GitLab

I am not sure what the correct solution is: either BLUE should be removed from the
list of 3d refinement rules.  Or somebody needs to implement the missing rules.
But I am not sure what they would be?  Would they be like HEX_BISECT_0_* ?

The old names clash with the placeholder constants of the same name
in libc++.

Oliver Sander authored Sep 29, 2015 and Oliver Sander committed Mar 01, 2016

This file is not used anywhere, it is not clear what it is good for,
and it hasn't been touched since 1996.

This was reported by Debian's lintian tool.

This addresses the following warning from GCC:

std_domain.cc: In function 'UG::INT UG::D2::BNDP_SaveBndP_Ext(void**)':
std_domain.cc:4296:24: warning: iteration 1u invokes undefined behavior [-Waggressive-loop-optimizations]
       dList[j] = pos[j];
                        ^
std_domain.cc:4295:5: note: containing loop
     for (j = 0; j < DIM; j++)
     ^

Remove XCode project

Cloned from https://github.com/ugfem/ug/issues/9

This was generated with XCode 3.9 and can no longer be read by recent
versions. As far as I can tell, it no longer serves a purpose.

See merge request !1

This was generated with XCode 3.9 and can no longer be read by recent
versions. As far as I can tell, it no longer serves a purpose.

Christoph's conversion to C++ must have messed it up somehow.

This is a left-over from the old build system.  For some explicitly given
architectures, UINT is not unsigned int, for example.  The special-casing
stopped working when we moved to the AutoTools long time ago, but hardly
anybody noticed because amd64 worked anyway.  Now this old code gives
us trouble building dune-grid on Debian ppc64el.  Since I don't even
know why these types need to be changed, and what the requirements on
the types are, I just kick out all the special cases.  If something
goes wrong we can then fix it in a portable way.

The pattern_data array was recent changed from char to unsigned char,
to silence C++11 compiler complaints.  However, the X method XCreateBitmapFromData
really does expect a pointer to char, so we have to introduce an explicit cast.

Indeed, I have no idea what it does or did.  In any case, its name does not
appear anywhere in ug/.  Hence I'll simply delete it.

Because the objects are Priorities semantically, so having them be the correct
type eases debugging.

Avoids gratuitous semicolon if REP_ERR_FILE is empty.

During the first run of 'uncrustify', there were a few files which the uncrustify
tool couldn't handle, mainly due to weird macro-magic effects.  Those files where
left as-is, with lots of weird formatting und tabs.  This patch now uncrustifies
them, too, by neutralizing the macro-magic with some manual intervention.

Transform forgotten folders dev/ps and dom/lgm/ngin2d.
Minor clean-ups.

Adjust all includes outside of ui/.

Add flags in Makefile.am to make all compile.