| Commit message (Collapse) | Author | Age |
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4299 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
and made it include libmpi_cxx
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4296 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4291 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4290 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4288 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4287 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
There seem to be new Darwin releases every week or so.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4283 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
| |
(old versions of g++) don't support restrict in arrays which are
structure members.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4282 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4281 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
| |
buggy compilers (e.g. old versions of gcc) do not support "restrict"
on arrays which are function arguments. On these compilers, we now
disable "restrict".
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4280 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
compile something. It seems as if AC_PROG_*_GNU does not clean up
after itself.
On case-insensitive file systems, having a file "conftest.c" around
while using cat to write to a file "conftest.C" means that the
resulting file is still "conftest.c", so that it is compiled as C and
not as C++ programme.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4279 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4278 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4274 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4273 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
line" requires a reconfigure of existing configurations.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4270 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4269 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
|
| |
up into make.config.defn so that the latter file has a chance of appending
necessary system libs to the linker command line.
This applies patch http://www.cactuscode.org/old/pipermail/patches/2006-March/000156.html with a slight modification: the make.link file is included directly
in make.config.defn, not in make.extra.defn (as proposed in the original patch).
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4268 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
The language must be either 'C' or 'Fortran', allow for case-independence.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4265 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4255 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4248 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4247 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4246 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
| |
Add support for gcc.
Update IBM compiler flags.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4242 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
* add proper detection of gfortran (95)
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4241 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
The other F77 flags are already overwritten.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4240 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4239 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
setting F77FLAGS to F90FLAGS. Since this change, compiling F77 code
with the IBM Fortran compiler seems to be broken. The compiler does
not recognise any more that F77 files are in fixed format. This
problem has been extensively discussed on the developers' list in
October 2005. This patch corrects that problem.
Some known-architecture files produce makefile fragments that add, at
run time, additional flags to F90FLAGS to set the source code form
correctly (fixed or free format). These fragments need to add the
same flags to F77FLAGS as well. This concerns the architectures AIX,
BGL, and Darwin.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4238 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
when using an F90 compiler.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4237 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4232 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
#define M2(a,b)
warn about the uses M2(x,) or M2(,y) or M2(,).
The empty arguments are still handled correctly. These warnings are only to
make it easier to port to systems where the system cpp does not handle empty
arguments and where one does not want to use the perl cpp instead.
Patch from Erik Schnetter.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4213 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4212 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
is actually 'restrict'.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4211 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
| |
convert the macro value to either 0 or 1 before returning it.
Otherwise things fail badly if the macro is defined but empty: The
returned value is then not an integer, and cpp.pl would keep both the
#if and the #else branch of an if statement.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4205 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
testsuites.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4203 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
| |
capabilities. This prevents name clashes with e.g. system header
files.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4201 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
(e.g. if a thorn's configuration.ccl has changed).
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4200 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
configuration bindings.
When reading the configuration settings from the output of the
configuration scripts (that provide capabilities), new lines were lost
in certain circumstances. This lead to makefile fragments like
PETSC_DIR = /usr/local/apps/petscPETSC_ARCH = linux
which are obviously incorrect (since these should be two lines, not
one), although this does not lead to errors or warnings.
This patch makes sure that all configuration settings have newline
characters appended or blanks prepended as they are concatenated into
the internal database (ConfigScriptParser.pl). When the database is
read out, the strings can just be printed to the file
(CreateConfigurationBindings.pl).
While doing this, this patch also removes some superfluous blanks and
newline characters from the output of the configuration settings.
This patch also removes a perl idiosyncracy when appending newline
characters to a string. It changes code like "#endif" . "\n" to
"#endif\n".
Tested on Mac OS and Red Hat Linux.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4199 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
| |
was left untouched if the new content was empty. This meant that
empty files were never generated, and existing non-empty files were
never made empty.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4198 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4193 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4192 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4191 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4189 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
| |
$MPICH_DIR/bin/tdevice. Unfortunately the tdevice command is broken, even for a
'ch_shmem' device it reports 'ch_p4'.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4187 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4185 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
${F90_OPTIMISE_FLAGS='$(F77_OPTIMISE_FLAGS)
in the PGI configuration section in lib/make/known-architectures/linux
causes trouble. It means "use as F90 flags whatever the F77 flags
will be set to". This doesn't work together with the a recent patch
to configure.in, which says "use the same flags for F77 as for F90".
Overall we now have a circular definition.
As a quick fix, I replace the above line with the value of
F77_OPTIMISE_FLAGS. Since there is no other code which sets a shell
variable (the left hand side) to a Makefile expression (the right hand
side), it may have been an oversight anyway -- probably the author of
the above line wanted to write ${F77_OPTIMISE_FLAGS} instead anyway.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4184 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
| |
overrides F77FLAGS with F90FLAGS. This patch also sets
F77_DEBUG_FLAGS and F77_OPTIMISE_FLAGS from the corresponding F90 flag
settings.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4183 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
| |
an if clause so it is only tested if CCTK_NEED_X is true.
The PETSc extras file has been updated to set this flag.
This will be reverted when the extras stuff is removed and
thorns are using the thorns in CactusExternal to detect things.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4182 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
| |
only read the file if it its size equals the length of the rdata string.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4181 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
|
| |
Similar to Windows, which is already handled as special case, it uses
a different extension for dynamic libraries, namely ".dylib". I test
for MacOS by grepping the output of uname for "^Darwin". This patch
is necessary to make Cactus configure with HDF5 on my notebook, where
I installed HDF5 via fink in the standard way.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4180 17b73243-c579-4c4c-a9d2-2d5706c11dac
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
the last macro argument is empty, it thinks there is one argument too
few. That is, the following code
#define M2(a,b)
M2(x,)
leads to an error message stating that M2 was used with only a single
argument. This is wrong, as it is called with two arguments, where
the second argument is the empty string. After reading the GNU cpp
manual, I think that the ANSI standard requires that empty arguments
must be handled.
The problem is that the routine SplitArgs that decodes the macro
arguments cannot distinguish between an empty last argument and no
argument, as both are represented by an empty string. I change the
routine so that an empty string always means "empty argument". This
means that SplitArgs cannot be called any more if there are no
arguments -- but this does not matter. I place a corresponding if
statement around the call.
Empty macro arguments may cause problems on some system. A later
patch will produce warnings or errors for them.
git-svn-id: http://svn.cactuscode.org/flesh/trunk@4179 17b73243-c579-4c4c-a9d2-2d5706c11dac
|