Trying to fix Patrick's problems when launching getdp on Windows...

When gmsh is started from the command line on Windows, cygwin does
not (for some obscure reason) prepend '/cygdrive/x/' to the path. Since
we check for '/cygdrive/' in SystemCall() in order to sanitize the
path before passing it to CreateProcess(), the system call fails.

I think the right way to deal with this is to actually remove the
path change in SystemCall() (since we only change the leading part),
and use the official cygwin API to change the path of every piece
that will eventually end up in the argument of CreateProcess().
And that's what "FixWindowsPath()" now does...

added an option to specify the maximum delay Gmsh will wait for a
solver to respond

forgot to commit this

1.56

work around the "window in window in window" MacOS X FLTK bug (ColorBar_Window
not showing completely) by replacing the option sub-windows with groups.

complete WriteView in parsed format (print strings)

draw the labels a little bit closer to the entities

reworked the "post-processing file format" section

added ability to save views in "parsed" format

added an option to draw the nodes per element (much slower, but permits to
visualize only a subset of the nodes)

slightly reworked interactive/non-interactove section

explain that "Duplicata View[i]" does not actually copy the data

forgot to add this

one last touch :-)

more fine-tuning

fine-tune size

fixed alignment problem when views have different View.Formats

better label

fix potential crash with TimeStepMin/TimeStepMax and empty views

when there is not enough space to diplay all the values, but there is
enough space to draw a middle value, do it

add an option to draw the post-processing scales horizontally + remove
Post_View.TransparentScales (never really worked well anyway)

compute Euler angles from rotation matrix

physical<->elementary

default partition number should be 1

renamed mshdensify to mshsort and modified the spec file to install it
in the rpm

A small utility that reads a Gmsh mesh file and makes the numbering
dense (without gaps in the node/element number sequences)

256 -> sizeof(string)

Correction of bug in List_LQuery (bad size of remaining list to be inspected)

(Mirroring bug fix from Patrick in GetDP)

p5

update BoundingBox doc to match what the code actually does + explain
relative/ansolute paths + small 'cosmetic' changes

better definition for holes (suggested by gkmohan@gmail.com)

Better handling of '-psn_XXX' command line arg. from the Mac Finder