Log of /trunk

more helpful raditive lifetime report

docs/latex/common/bibliography2.bib
  - Updated NIST ASD reference.

further limit number of elements for load leveling.  There were five to ten percent changes in monitored values due to change in cooling.

varlues have drifted a bit, updated H3+ branching ratio in mole_co_base

tsuite/auto/fix_input_scripts_wl.pl
  - Ported changes to look for suggested wavelength, instead of assuming it is
    always 3 lines after the emitted warning.  First implemented on the Priyanka
    branch in late Aug 2020.

docs/LineLabels.txt
  - Updated to current wavelengths.

test new ISM abundances mixture

ADS link to published Morisset paper

further update to McCall CR-related rates from r14260

remove two CR rates since moved to
mole_co_base.dat and updated to McCall 2004, GS 

rollback r14255-14256 until we understand differences in tsuite better

Gargi Shaw removed old duplicate reaction to silense comment

Trap for reported undefined behaviour

Add const annotation

diagnostics for physical state of sim upon abort or reporting arrays

update H3+ rates to Indriolo+2012

discern big botch gt 9 sigma

fix user suppled density label dat -> sden

double limit on NCHLAB as per Fran comment o thread Buffer Overflow with save species densities

want total (blnd) intensity of line

update CR H2 dissoc as per  2020RNAAS...4...78S

add aastex63.cls, the current version,
and copy needed sumbols to macros.tex

add Dalgarno ref to CR secondaries

data/ionization_potentials.dat
  - Updated ionization potentials to NIST values on May 01, 2020.

source/atmdat_readin.cpp
  - Updated magic number to current version: 20200501.

PS-M non-degenerate case was depending on the species (nelem) instead of on the iso-sequence (ipISO). This is now corrected.

typo fix

added crossreference to Vrinceanu2017

change yahoo to groups.io

update line label list

reintegrated branch IsoFujiandVR62 for Fujimoto and VR62 collisional theories

docs/latex/hazy1/conout.tex
  - Explain that level-resolved output of the 'save species' family of commands
    is reported in the same order as in the input database models (without the
    sorting done internally).

Smyth+19 paper has appeared

add discussion
and figure showing effects of cosmic rays on low density gas

Add another incident continuum point - these really should be exported to an external file.

patch to fix thermal command in iso-sequences

changed monitors from update 12.8 micron cs
We had used very old Hollenbach+ H0 collision rates.  The new values are
significantly smaller.  The electron CS are unchanged.  As a result
the line became a bit weaker in x-ray dominated gas where warm H0 excited
the line.

update e and H0 collision rates for Ne II 12.8 micron
# atomic hydrogen
# Wan, Y., Leiberman, P., Buenker, R., Loch, S. D., Schultz, D. R., & Stancil, P. C.  2019, \apj, 881, 3
# https://ui.adsabs.harvard.edu/abs/2019ApJ...881....3W

# electron collisions <=2000 K
# Wang, Q., Loch, S. D., Li, Y., Pindzola, M. S., Cumbee, R. S., Stancil, P. C., McLaughlin, B. M., & Ballance, C. P.  2017, \mnras, 469, 1225-1232
# https://ui.adsabs.harvard.edu/abs/2017MNRAS.469.1225W
# electron collisions >2000 K
# Electron CS: Griffin,D.C., Mitnik,D.M. & Badnell, N. R.,2001,JPhB,34,4401

the original data file for now

atmdat_chianti.cpp - user friendly error message
ne_2.coll - update H0 and e collisions to  https://ui.adsabs.harvard.edu/abs/2019ApJ...881....3W

atmdat_adfa.h
yield.h
  - Use LIMELM parameter instead of bare 30.

bugfix - trans class not defined when Fe turned off

bugfix for Fe Ka.  We did not include two electron Fe K.  The labels and wavelengths have been update to better represent physics of production and observed wavelength.  The new labels and wavelengths are:
175369	FeKa      1.78000A		# type: i, total intensity of Fe K-alpha line, grain, cold, hot, 1 and 2 electron
175370	FeK1      1.78177A		# type: i, H-like one-electron Ka
175371	FeK2      1.85000A		# type: i, He-like two-electron Ka
175372	FeKH      1.80000A		# type: i, fluorescent hot iron, Fe 18 - 23 times ionized
175373	FeKC      1.75000A		# type: i, fluorescent cold iron, less than or 17 times ionized
175374	FeKG      1.75000A		# type: i, grain production of cold iron

document source for two Fe II bands

make it clear that DATABASE NAME MASTERLIST
replaces the default masterlist rather than supplement it

update LineLabels - comments have changed

cross references for SPHERE command

add two Jin AGN SED papers cited in c19

LineList_BLR_Fe2, Arnab Sakar add missing Fe II bands

FeII_bands.ini, improve comments in refer to Baldwin+04

the four SEDs discussed in Ferland+19 MNRAS submitted
reference will be added when available

two papers on IR CS

line format explain

updated Hazy 1 to the new n-changing collisions options

split He I 2 ^3P_J

rm char some fine offensive

logic assumed that SDIV would not return negative

source/hydro_vs_rates.cpp:
  - Fix FP exceptions when running with -DFLT_IS_DBL.

fix bug on GetHelikeCollisionStrength(), Percival and Richard collisions are not resolved in l

source/grains.cpp:
source/grainvar.cpp:
source/grainvar.h:
  - Remove estimate of grain charge based on variation with Te during search phase.
    This code could lead to a failed assert and probably did not much good anyway.

