Notes:

Release Name: 0.2.12

Notes:
Next release of MRtrix, with the following key enhancements:

  • DICOM multi-frame support
  • enhanced ROI drawing tools in MRView (courtesy of Richard Beare)
Other bug fixes and enhancements
  • Clean compile on MacOSX 10.9
  • no need to set LD_LIBRARY_PATH or ld.so.conf (linker runpath is now set in executable)
  • build script now works under Python 2 & 3 without modification

Changes:
------------------------------------------------------------------------
r409 | jdtournier@gmail.com | 2014-01-27 22:10:31 +0000 (Mon, 27 Jan 2014) | 2 lines

remove unnecessary 'typename', prevented compilation with certain compilers (Debian in particular)
------------------------------------------------------------------------
r408 | jdtournier@gmail.com | 2014-01-23 22:36:24 +0000 (Thu, 23 Jan 2014) | 2 lines
streamtrack: -nomaskinterp also applies to seed regions
------------------------------------------------------------------------
r407 | jdtournier@gmail.com | 2014-01-15 12:26:27 +0000 (Wed, 15 Jan 2014) | 1 line
make linker command use shell on MacOSX - doesn't seem to work otherwise
------------------------------------------------------------------------
r406 | jdtournier@gmail.com | 2014-01-15 11:20:28 +0000 (Wed, 15 Jan 2014) | 1 line
fix linker flags on MacOSX
------------------------------------------------------------------------
r405 | ben.jeurissen | 2014-01-15 06:32:59 +0000 (Wed, 15 Jan 2014) | 2 lines
Correctly interpret non-normalized gradient orientations.
------------------------------------------------------------------------
r404 | jdtournier@gmail.com | 2013-12-05 10:58:53 +0000 (Thu, 05 Dec 2013) | 1 line
minor change to ensure MRView runs clean under Valgrind
------------------------------------------------------------------------
r403 | jdtournier@gmail.com | 2013-11-21 11:17:50 +0000 (Thu, 21 Nov 2013) | 10 lines
Update build script to add linker runpath into executables
This avoids any needs to set LD_LIBRARY_PATH or editing of ld.so.conf files.
The runpath is set to ../lib, so dynamic library must reside there relative
to the executables for this to work.
Tested on Ubuntu 12.04, needs testing on MacOSX.
No need for testing on Windows, since dll search path includes folder where
executable resides.
------------------------------------------------------------------------
r402 | jdtournier@gmail.com | 2013-11-15 13:58:27 +0000 (Fri, 15 Nov 2013) | 1 line
fix strange bug in name parser that caused issues when creating split nifti images
------------------------------------------------------------------------
r401 | jon.clayden@gmail.com | 2013-10-28 17:51:02 +0000 (Mon, 28 Oct 2013) | 1 line
Using Clang compiler by default on Mac OS X
------------------------------------------------------------------------
r400 | jon.clayden@gmail.com | 2013-10-28 17:49:44 +0000 (Mon, 28 Oct 2013) | 1 line
Disambiguating calls to round() to keep Clang happy
------------------------------------------------------------------------
r399 | jdtournier@gmail.com | 2013-10-11 14:52:58 +0100 (Fri, 11 Oct 2013) | 1 line
replace all uses of NAN, INFINITY, finite() & isfinite() with GSL portable equivalents
------------------------------------------------------------------------
r398 | jdtournier@gmail.com | 2013-10-10 10:43:35 +0100 (Thu, 10 Oct 2013) | 1 line
patch from Robert Smith to ensure transform matrix is correctly set for DICOM multi-frame data
------------------------------------------------------------------------
r397 | jdtournier@gmail.com | 2013-10-08 08:38:15 +0100 (Tue, 08 Oct 2013) | 1 line
remove stray 'raise' clause from build script that caused build to abort when subversion is not set up
------------------------------------------------------------------------
r396 | jdtournier@gmail.com | 2013-10-07 09:06:55 +0100 (Mon, 07 Oct 2013) | 1 line
remove all uses of variable-length arrays since these are technically not in the C++ standard (even though supported by most compilers). Newer versions of g++ seem to become progressively stricter, so this will need to be done at some stage.
------------------------------------------------------------------------
r395 | jdtournier@gmail.com | 2013-10-03 10:25:33 +0100 (Thu, 03 Oct 2013) | 1 line
make sure streamtrack honours the -quiet option
------------------------------------------------------------------------
r394 | jdtournier@gmail.com | 2013-10-03 10:12:10 +0100 (Thu, 03 Oct 2013) | 1 line
fix matlab write_mrtrix function's handling of comments
------------------------------------------------------------------------
r393 | ben.jeurissen@gmail.com | 2013-08-21 05:05:56 +0100 (Wed, 21 Aug 2013) | 2 lines
Default bvalue threshold wasn't that great
------------------------------------------------------------------------
r392 | ben.jeurissen@gmail.com | 2013-08-12 08:25:57 +0100 (Mon, 12 Aug 2013) | 2 lines
Add support for data acquired on multiple shells.
------------------------------------------------------------------------
r391 | jdtournier@gmail.com | 2013-05-30 07:56:54 +0100 (Thu, 30 May 2013) | 1 line
hopefully final changes to the build script for Python 2.6... and changed GDK_KEY_X entries to GDK_X for backward compatibility with older versions of GDK
