Activity by Clock Time
Removed unnecessary ';'.
0 lines of code changed in 2 files:
Hack... had to cast to (int) so the SGI compiler would know what to do... :-(
0 lines of code changed in 2 files:
M Core/Datatypes/PrismVolMesh.h
M Core/Datatypes/TetVolMesh.h
I think I have finally found a syntax that the SGI compiler will take
and that will not cause problems for other compilers. The SGI
compiler has been very picky... and troublesome. In investigating
this fix, I've had CC (on the SGI) give me a compiler error, then I
monkey with the syntax, and it gives me another error, then I put it
back to the original and it compiles... However, when I then try to
compile the code using a different (fresh) build, it breaks again.
M Dataflow/Ports/PathPort.cc
M Dataflow/Ports/ColorMap2Port.cc
M Dataflow/Ports/TexturePort.cc
M Dataflow/Ports/MatrixPort.cc
M Dataflow/Ports/NrrdPort.cc
M Dataflow/Ports/StringPort.cc
M Dataflow/Ports/MatrixPort.h
M Dataflow/Ports/BundlePort.cc
M Dataflow/Ports/ColorMapPort.cc
M Dataflow/Ports/ImagePort.cc
M Dataflow/Ports/FieldPort.cc
Moved the template instantiation to below the explicit setting of the
port_type_/port_color_. This is necessary to get the SGI compiler to
compile this code... and doesn't seem to make any difference to any
other compiler.
0 lines of code changed in 26 files:
Warning messages in constructors (using 'cerr') cause a core dump (on at least my 10.3 Mac). Replaced with 'printf'.
0 lines of code changed in 2 files:
Cosmetic updates. Removed unnecessary ';'s.
0 lines of code changed in 2 files:
Cosmetic updates.
0 lines of code changed in 2 files:
Fix 'non' return of 1. Remove redundant comments.
3 lines of code changed in 1 file:
Second attempt to get SGI and other OSes to compile (and run) this stuff. I don't understand why this typecast appears to work... to the best of my knowledge the docs on hash_multiset say the first parameter is the 'type' of the map...(not the size?) However, preliminary results indicate that this works. A little more testing will be necessary to confirm it.
17 lines of code changed in 2 files:
Fixes (hack) to the new Basis stuff to allow compilation on the SGI.
Core/Datatypes/CurveMesh.h
Had to give type_name a parameter because for some reason the SGI
compiler couldn't find the correct function without it...
Core/Datatypes/PrismVolMesh.h
Fixes to allow hash_multisets to compile on the SGI... this effects
the linux/mac build as they use the same code, and I am not sure how
it compiled on linux before...? However, this 'hack'/fix appears to
be the correct way to construct the hash_multisets, so it should just
work everywhere now. More information on this change is in the file
(search for hack_eit).
Core/Datatypes/TetVolMesh.h
Can't use #include <cmath> on the SGI (and a few other
platforms)... changed to good ol' #include <math>.
Did some indenting on the #if statements to help follow them. Added
// end class name comments to help figure out what are nested classes
and where they end. Added hash_multisets hack as described about and
in this file.
Dataflow/Network/PackageDB.cc
Added a check for an error with looking up the *.xml files so that a
useful error message gets printed. It appears to me that 'files' will
always be non-nil and thus the first check doesn't really do anything.
Also, added a warning that it looks like 'files' is a (small) memory
leak... hopefully someone with more time will track that down and fix
it.
StandAlone/utils/TestBasis.cc
Added a dummy var (test_vector) that instantiates a vector of vector
of Point so the SGI compiler doesn't complain about not being able to
find it. (More info in the file itself.)
StandAlone/utils/UnitElementMesh.cc
Same hack/fix as in TestBasis.cc above.
105 lines of code changed in 8 files:
Remove extra ;
1 lines of code changed in 1 file:
Removed extra ;
1 lines of code changed in 3 files:
These should fix the SGI compilation... I think. I have one other
hack in another file that may be necessary, but the compilation seems
to work (sometimes) without it, so I am leaving it out for now.
Core/Basis/Constant.h
Fixed SGI compilation by making the array non zero in size. I added a
comment in the file about what I did, but it would be great if someone
who really knows what is going on would update the comment to reflect
that knowledge.
Removed a argument to derivate to quite the compiler. Don't know if that
arg should be used or not, and put in a comment to this effect. Again,
if someone who knows this code could update the comment, that would be
great.
Core/Basis/NoData.h
Same non-zero array fix as reported above.
22 lines of code changed in 2 files:
Fix compiler warnings.
1 lines of code changed in 1 file:
Organize #includes. Fix typos in comment... though I still don't know what it means. Ident #ifs.
14 lines of code changed in 1 file:
consolidate the stl #includes
3 lines of code changed in 1 file:
Fix compiler warnings as I continue to track down SGI problem.
2 lines of code changed in 1 file:
Fix (at least) sgi compile by fully qualifying Cell and Elem.
6 lines of code changed in 1 file:
Remove more warnings as I try to figure out why the SGI build is broken.
1 lines of code changed in 1 file:
Added #include of Point.h... since Point is used in this file... This
allows SGI compile to complete... strange that the SGI compiler can't
figure out what a Point is and the other compilers can... either way,
it is a good idea to #include files that you use directly and not hope
that they are indirectly brought in for you.
1 lines of code changed in 1 file:
Commented out an unused var.
0 lines of code changed in 2 files:
(173 more)