Heavy ions impact excitation is not treated any more for Percival and Richards 1978 ion-ion collisions. It was treated at the same level as the electrons

update line list with Fe III multiplet

fix missed xref to deleted command, give correct

Merge cdTemp branch back into mainline

increased number of levels for Fe III, drift over time, use mean of grid_extreme values

enough levels to caputer Fe III UV34

source/grains_mie.cpp:
  - Add support for the Maxwell Garnett (1904) EMT.
docs/latex/common/bibliography2.bib
docs/latex/hazy1/vanhoof_grain_model.tex
  - Describe the new EMT.
data/fluffy.mix:
  - Update the EMT to a more sensible choice.
  - Fix comment.

Maryam add new Lodders SS abundances, 
use them in two test cases

source/hydro_vs_rates.cpp:
  - Fix FP exception when running with -DFLT_IS_DBL.

source/atom_hyperfine.cpp
  - Revert to setting Lya occupation number to 0 if negative.  Add more
    detailed explanation to fixit().
tsuite/programs/hizlte/hizlte.txt
  - Update output of hizlte sim.

source/cpu.h:
  - Let NORETURN macro support the Oracle Studio compiler as well.

the Tayal 2018 dataset

changed monitors to ratio with case b Hummer and Storey predictions

added case b simulations for 30000K: the highest temperature of Hummer & Storey

add Bautista and Smyth data sets

data/*.opc:
  - Update grain opacity files as a result of r12781.

source/atom_hyperfine.cpp
  - Remove if-test against non-existent flag conv.lgLyaMaseCommentDone.

source/parse_table.cpp:
  - Bug-fix: pad keyword with spaces so that unit is always correctly recognized.
source/parser.cpp:
  - Use real path to data file in error messages.
source/cpu.cpp:
source/cpu.h:
  - Export actual location of data file via open_data() if requested.
data/SED/pl1.sed:
  - Extend data range.
  - Exercise NUFNU and UNITS keywords.

source/Makefile:
  - Bug-fix: when creating dependencies, we should also use the -std=c++11 flag.

Add max field width to quieten g++ 9.1.1 warning

BPT paper

source/parse_agn.cpp:
  - Fix bug reported by David Carr: optimizing the AGN command sets wrong initial
    temperature.

source/helike_einsta.cpp:
source/helike_einsta.h:
source/iso.cpp:
source/iso.h:
source/iso_create.cpp:
source/parse_atom_iso.cpp:
source/prt_lines_helium.cpp:
  - Remove code for He-like fine-structure mixing. This code was not working
    correctly and would also not have any impact for low-Z ions. The situation
    is different for high-Z ions like e.g. Fe XXV where the individual fine-
    structure components can be spectroscopically resolved, but that requires
    major changes in the code. See ticket #425 for details.

docs/latex/hazy1/compos.tex:
docs/latex/hazy2/observed.tex:
  - Make some minor improvements to the discussion of the emergent spectrum.

source/abund.h:
source/abundances.cpp:
source/elementnames.cpp:
source/elementnames.h:
source/init_defaults_preparse.cpp:
source/input.cpp:
source/input.h:
source/parse_abundances.cpp:
source/parse_element.cpp:
source/parser.cpp:
source/parser.h:
  - Rewrite ABUNDANCES command such that the element symbol is mandatory to
    identify the element the abundance belongs to. This also implies that
    abundances can now be given in arbitrary order and the list of elements
    need not be complete.
  - Remove the ELEMENTS READ command. This is no longer needed as the info
    it gathers is now implicitly contained in the ABUNDANCES command.
  - Modify routine nelem_symbol_to_index() to be completely case-insensitive.
  - Create new methods t_input::peekarray() and Parser::peekNextCommand() to
    test what next command will be (needed to check for CONTINUE lines).
data/c84.ini:
data/fast.ini:
data/ism.ini:
docs/latex/hazy1/cmdintro.tex:
  - Remove ELEMENTS READ command.
docs/latex/hazy1/compos.tex:
  - Update documentation of the ABUNDANCES command.
  - Remove description of the ELEMENTS READ command.
  - Fix typo in the "Variable grain abundances" section.
tsuite/auto/*.in:
tsuite/slow/*.in:
  - Update syntax of ABUNDANCES commands (include '=').