------------------------------------------------------------------------
r390 | jdtournier@gmail.com | 2013-05-30 03:53:35 +0100 (Thu, 30 May 2013) | 1 line
and another fix for the build script under Python 2.6...
------------------------------------------------------------------------
r389 | jdtournier@gmail.com | 2013-05-30 03:47:51 +0100 (Thu, 30 May 2013) | 1 line
yet another attempt at fixing up the build script for Python 2.6...
------------------------------------------------------------------------
r388 | jdtournier@gmail.com | 2013-05-30 03:09:25 +0100 (Thu, 30 May 2013) | 1 line
trying to fix up the build script for Python 2.6...
------------------------------------------------------------------------
r387 | addo81@gmail.com | 2013-05-24 04:50:55 +0100 (Fri, 24 May 2013) | 2 lines
Utility to merge track files. First argument is the output, remaining arguments are track files to concatenate.
------------------------------------------------------------------------
r386 | jdtournier@gmail.com | 2013-05-24 03:37:34 +0100 (Fri, 24 May 2013) | 1 line
fix silly typo in build script
------------------------------------------------------------------------
r385 | jdtournier@gmail.com | 2013-05-24 03:34:53 +0100 (Fri, 24 May 2013) | 1 line
added -addcomment option to mrconvert to add comments to the header.
------------------------------------------------------------------------
r384 | jdtournier@gmail.com | 2013-05-24 03:27:42 +0100 (Fri, 24 May 2013) | 1 line
added -nocomment option to mrconvert to strip comments from header. Provides a quick and dirty way to anonymise data...
------------------------------------------------------------------------
r383 | jdtournier@gmail.com | 2013-05-17 13:02:11 +0100 (Fri, 17 May 2013) | 1 line
changes to build script to compile on non-unicode environments, and other changes to get around missing includes in recent update to GTKmm
------------------------------------------------------------------------
r382 | jdtournier@gmail.com | 2013-05-10 01:58:31 +0100 (Fri, 10 May 2013) | 1 line
fix build script to handle non-SVN installs and compilation against separate projects
------------------------------------------------------------------------
r381 | jdtournier@gmail.com | 2013-05-08 06:42:21 +0100 (Wed, 08 May 2013) | 1 line
added missing newline
------------------------------------------------------------------------
r380 | jdtournier@gmail.com | 2013-05-08 02:37:20 +0100 (Wed, 08 May 2013) | 1 line
modified build script to work under both python 2 or 3. Expect bugs.
------------------------------------------------------------------------
r379 | richard.beare@gmail.com | 2013-04-17 05:13:33 +0100 (Wed, 17 Apr 2013) | 5 lines
Set editing=false prior to early return from copyslice
Added some breaks to a switch statement for clarity/future safety.
They come immediately after return statements, and are therefore unnecessary.
------------------------------------------------------------------------
r378 | jdtournier@gmail.com | 2013-04-11 08:30:07 +0100 (Thu, 11 Apr 2013) | 1 line
minor modifications of Richard Beare's ROI editing extensions. Added buttons with tooltips to make the functionality more visible. This also required modifications to the way the flood fill works: it now fills from the current crosshairs position, rather than the actual cursor - this was required to allow the fill button to actually do something sensible.
------------------------------------------------------------------------
r377 | richard.beare@gmail.com | 2013-04-11 06:09:15 +0100 (Thu, 11 Apr 2013) | 3 lines
Changed offset arrays to int. Previously incorrectly set as unsigned.
------------------------------------------------------------------------
r376 | richard.beare@gmail.com | 2013-04-11 05:31:52 +0100 (Thu, 11 Apr 2013) | 16 lines
Extensions to mrview roi painting capabilities.
These are loosely modelled on options available in MNI Display.
1) Turning 3d on and off - this gives a circular brush in the view plane.
2) An isotropic brush option, so that brush size is approximately in mm, for highly anisotropic scans.
3) Fill (press f/F) - this fills with the inverse of the mask voxel under the pointer, so it can be used to delete blobs too.
4) Slice copy from previous/next (press n/N or p/P)
5) undo/redo ctrl-z, ctrl-shift-z (in principle, cmd on the mac). There is currently a limit of 10 steps, but that is just a constant.
There are likely to be issues if oblique masks are used in combination with isotropic brushes.
The history of this development is in my github repository github.com/richardbeare/mrtrix . I haven't set up ways of pushing directly
from my git repo into the googlecode svn repo, but simply copied files manually. Hopefully this hasn't caused problems.
------------------------------------------------------------------------
r375 | jdtournier@gmail.com | 2013-03-27 06:20:27 +0000 (Wed, 27 Mar 2013) | 1 line
add sanity check in case of missing frames in DICOM header - avoids unexplained segfaults for incomplete/corrupt DICOM folders
------------------------------------------------------------------------
r374 | jdtournier@gmail.com | 2013-03-13 03:22:56 +0000 (Wed, 13 Mar 2013) | 1 line
add some error checking to file I/O in track file writer