declaration/definition of static variables in templatized classes.
There is an irreconcilable difference between cygwin and Solaris, that
I handled by creating a compiler define in config.h.
solaris.
preconfig now sets it to null.
Unless the user has set it to something nonnull,
configure then resets it to -fPIC for all platforms except for Solaris,
where it set to null. Solaris doesn't require -fPIC.
a FULL python installation, but will inevitably fail because neither the numarray
or numeric package has been installed.
Now, a check is made during the raw preconfig run to see if the include
files for the numarray package can be found. If they can't be found and the
python installation option is set to default, then the python build variable
is set to 1, reducing python to a minimal installation.
In most cases, the raw out-of-the-box
preconfig
make
make test
steps will now work. Whereas before, there were compile errors occurring because
the numarray include files were not able to be found.
More information about the SUNDIALS installation is now printed out.
CVS
the macro AC_PROG_INSTALL to the program config/install-sh, which was
the backup option for AC_PROG_INSTALL anyway.
The driving reason for this was that the default /usr/bin/install
program on linux had problems when directories under $(CANTERA_ROOT)/build
weren't owned by installing user. Note, this can happen in a
collaborative environment when multiple people have group rw
on the source/build directory.
However, in general, this should help cross-platform variability
issues, since there is less variability now.
The user can override config/install-sh choice by setting the
environmental variable INSTALL_BIN in the preconfig script.
of the Makefiles in the code. This variable will now be used
to specify the position independent code compiler option within
Cantera. Note, this bug fix was needed for a 64 bit full python compile
on linux.
-> Added a pureFluidTest problem that only gets carried out if
the WIT_PURE_FLUIDS conditional compile takes place.
-> added more to testWaterTP, so that it queries the unstable
water region up to the spinodal curve.
Fixed an error in IdealMolalSoln -> density wasn't beeing recalculated.
Updated the test problem with the fix.
Worked CANTERA_DEBUG_MODE into the Makefiles. It wasn't being used.
Split the multiproblem DH_graph_1 test into multiple directories, one per test
Worked on getting all of the test problems to have the same look, feel, and printout.