two bug fixes

source/prt_lines.cpp:
docs/latex/hazy1/miscell.tex:
  - Support molecular species when using SET BLEND QUIET.
source/prt_lines_molecules.cpp:
data/blends.ini:
  - Move H2 blend to blends.ini.

docs/latex/hazy1/conout.tex:
source/mole_h2_io.cpp:
  - Fix incorrect description of the SAVE H2 PDR command.

source/cddrive.cpp:
source/maincl.cpp:
  - Bug-fix: this is the third stab at reading the standard blends. The first
    implementation in r12725 didn't work in optimizer runs, and likely also not
    when running Cloudy as a subroutine. The second implementation in r12727
    fixed the first problem but not the second. This third implementation should
    work in all setups.

source/maincl.cpp:
source/parse_commands.cpp:
source/parse_init.cpp:
source/parser.h:
  - Bug-fix: the solution for reading the blend file adopted in r12725 doesn't
    work in optimizer runs. Fix that here.

orion paper

source/parse_save.cpp:
  - Always save to separate files when using the SAVE TRANSMITTED CONTINUUM
    command in a grid run. The code cannot read concatenated files back in.
source/parse_table.cpp:
  - Rewrite parser of the SAVE TRANSMITTED CONTINUUM using DataParser. This
    avoids problems when reading output generated in a grid run where the top
    comment line is missing for the second and following grid points.
source/save_do.cpp:
  - Ignore the UNITS keyword when writing SAVE TRANSMITTED CONTINUUM output.
    Support for that was not complete and the option is not really needed
    anyway.
docs/latex/hazy1/conout.tex:
  - Update documentation of the SAVE TRANSMITTED CONTINUUM command.

docs/latex/hazy1/cmdintro.tex:
  - Remove weird invisible character.

a work in progress

source/lines.cpp:
source/prt_lines_helium.cpp:
  - Optimize setting up He-like blends by avoiding the use of findline().

tsuite/auto/func_LineList_BLR.in:
tsuite/auto/func_LineList_BLR_Fe2.in:
tsuite/auto/func_LineList_HII.in:
tsuite/auto/func_LineList_HeH.in:
tsuite/auto/func_LineList_He_like.in:
tsuite/auto/func_LineList_NLR.in:
tsuite/auto/func_LineList_PDR.in:
tsuite/auto/func_LineList_PDR_H2.in:
tsuite/auto/func_LineList_strong.in:
tsuite/auto/h2_pdr_leiden_f1.in:
  - Missed these scripts in r12703. They also need blends.ini as they read line
    list files.

head_info.tex - silence newly installed grumpiness in LaTex 2018

compos.tex - improved emergent spectrum discussion in section 7.9.6

observed.tex -  improved emergent spectrum discussion in section 2.10

stopping.tex - add label for xref

copyright year, to be parallel with code copyright

improve comments on NGC 5548 SED file

source/save_do.cpp:
  - Fix incorrect return value.

docs/latex/common/bibliography2.bib:
docs/latex/hazy1/cont-shp.tex:
  - Update reference to Khaire & Srianand (2019).

source/cont_ipoint.cpp:
source/opacity_createall.cpp:
source/sanity_check.cpp:
  - Fix various tests in the code as discussed in ticket #421.
  - In ipLineEnergy(), allow line to occupy the same frequency cell as the
    ionization edge.

source/atmdat_char_tran.cpp:
  - Move last of the Cobalt Blue translated common blocks into external data files.
data/ctiondata.dat:
data/ctrecombdata.dat:
  - New data files holding the charge transfer data.

Updated NIST reference

tsuite/auto/blr_level2.in:
tsuite/auto/rrc_o.in:
  - Fix wavelengths of RRC lines as a result of r12609.

updated bibliography

This commit implements line disambiguation by optionally supplying either the
lower and upper level index, or the energy of the lower level. This is supported
by the following commands:

SAVE LINES EMISSIVITY
SAVE LINE LIST
PRINT LINE SUM
TABLE LINES
MONITOR LINE
OPTIMIZE LINES

The command SAVE LINE LABELS has been modified to include the necessary info
needed to do the line disambiguation. Also the following routines support line
disambiguation by accepting a parameter of type LineID:

LinSv::addComponent()
t_LineSave::findline()
cdLine()
cdEmis()
cdGetLineList()
prt_line_err()

Additionaly some line labels were modified to aid with disambiguation:

