"On fixing the pointer access I discovered that reading osga archives
containing ive files went into a cpu loop. This turned out to be a
problem with proxy_streambuf on Solaris. Public methods in the Solaris
streambuf standard library implementation rely on the gptr() being set,
which proxy_streambuf was not doing. So I have modified
proxy_streambuf to set the input sequence pointers, and have also
aligned it more with the standard library streambuf implementation
where all input is through underflow(), not uflow() which merely calls
underflow() and advances the pointer."
From Robert Osfield, change from using pointer cast and assignment to using
a templated _write and _read method to avoid pointer aliasing to 2/4/8
byte boundaries that some computer system may produce. These changes
where inspried by Colin McDonalds change to using memcpy, these
changes weren't merged as memcpy is not as clear in naming as _read,
_write and memcpy will incurr a function call just for copy a
uint.
"Some makedefs fixes for Solaris:
For multithreaded applications the -mt option must be specified on both
the compile and link steps, to ensure correct behaviour. According to
the Sun compiler documentation it sets REENTRANT flags in the system
header files, and links the -lthread library in the correct order.
When compiling shared libraries the -KPIC option should be specified.
Although Solaris will handle shared libraries without
position-independent code there is a performance penalty. The linker
reference manual says: 'If a shared object is built from code that is
not position-independent, the text segment will usually require a large
number of relocations to be performed at runtime. Although the runtime
linker is equipped to handle this, the system overhead this creates can
cause serious performance degradation'."
(http://openscenegraph.org/archiver/osg-users/2005-June/0575.html);
after switching children of a Switch node off and on again, they become
unpickable. This issue occurs first in 0.9.9, with 0.9.8 everything is fine.
My fix involves calling dirtyBound() every time the on/off-values of the
Switch are changed"
NSLookupSymbolInModule. The former call would lookup the named
symbol NOT in the current dynamic library, but in the entire running
program while the call NSLookupSymbolInModule, takes the handle to
the library the symbol should be found in. This means the current
code will fail if one loads multiple bundles at runtime and attempts
to load the same named symbol from each one."
in removes any options beginning with "-psn" from argv on OSX by
calling the "remove" method. When a .app run is created in OSX,
which is required to get a fully functioning UI application, the OSX
finder passes a -psn_XXXX option to the application where the XXXX
refers to a unique process number. An example option would be "-
psn_0_37617665". The argument parser was choking on this option in
all the osg example applications."
"I took a closer look at the conditional code in
SeamFinder::seamReplacement().
Because _info.minRange is a double and lod->getMinRange(0) is a float,
the difference will be calculated with double precision. If
_info.minRange is cast as a float it is exactly the same value as
lod->getMinRange(0) and the difference is exactly zero.
So if you change
if((fabs(_info.minRange-lod->getMinRange(0))<0.001)&&(fabs(_info.lod0Range-lod->getMaxRange(0))<0.001))
to
if((fabs((float)_info.minRange-lod->getMinRange(0))<0.001)&&(fabs((float)_info.lod0Range-lod->getMaxRange(0))<0.001))
it works a lot better."
matrixd. It was returning the values of the diagonal
of the matrix, which only returns the scale if there
is not a rotation. I fixed this by returning the
length of the vectors that form the basis.
I also added a function to orthonormalize the
rotation component of the matrix. I seem to always run
into situations where non uniform (or even uniform)
scale complicate my calculations, and I thought other
members of the community could use this function as
well."
