Added the configuration command line and CFLAGS/CXXFLAGS to the snapshots and greeting message.

Does what it says on the tin. Just wanted to run it pass you in case you have a better way of doing this.

Thanks ! Branch can be removed.

The start of the output now looks like this:

 Welcome to the cosmological hydrodynamical code
    ______       _________________
   / ___/ |     / /  _/ ___/_  __/
   \__ \| | /| / // // /_   / /   
  ___/ /| |/ |/ // // __/  / /    
 /____/ |__/|__/___/_/    /_/     
 SPH With Inter-dependent Fine-grained Tasking

 Version : 0.4.0
 Revision: v0.4.0-509-g2474a3c2-dirty, Branch: master
 Webpage : www.swiftsim.com

 Config. options: '--disable-vec'

 Compiler: GCC, Version: 5.4.1
 CFLAGS  : '-O3 -fomit-frame-pointer -malign-double -fstrict-aliasing -ffast-math -funroll-loops -march=haswell -mavx2 -fno-tree-vectorize  -Wall -Wextra -Wno-unused-parameter -Werror'
 CXXFLAGS: ''

 HDF5 library version: 1.8.16
 FFTW library version: 3.x (details not available)

[00000.0] main: CPU frequency used for tick conversion: 2693658262 Hz
[00000.0] main: Running on: matthieu-laptop

mentioned in issue #219 (closed)

Added 4 commits:

• 27174568 - Make hydro scheme and dimensionality configure options
• 32298b94 - Merge branch 'master' into autotools-update
• 3652fec1 - Merge branch 'autotools-update' into configure_options_in_code
• f38e3714 - Selection of the different modules is now done in the configure script.

Ok, I have now gone all the way and moved all the module selection to the configure script.

Some non-standard options are still in const.h. I need to think about how to handle these special cases.

Added 1 commit:

• 22f1f685 - Also moved the Riemann solver choice to the configuration script.

Added 1 commit:

• 5825c624 - Missing files...

All looks good to me, except why are we recording CXXFLAGS? These will only be used if we start coding with C++. I'd say remove those.

I thought @jwillis was using them to pass some vectorization flags. If not, agreed, we can drop it.

I guess we don't have any reason to keep the LDFLAGS ?

I use CPPFLAGS sometimes but not CXXFLAGS, I might have used it in the past when we tried to overload the vector operations on the KNL.

Ok. I'll drop the CXXFLAGS for now then.

Note that this will work only for the main code anyway. Whatever happens in the test directory cannot be recorded using that mechanism. Especially hand-editing the Makefiles.

I wouldn't bother with LDFLAGS, the linked against libraries can be discovered from the binary, as should the paths used to locate them (since we use rpath).

Added 1 commit:

• f1788d3d - Record CPPFLAGS (for @jwillis) and not CXXFLAGS.

Ok, replaced the CXXFLAGS with CPPFLAGS.

Hmm, CPPFLAGS still means we are using C++, so unless we intend to allow that permanently I don't see the reason to keep that either.

Title changed from Added the configuration command line and CFLGAS/CXXFLAGS to the snapshots and greeting message. to Added the configuration command line and CFLAGS/CXXFLAGS to the snapshots and greeting message.

My understanding is that CPPFLAGS are for the pre-processor and is language agnostic.

I thought that also, I have been using it to define macros

But you only need them when working with more than one language, so these share flags between C and C++ for instance. Without C++ they are not very useful.