GrGC -> GrCC, GrCH
Inwd -> InwdBnd (for continuum bands)


source/atmdat.h:
source/atmdat_chianti.cpp:
source/atmdat_lamda.cpp:
source/prt.cpp:
source/prt_lines_hydro.cpp:
  - Update comment for SAVE LINE LABELS output.
source/cddrive.cpp:
source/cddrive.h:
source/cdgetlinelist.cpp:
source/prt.h:
  - Add interfaces that accept a LineID.
source/conv_itercheck.cpp:
  - Fix capitalization.
source/cool_dima.cpp:
  - Bug-fix: we should test that level 2 lines have upper level energy larger
    than the highest level in Stout/Chianti. This is still not safe as level 2
    lines have inaccurate OP energies.
source/init_defaults_preparse.cpp:
source/iter_startend.cpp:
source/lines.cpp:
source/lines.h:
source/monitor_results.cpp:
source/optimize.h:
source/optimize_func.cpp:
source/parse_optimize.cpp:
source/parse_save.cpp:
source/parse_table.cpp:
source/parser.cpp:
source/parser.h:
source/prt_linesum.cpp:
source/save.h:
source/save_do.cpp:
source/save_line.cpp:
  - Implement line disambiguation.
source/lines_service.cpp:
  - Fix typos.
source/prt_final.cpp:
  - Bug-fix: Change "He 1" 5875.64 -> "Blnd" 5875.66.
source/prt_lines_grains.cpp:
  - Change GrGC -> GrCC, GrCH.
source/prt_lines_helium.cpp:
  - Bug-fix: remove duplicate He 1 lines from the line stack.
source/prt_lines_molecules.cpp:
  - Remove trailing spaces in line labels.
source/species_pseudo_cont.cpp:
  - Change Inwd -> InwdBnd for continuum bands.
source/taulines.h:
  - Remove unused constant.
source/TestDataParser.cpp:
  - Update test of getLineID().
source/TestParser.cpp:
  - Add test of getLineID().
docs/latex/hazy1/conout.tex:
docs/latex/hazy1/define.tex:
docs/latex/hazy1/miscell.tex:
docs/latex/hazy1/optimize.tex:
  - Document the new syntax.
  - Document removal of the limits on the number of lines for the PRINT LINE SUM
    and SAVE LINES EMISSIVITY command.
  - Document the inward component of continuum bands.
docs/latex/hazy2/sub.tex:
  - Document the modified API for cdGetLineList().
  - Document the additional API for cdLine().
tsuite/auto/feii_t4n8_rt.in:
tsuite/auto/grains_hot.in:
tsuite/auto/grains_hot_wd01.in:
tsuite/auto/ism_opacity.in:
tsuite/auto/limit_caseb_h_den2_temp37.in:
tsuite/auto/limit_caseb_h_den2_temp4.in:
tsuite/auto/limit_caseb_h_den2_temp5.in:
tsuite/auto/limit_caseb_h_den2_temp5_coronal.in:
tsuite/auto/limit_caseb_he2_den8.in:
tsuite/auto/orion_hii_open.in:
tsuite/programs/mpi/mpi.cpp:
  - Disambiguate lines where needed.
  - Change GrGC -> GrCC.
  - Change Inwd -> InwdBnd for Fe 2 continuum bands.
data/LineList_BLR.dat:
data/LineList_BLR_Fe2.dat:
data/LineList_NLR.dat:
  - Disambiguate the Al 2 1670.79 line.
  - Change Inwd -> InwdBnd for Fe 2 continuum bands.

source/iso_level.cpp:
  - Bug-fix. See comment 8 in PR #358 for a detailed discussion.
tsuite/auto/pdr_minimal.in:
  - Add new test suite sim to guard against regression.

source/parse_metal.cpp:
  - Fix bug where METALS DEPLETE command would reset abund.ScaleMetals to 1. This
    made the effect of the METALS DEPLETE and the regular METALS command dependent
    on the sequence of the two commands.
docs/latex/hazy1/compos.tex:
  - Add clarification that the METALS DEPLETE and the regular METALS command can
    be combined to modify the depletion effect.

source/input.cpp:
source/input.h:
source/monitor_results.cpp:
source/parse_grid.cpp:
  - The vector form of input_readvector() was far too restrictive concerning the
    format of the file it was reading. Replace routine with new version using
    the DataParser so that comments are allowed now.
  - Remove old-style array version of input_readvector().
docs/latex/hazy1/grid.tex:
docs/latex/hazy1/miscell.tex:
  - Document that arrays are allowed now.

source/parse_table.cpp:
  - Fix bug where TABLE SED files that have wavelengths in increasing order are
    rejected. This is not documented behavior.
docs/latex/hazy1/cont-shp.tex:
  - Explicitly state that both increasing and decreasing order of frequencies or
    wavelengths are allowed.
data/SED/Rubin.sed:
  - Reorder this file so that is has wavelengths in increasing order and
    exercises the new code.

tsuite/auto/func_grid_list.in:
  - This sim should have been sequential.

source/cddefines.h:
  - Start using nullptr.

