. Hence when EGSnrc is cloned to a Mac, changes are automatically applied to the repository because of duplicate image file names (I6702 and i6702). Although not critical, it is annoying. Change file names to avoid conflicts.
This is a format-only commit. The egs++ source code before and after this commit compiles to identical object files. The coding style is not consistent throughout the egs++ library. Use the "Artistic Style" (astyle) code formatter to clean up the egs++ source code: make indentation uniform at 4 spaces, split compound statements, adjust braces and parenthesis placement and spacing, and indent preprocessor code blocks. It is better to do this in one swoop so as to avoid a large number of format-only commits polluting the code history later on. Add scripts/egs_style_standard file which contains the astyle options to apply to new code, or after editing existing code: astyle --options=$HEN_HOUSE/scripts/egs_style_standard "*.cpp" "*.h" For reference, the options used for astyle 2.05 for this commit were: --convert-tabs --indent=spaces=4 --style=attach --unpad-paren --pad-header --align-pointer=name --break-closing-brackets --add-brackets --indent-preproc-block
The C++ configuration script sed commands tacitly relied on a trailing space in empty variables in the EGS_CONFIG file. Add a "\?" switch in sed to match cases where the variable is not defined; otherwise one ends up working with the variable name instead of its value. This is the root cause of issue #20.
Fix bug in cshrc additions script
GUIs now use default fonts defined in .gui_defaults. Previously myDefaultFont, defined in beamnrc_gui.tcl and dosxyznrc_gui.tcl, overrode any other font definitions. Now myDefaultFonts is only enforced on main page of GUIs.
Prior versions of sprrznrc did not have this capability. Should automatically combine .egsdat files at end of run but the user can also run with IRESTART=parallel to manually combine these files. Also necessitated including egs_parallel.mortran in sprrznrc.make.
The changes in commit c67fc15 use a regular expression which is incompatible with the FreeBSD version of sed version distributed with OS X. Update with a regular expression which should be compatible across the board.
First line of these files changed from blank to a single "#". Subroutine egs_open_units chokes on blank lines and fails to read the rest of the .io file. The result was that some units are unassigned (eg unit 1 to the .egslst file), resulting in mysterious fort.* on output.
Warnings in dosrznrc.mortran, flurznrc.mortran, sprrznrc.mortran were due to these codes passing TITLE, an 80-element array of CHARACTER*1, to GRAPHTITLE, a CHARACTER*(*) variable in SUBROUTINE XVGRPLOT. Mismatch was eliminated by first copying TITLE into a local version of CHARACTER*80 variable, GRAPHTITLE, before passing to XVGRPLOT.
This folder contains the tools for packing the GUIs archive inside the egs_configure GUI.
Target is placed in $HEN_HOUSE/pieces/linux and library location made dependent to the LD_LIBRARY_PATH definition.
- Briefly describe Bielajew's EMF macros (emf_macros.mortran) - Describe how to turn on EMF transport in user codes - Explain how to turn on EADL atomic relaxations
- Use 1 instead of -IQ(NP) to convert E field from V/cm to 1/cm. - Use RM instead of hard-coded 0.511 for the electron rest mass.
ftessier
added a commit
that referenced
this pull request
Jan 21, 2016
Merging develop branch into master
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.