NMRViewJ 8.0.a2 now available

[Bruce Johnson <nmrview@xxxxxxx>]

NvJ 8.0.a2 has a number of significant changes including:

Documentation
	There is now a new version of the guide to RunAbout (this is still  
not quite consistent with the latest version, but is much better than  
before) and a brief overview of protocol for using titration analysis  
tool.
At present, the docs are only available within NvJ from the Help menu  
and some of the images are not rendering properly (I'll fix this in  
8.0a3)

Testing
  	We've now got a much better test suite for some core  
functionality (and have fixed a number of bugs that the tests  
found).  One test, for example, reads into memory one hundred 130 Mb  
NOE files at the same time.  NvJ now handles this without any trouble.

Spectrum Bindings
	We've added a number of new standard bindings to spectra (like "e"  
for expand window, "f" for full etc.).  These are documented (Section  
on Creating and Using Spectral Display Windows -> Using Spectral  
Windows -> KeyPad Navigation).  Most of the bindings and some new  
related procs were incorporated from Mike Summers' lab at UMBC).

Mouse Modes
	The three button mouse mode is now more like the old NvC code, in  
that you can select the vertical or horizontal crosshairs by clicking  
close to them.

RunAbout
	It was possible in RunAbout to get in an endless loop if you formed  
a "cyclic" fragment while confirming cluster links.  This is no  
longer possible.

Memory Usage:
	Recent additions to the internal code for peaks, for better  
compatibility with CCPN and NMRSTAR3, led to signficicant bloating of  
the peaks.  With large peak lists (tens of thousands of peaks) this  
led to potential out of memory problems).  We've significantly  
reduced memory usage of peaks in 8.0.a2.

STAR3 files
	NvJ's STAR3 files store a list of opened datasets.  On Windows,  
there was a problem rereading the STAR files (because of Window's  
stupid backslash path separator).  When writing STAR3 files the  
dataset path separator is now always changed to a forward slash  
(which Java reads read properly on all platforms).

Titration Tool
	The tool can now use a set of peaks you've selected in the spectrum  
window, in case the automatic path finder doesn't work.



Note as indicated by the "a2" designation we still consider this an  
"alpha" version of the software, indicating that before the official  
8.0 release there may be some incompatible changes, and signifcant  
bugs may remain (that said, it should be better than any previous  
version).


The complete (mostly) list of changes (since version 7.1.17) is below:


Set spectrum attributes window to not be always on top

Remove QD dummy atom from tyr.prf
Its incorrect attachement to CD1, rather than CG, was causing
the calculation of equivalent atoms to be incorrect.

Don't save titration window when saving state.

Change meaning of min matches field in cbca2.tcl to be >= rather than >
so that you enter a "3" if 3 matches are required (before fix) you  
entered
a "2" to require 3 matches.
Update cluster display after adding a peak.

Add Pick button and method to pick selected datasets in Dataset table.

Add more default key bindings to spectra (e, f, p, c, v x, y r, w)
Automatically change first character of toplevel window names to  
lower case.

titration.tcl
Rename toplevel from "Rate Analysis" to "Titration Analysis"
Add "Use Selected" and "Use Table" modes
Remove some unused buttons from controls

specat.tcl
Check dataset dimensions to make sure they all have unique labels  
prior to peak picking.
Fix zoomTo to use new crosshair mode 4
Add copy/paste xslicer/yslicer, getLimits/setLimits procs (these are  
modified and renamed from UMBC scripts)

star3.tcl: Automatically add ".str" to star3 file names if not  
specified by user.

Fix problem with setting crosshair positions and change crosshair so  
that
in 3 button mode one can select horizontal or vertical crosshair by  
clicking near them.

Fix problem (in PeakDim.java) with setting labels when reading star2  
files

set default dataset scale to 1.0e6
Remember last dataset and block id for getBufferVector (was returning  
wrong data sometimes when it
only checked last block id.
Store dataset path in star3 file in uri format (with forward  
slashes).  Fixes problem in reading
dataset paths in star3 files on Windows as the back slash path  
separator was seen as an escape
sequence.

PeakClusterCmd:  don't allow making links that will cause a cyclic  
fragment.
Only allow one link from a cluster (will have to make duplicate  
clusters to have multiple links now)

MoleculeTable:
Fix error when changing ppm (threw casting error because ppm column  
was Double, not String) )from atom assignments table.

STAR3: Catch errors and continue if datasets recorded in star3 file  
can't be found.

Fix menubar in swkcon (wasn't showing up do to changes swanks menubar  
stuff)

Make getBufferVector static (fixes problem with multiple threads  
trying to get buffers).