source/conv_init_solution.cpp:
source/dense.cpp:
source/dynamics.cpp:
source/ion_trim.cpp:
source/mole.h:
source/mole_drive.cpp:
source/mole_eval_balance.cpp:
source/mole_priv.h:
source/mole_reactions.cpp:
source/mole_solve.cpp:
source/mole_species.cpp:
source/parse_species.cpp:
source/parser.h:
source/prt_zone.cpp:
source/radius_next.cpp:
source/save_line.cpp:
source/transition.h:
  - Replace count_ptr with shared_ptr.
  - Replace naked pointer in chem_nuclide with weak_ptr.
  - Add method el() to chem_nuclide to access weak pointer.
source/count_ptr.h:
  - Deleted.

docs/latex/hazy1/ControllingModelAtoms.tex:
docs/latex/hazy1/cmdintro.tex:
docs/latex/hazy1/conout.tex:
source/TestDataParser.cpp:
source/TestThirdparty.cpp:
source/atmdat_chianti.cpp:
source/cddrive.cpp:
source/cdgetlinelist.cpp:
source/cpu.h:
source/helike_einsta.cpp:
source/iso.cpp:
source/parse_abundances.cpp:
source/parse_commands.cpp:
source/parse_interp.cpp:
source/parser.cpp:
source/parser.h:
source/prt_lines_helium.cpp:
source/thirdparty.cpp:
source/thirdparty.h:
tsuite/auto/agn_lex00_u0.in:
tsuite/auto/agn_lex00_u1.in:
tsuite/auto/agn_lex00_um1.in:
tsuite/auto/agn_warm_absorber_hiU.in:
tsuite/auto/blr_kk81.in:
tsuite/auto/blr_rnfa.in:
tsuite/auto/blr_rnfb.in:
tsuite/auto/coll_t5.in:
tsuite/auto/dynamics_wind.in:
tsuite/auto/func_abund_fluc.in:
tsuite/auto/func_stopline2.in:
tsuite/auto/hii_Um3.in:
tsuite/auto/hii_Um35.in:
tsuite/auto/hii_Um385.in:
tsuite/auto/limit_caseb_he_den2_temp4.in:
tsuite/auto/limit_caseb_he_den4_temp4.in:
tsuite/auto/limit_conserve.in:
tsuite/auto/limit_lte_he1_coll.in:
tsuite/auto/limit_lte_he1_nomole_iso_force.in:
tsuite/auto/limit_lte_he1_nomole_ste.in:
tsuite/auto/limit_lte_he1_nomole_ste_nocoll.in:
tsuite/auto/limit_lte_he1_nomole_ste_nocoll2.in:
tsuite/auto/limit_lte_he1_ste.in:
tsuite/crash/limit_lte_he1_nomole_ste_nocoll2.in:
tsuite/experimental/cosmo_recom.in:
tsuite/mpi/func_test_cycle2.in:
tsuite/mpi/run_parallel.pl:
  - This is the final merge from the (now poorly named) newcomment branch. This
    implements the first update of the input parser to make it check the scripts
    more strictly. From now on all characters in a command name that are typed
    will be checked. Abbreviating commands is still allowed, and the code now
    automatically determines the shortest unambiguous abbreviation (with a
    minimum of four. If a command is not recognized, but close to other commands,
    these will be suggested as alternatives.

source/atmdat.h:
source/atmdat_chianti.cpp:
source/atmdat_lamda.cpp:
source/atmdat_readin.cpp:
source/atom_hyperfine.cpp:
source/cddefines.h:
source/cddrive.cpp:
source/cdinit.cpp:
source/cont_createmesh.cpp:
source/cont_createpointers.cpp:
source/cont_setintensity.cpp:
source/continuum.h:
source/geometry.h:
source/grains_qheat.cpp:
source/grid.h:
source/grid_xspec.cpp:
source/hcmap.cpp:
source/hcmap.h:
source/hyperfine.h:
source/init_coreload_postparse.cpp:
source/ion_solver.cpp:
source/iso.cpp:
source/iso.h:
source/iso_continuum_lower.cpp:
source/iso_cool.cpp:
source/iso_create.cpp:
source/iso_solve.cpp:
source/iterations.cpp:
source/lines.h:
source/lines_service.cpp:
source/mole.h:
source/mole_h2_create.cpp:
source/mole_reactions.cpp:
source/mole_species.cpp:
source/monitor_results.cpp:
source/opacity.h:
source/parse_atom_iso.cpp:
source/parse_crashdo.cpp:
source/parse_save.cpp:
source/parse_table.cpp:
source/prt_lines_continuum.cpp:
source/prt_linesum.cpp:
source/rfield.h:
source/rt_diffuse.cpp:
source/sanity_check.cpp:
source/save_average.cpp:
source/save_do.cpp:
source/save_fits.cpp:
source/secondaries.cpp:
source/secondaries.h:
source/service.cpp:
source/species.cpp:
source/species2.cpp:
source/taulines.h:
source/temp_change.cpp:
source/timesc.h:
source/zero.cpp:
  - Remove all remaining instances of MALLOC. Apart from specially aligned memory
    for SIMD instructions, we no longer use malloc(). The code is now leak-free!
  - The CRASH UNDEFINED STATIC command has been removed. Since static memory is
    always initialized, this command makes no sense.
  - The default CRASH UNDEFINED command has also been removed as we no longer
    allocate raw data on the heap.
  - The CRASH UNDEFINED STACK | AUTO command has been renamed to plain CRASH
    UNDEFINED as it is now the only variant left.
  - The CRASH BOUNDS HEAP command has been removed since we no longer allocate
    arrays this way.
  - The CRASH BOUNDS ARRAY command has been added to test the STL class "array".
tsuite/crash:
  - The crash test suite has been updated to reflect the changes above.
docs/latex/hazy1/miscell.tex:
  - Document the changes in the CRASH commands.

