Fix typo in Hazy2

See merge request !56

While updating the Hazy covers, a string '-2' had been copied from an
online source.  That was not recognized on Mac, and led to failed Hazy
compilations.  Replace that minus with an ASCII equivalent.

Use NIST ionization potentials, not phfit.dat

See merge request !54

Update NistExtractor to current URL interface

See merge request !53

Previously, for H and He iso-sequence ions, the ionization potential was
obtained from the Stout energy files, ending in .nrg.dat.  A comparison
shows that these values are identical with the NIST values found in
ionization_potentials.dat.

Remove the IPs from the .nrg.dat files, and modify the data parser (in
iso.cpp) to use the NIST values instead.

Previously, NIST ionization potentials were not used when computing
photoionization cross sections; instead the data in phfit.dat were used.

Define method getEthresh to switch between using NIST values for valence
shells, and the Verner data for inner shell ionization thresholds.

Previously, the code used function ph1() to access the valence
threshold energies (ionization potentials) stored in file phfit.dat.

Define method atmdat.getIonPot() to access the NIST IPs (read from file
ionization_potentials.dat), and use it in lieu of the call to ph1(),
where possible.

Convert all REAME files to Markdown

See merge request !52

Update Hazy 2, 3, and QSG front-page images

See merge request !51

Fix bug with revision string sans git repo

See merge request !48

Fix bug in gitversion.sh

See merge request !47

The previous implementation assumed a .git repository be present.  This
is obviously not true for release tarballs.  In that case, gitversion.sh
emits git's fatal warnings that a repository was not found, the revision
is set to the empty string, and the smoke test crashes in version.cpp.

Add a test in gitversion.sh to gracefully handle check for a git
repository being present, and modify version.cpp to handle empty
revision strings.

Reported-by: Peter van Hoof

When a release tag is cloned, the branch default to '(no branch)'.
The revision generated for the branch contains an unbalanced parenthesis
(i.e., it is "(no-c22.00"), which causes compilation to abort.

The present fix causes the revision on a cloned release tag to be, e.g.,
'c22.00'.

Recall, this appears in the first line of the main Cloudy output.

NB NB This patch has been independently applied to the release branch.

New day dawns

See merge request !46

Gari removes redundant reactions from mole_deuteriaum.dat

See merge request !45

We had not tested master with D turned on. Even the primordial composition
set does not include D. Gargi ran the tsuite suite with D enabled and
found converge aborts and botched monitors on master. She carefully reviewed
the chemistry data and found that many D reactions were duplicated.
She fixed that and I am committing that one set of changes to
this branch.

I will propose merging it to master with D left off. The corrected
D chemistry will not be used by default, just as it is right now.
We will then merge from master with these changes to her HDgargi branch
and continue other HD development there.

There is no functional change in master as a result of this
set of changes, since D is not turned on by default.

Fix bug in run_programs.pl

See merge request !44

The script instructs to run it with, say, sys_llvm.  However, when
queried about compilation flags, the Makefile in sys_* directories
returns 2 lines of output instead of the 1 line returned by the
Makefile in source.  This causes commands to be 4 lines instead of
1 for the sys_* directories, and the compilation to fail.

Capture only the last line of the Makefile output; compilation
commands are 1 line then, and the script works as intended.

Make small improvements re HFS lines

See merge request !41