Specify values for CSI buttons

Faster updating for relaxation table (don't refresh every addline)

Fix bug in "nv_win pos -dataset datasetName value" (using wrong  
argument for value)

hetnoe.tcl :  Destroy table if it exists when recreating interface
Faster updating of table (don't update with every inserted line, just  
at end)

Throw error in "vecmat max" if vector empty

Remove peakDimContribs from resonance when deleting peaks (ensures  
that unused resonances are garbage collected)

Put "addline" in table model onto event thread

Apply dataset scale when doing analyzeRegion

Add BufferVector back in (accidently removed in previous commit)
Fix in checkForPeak for bug that rejected real peaks.

cbca2.tcl: Check isViable when doing extend.

Fix bug with new Resonance and PeakDimContribs iterator (unlink not  
working right)

Change "nv_win draw peaks" to accept shorter versions of "peaks",  
like "nv_win draw pea"

Changes to Spectrum.java (and related files) to move widget changes  
onto event thread

Add code to eliminate ArrayList in PeakDim and Resonance if not  
needed (if only one PeakDim linked to Res)
Peaks now use less memory

Update peaks files to pass new test suite.
Update dataset and peak tests (and remove dataset2.test and peak2.test)

Add code for "registerLoadFileListener" so that procedures can be  
executed when file is loaded.

Allow setting username/password for connections
Up heap memory to 384M
[project @ vec.tcl: Change from jadd to add cascade in Filer]

[project @ nvtclExt/crosswin.tcl:  set to be "alwaysontop"]
[project @ Resonance.java: when setting atom, also set name (to atoms  
"shortName")]

[project @ cntrl.tcl:  Change help menu to use new jhelp.]

[project @ cntrl.tcl:  Move STAR2 menu entries to a STAR2  
(deprecated) sub-cascade]

[project @ ResonanceCmd.java:  Add sub-option to set label.]

[project @ ResonanceSet:  set resonance names from atom when  
assigning atom to ResonanceSEt]

[project @ STAR.java: Reading STAR2 files was not getting peak label  
field (was getting reset when resonance field read)]

[project @ Graph.java:  Wasn't working on Java 6 because  
createTransformedShape returning Rectangle2D.Double (not Float)]

[project @ gui.tcl:  Remove if test of javaMode (and non javaMode  
code), replace many calls to set Nv_Priv to one array set command,  
remove splashProgress calls.]

[project @ init.tcl:  Initialize TJC]

[project @ specat.tcl:  Add rotateView, changeLevel and zoomTo methods.]

[project @ specat.tcl:  Set spectrum attributes window to be always  
on top]

[project @ specat.tcl:  Remove calls to splash stuff (just slows app  
down)]

[project @ specat.tcl:  Remove "widget object" commands (no longer in  
Swank)]

[project @ spectrum.tcl Add key bindings for changeLevel, rotateView  
and zoomTo]

[project @ start.tcl:  Remove '[. object] repaint' clall (widget  
object subcommand removed from Swank)]

[project @ start.tcl: load TJC compiled code.]

[project @ start.tcl:  Remove calls to splashProgress, splashText,  
they just slow app down.]

[project @ strips.tcl:  Remove use of widget object subcommand (which  
was removed from swank because it's bad!!!)]

[project @ table.tcl:  Move all procs in table.tcl out of table  
namespace {}, and rename with FQN ::table::*]
This is necessary for TJC to find the procs

[project @ specat.tcl:  Move all procs out of spectrum namespace {}  
and rename with FQN ::spectrum::*]

[project @ Move procs that were not in spectrum namespace from  
specat.tcl to spectrum.tcl]

[project @ Coupling.java: Remove debugging message]

[project @ specat.tcl: Add rotateView proc]

[project @ crosswin.tcl:  Fix problem when dataset doesn't exist.]

[project @ Datasets:  Improve speed by reading from Buffers in whole  
blocks, not point by point.]

[project @ Spectrum*.java:  Update to be consistent with coding in  
Swank widgets (especially for configure, cget etc.)]

[project @ NvTableModelCmd.java:  Fix threading issue (model assign  
not on event thread)]

[project @ peakid.tcl: Setting x and y ppms backwards on second  
spectrum.]

[project @ nvtclJ/init.tcl:  Don't update (splash screen) faster,  
update version to 7.2.0]

[project @ cntrl.tcl:  Add STAR3 menus]

[project @ cntrl.tcl: Change to using jmenubar (not menu).]

[project @ noe.tcl:  nv_noe elem list now returns listname not number  
(so changed noe.tcl to support this), fixed color in noe plot]

[project @ noe.tcl:  Change to using NvFillPeakMenu]

[project @ specat.tcl:  Don't preset geometry and remove some update  
calls (now faster and less "jumpy")]

[project @ cbca2.tcl:  More work on Tocsy support]

[project @ access.tcl:  Change use of [combo object] removeAllItems  
to remove all]

[project @ analyzer.tcl and adjust.tcl  Add support for Sparse NNMF  
(also code to ignore PC* fields in selector)]

[project @ molecule.tcl:  Remove use of [menu object] removeAll  
(change to menu remove all)]