source/atmdat_readin.cpp:
source/atom_leveln.cpp:
source/atoms.h:
source/cddefines.h:
source/cont_createmesh.cpp:
source/dynamics.cpp:
source/dynamics.h:
source/h2.cpp:
source/h2_priv.h:
source/init_sim_postparse.cpp:
source/ion_recomb_Badnell.cpp:
source/iso_create.cpp:
source/iter_startend.cpp:
source/iterations.cpp:
source/iterations.h:
source/mole.cpp:
source/mole.h:
source/mole_h2.cpp:
source/parse_commands.cpp:
source/parse_print.cpp:
source/parse_radius.cpp:
source/parse_set.cpp:
source/parse_stop.cpp:
source/species2.cpp:
  - Remove more instances of MALLOC().

source/ionbal.cpp:
source/ionbal.h:
  - Remove more calls to MALLOC().

source/iso.cpp:
source/iso.h:
source/parse_atom_iso.cpp:
source/prt_lines_hydro.cpp:
  - Create new command DATABASE H-LIKE KEEP FINE STRUCTURE to retain the fine
    structure of H-like lines on the line stack.
docs/latex/hazy1/ControllingModelAtoms.tex
  - Document the new command.
tsuite/auto/limit_caseb_h_den2_temp37.in:
  - Exercise the new command and monitor some fine structure lines.

source/monitor_results.cpp:
  - Fix alignment of headers.

source/helike_einsta.cpp:
source/helike_recom.cpp:
source/ion_recomb_Badnell.cpp:
source/iso_create.cpp:
source/iso_radiative_recomb.cpp:
source/iso_solve.cpp:
source/prt_alltau.cpp:
source/prt_comment.cpp:
source/prt_header.cpp:
source/prt_lines_hydro.cpp:
source/prt_zone.cpp:
source/rt_tau_init.cpp:
source/sanity_check.cpp:
  - Replace QuantumNumbers2Index[] with QN2Index().
  - Don't use data that are hardwired to a specific level sequence.
  - Fix incorrect test of quantum numbers in ion_recomb_Badnell.cpp.
source/iso.h:
source/iso.cpp:
source/helike_energy.cpp:
  - Modify QN2Index() such that it also matches unresolved levels with resolved
    parameters. This makes it more compatible with QuantumNumbers2Index[].
  - Refactor QN2Index() to take g as parameter, rather than j.
  - Make QN2ind() safer to use.
  - Refactor energy methods to take g as parameter, rather than j. Otherwise
    they cannot be used for H-like species.
  - Add new method hydro_energy().
source/prt_lines_helium.cpp:
tsuite/auto/pn_paris_fast.in:
  - Fix bug in generating wavelengths for He I Ca B lines.
  - Don't use data that are hardwired to a specific level sequence.
  - Replace QuantumNumbers2Index[] with QN2Index().
  - Refactor GetStandardHeLines() to use the DataParser.
  - Fix test suite sim to use the correct Ca B wavelength.
source/parse_commands.cpp:
  - The NO TIME command should suppress printing the PRNG seed.
docs/latex/hazy1/ControllingModelAtoms.tex:
  - Add caution when using small model atom with He-like species.
source/helike.h:
data/he1_case_b.dat:
  - Update magic number.
  - Update wavelengths.
  - Change level indices to quantum numbers.

source/rt_continuum_shield_fcn.cpp:
  - Remove two unused classes. These were causing compilation warnings with the
    Intel compiler.

source/hydrocollid.cpp:
  - Remove dependency on QuantumNumbers2Index[] by using quantum numbers instead
    of level indices.
source/atmdat_adfa.cpp:
source/atmdat_adfa.h:
  - Refactor routines working with hydrogen collisional data to use quantum
    numbers instead of level indices.
data/h_coll_str.dat:
  - Change level indices to quantum numbers.

source/helike_einsta.cpp;
  - Refactor he_1trans() such that it no longer needs ipLo, ipHi.
  - Remove getG() as the source of the bogus J values has been fixed.
  - Start converting ForbiddenAuls() to remove dependency on ipLo, ipHi.
source/helike.h:
source/helike_energy.cpp:
data/helike_energies.dat:
  - Move He-like level energies and ionization potentials into separate file
    (helike_energies.dat). The data are now stored in the iso_sp class.
  - Refactor helike_energy() such that it no longer needs ipLo, ipHi.
source/iso.cpp:
source/iso.h:
  - Add routine for reading helike_energies.dat called iso_init_energies().
  - Add methods for retrieving the energies for a given set of quantum numbers
    from iso_sp, called energy() and energy_ioniz().
source/iso_create.cpp:
  - Call iso_init_energies().
source/helike_cs.cpp:
  - Improve error checking.
source/parser.h:
source/TestDataParser.cpp:
  - Add new checkEOF() method.
source/physconst_template.h:
  - Add constants ELECTRON_MASS_U, PROTON_MASS_U, ALPHA_MASS_U, H_RYD_FACTOR,
    and HE_RYD_FACTOR.

source/helike_cs.cpp:
  - Refactor HeCS array into map so that it no longer depends on QuantumNumbers2Index.
  - Rewrite parser for he1_cs.dat using DataParser.
source/helike_cs.h:
  - Do not export CSTemp and HeCS. This is not needed.
source/helike_einsta.cpp:
source/hydro_bauman.cpp:
source/iso.h:
  - Refactor Index into QNPair and QNPack structs.
  - Get rid of preprocessor macros.
source/iso_create.cpp:
  - Add routine getL() for converting e.g. 'P' -> 1.
source/parser.h:
  - Fix typo in comment.
source/thirdparty.h:
source/thirdparty_interpolate.cpp:
  - Refactor linint() so that it can work on realnum arrays as well.
source/helike.h:
  - Get rid of preprocessor macros.
  - Update magic number for he1_cs.dat.
data/he1_cs.dat:
  - Added J values in the file.

This is the third and final commit from the backtrace branch. Here the lgAbort
variable is replaced with throwing a cloudy_abort. This cleans up the code and
prevents it from producing lots of meaningless output when it soldiers on after
the abort. This behavior is no longer needed as we have a new method to create
stub output files for grid points that failed.

This commit also works around a mysterious bug where the headers of the gathered
output of a grid run would be missing. This was caused by the fact that a file
that was opened was positioned at the start, despite being opened in append mode.
This is worked around by explicitly positioning the file at the end. This should
not be needed, but it fixes he problem...

source/lines.cpp:
  - In r12481 the test suite slowed down by about 1%. The test case
    grid_h2coronal.in was badly affected. This is an attempt to fix this.

source/ran.cpp:
source/ran.h:
source/thirdparty.cpp:
  - Bug-fix: fill_ran_pool() was hardwired to the global ran instantiation. This
    violates the intended design where it is OK to have multiple local
    instantiations. Remove this routine and pass a pointer to the routine for
    filling the pool to the ran_pool class instead.
  - Upgrade the xoroshiro128+ algorithm to v1.0. We had the 2016 version.
  - Clean up the classes and fix some errors in the comments.
source/TestRandom.cpp:
  - Add test to check that random number sequence is as expected. This validates
    that the PRNG algorithms in Cloudy are still identical to the ones posted by
    Blackman & Vigna.

source/ran.cpp:
source/ran.h:
source/thirdparty.cpp:
source/thirdparty.h:
  - Add second PRNG algorithm named xoshiro256**. Blackman & Vigna now promote
    this as their all-purpose, rock-solid generator. So it would be strange if
    Cloudy did not support it. The previous algorithm xoroshiro128+ is retained
    as a second option.
source/ran.h:
source/mpi_utilities.cpp:
source/optimize_phymir.cpp:
  - Rename t_ran::reseed() -> t_ran::new_rank() as this method does not change
    the seed, so the old name was inappropriate.
  - Replace preprocessor functions with inline functions.
source/capabilities.pl:
  - Remove obsolete tests for g++ version as our minimum requirement is 4.8.1.
docs/latex/hazy1/cmdintro.tex:
  - Document the change in PRNG algorithm.

bibliography2.bib:
  - Add Marsaglia & Tsang (2000).

source/ran.cpp:
source/ran.h:
source/mc_escape.cpp:
source/TestRandom.cpp:
source/TestVectorize.cpp:
  - Fix compilation problems with -DFLT_IS_DBL in effect.
  - Rename ran.flt() -> ran.rnm() to make it clearer that it returns a realnum.
  - Reduce size of unit tests for random numbers to speed them up.

source/prt.cpp:
  - Remove print statement. This was producing output during unit testing, which
    should be avoided. In this case the statement was not concidered to be
    sufficiently important to retain, so the simple fix is to remove it.

source/mpi_utilities.cpp:
source/optimize_phymir.cpp:
source/ran.cpp:
source/ran.h:
  - The random number class would seed MPI ranks correctly (i.e. making sure
    that each rank generates a different sequence). This was not the case for
    forked runs though, so fix that here.
  - Bug fix: aligned memory should be freed with posix_memalign_free().
  - Bug fix: make sure that the command parser only seeds the random generator
    once. Poor randomness could otherwise result in grid or optimizer runs.
  - Delete copy ctor and assignment operator for ran_pool and t_ran classes.
    These would not be safe.
  - Use /dev/urandom instead of local time (when available) to create random
    seed. This works better if multiple Cloudy jobs are started simultaneously
    in a script. They might otherwise get the same seed.
source/vectorize.h:
source/cpu.h::
  - Move definitions of posix_memalign() and posix_memalign_free() from
    vectorize.h to cpu.h since that seems a better place for them.
source/cloudyconfig_vs.h:
source/configure.sh:
  - Add test to determine if /dev/urandom is available to obtain a seed.

source/hydro_bauman.cpp:
  - Fix compiler warning about unused variable when compiling with -DNDEBUG.

source/TestMain.cpp:
source/TestMesh.cpp:
source/TestRandom.cpp:
source/TestService.cpp:
source/TestThirdparty.cpp:
source/TestVectorize.cpp:
source/cddefines.h:
source/cdinit.cpp:
source/cdstd.h:
source/cpu.h:
source/ion_recomb_Badnell.cpp:
source/iso.h:
source/mc_escape.cpp:
source/mole_h2_create.cpp:
source/mole_h2_io.cpp:
source/parse_atom_h2.cpp:
source/parse_atom_iso.cpp:
source/parse_crashdo.cpp:
source/parse_set.cpp:
source/ran.cpp:
source/ran.h:
source/random_crash.cpp:
source/service.cpp:
source/thirdparty.cpp:
source/thirdparty.h:
  - This is a partial merge from the spindust branch to bring in the new random
    number generator class. The core of the class is in the files ran.h and
    ran.cpp. The existing routines based on the Mersenne twister algorithm have
    been removed and the code is moved over to the new class where possible. The
    code is based on the xoroshiro128+ algorithm which is both faster and better
    than Mersenne twister. The code is fully vectorized on supported platforms
    and is also aware of parallelization making sure that every rank generates a
    different random number sequence. RandGauss() has been replaced by a faster
    version based on the Ziggurat algorithm.
  - Add unit tests for the new random number infrastructure.
  - Add an improved implementation of the Dawson function.

source/hydro_bauman.cpp:
source/service.cpp:
source/service.h:
  - Speed optimization of H_Einstein_A(). The new version is almost a factor of
    five faster.

source/atmdat_gaunt.cpp:
source/atmdat_gaunt.h:
source/init_defaults_preparse.cpp:
source/parse_commands.cpp:
source/parse_drive.cpp:
source/parser.h:
source/prt_lines.cpp:
source/rt.h:
source/rt_continuum_shield_fcn.cpp:
source/trace.h:
  - Remove all DRIVE commands. They are obsolete and may no longer work.
docs/latex/hazy1/miscell.tex:
  - Remove the documentation.

source/cddefines.h:
source/cpu.cpp:
source/cpu.h:
source/maincl.cpp:
source/parse_print.cpp:
source/parse_set.cpp:
source/service.cpp:
  - Remove the '-a' command line flag, the SET ASSERT ABORT command, and the
    OLD_ASSERT compiler flag. These all had the aim to modify the ASSERT macro
    such that it aborts rather than throw a C++ exception. This behavior was
    deprecated.
docs/latex/hazy1/miscell.tex:
tsuite/crash/ReadMe.txt:
  - Document the changes.
tsuite/crash/assert_abort.in:
  - Deleted.

source/sys_llvm/Makefile.conf:
  - Silence warning about unused command line arguments.

tsuite/crash/*.in:
  - Make sure scripts are complete and up to date.
tsuite/crash/runall.pl:
  - Deleted.
tsuite/crash/run_parallel.pl:
tsuite/crash/run_sequence.pl:
tsuite/crash/checkall.pl:
  - Use standard scripts for running the test suite.
tsuite/crash/ReadMe.txt:
  - Update description of files.

source/Makefile:
  - Fix bashism in Bourne script. This would not run with /bin/dash (e.g.,
    under Ubuntu 18.04, /bin/sh is linked to /bin/dash).
source/configure.sh:
  - Disable creating position independent executables. This breaks addr2line.

source/iter_end_chk.cpp:
  - Remove inappropriate call to TotalInsanity().

This is the first merge from the backtrace branch. This implements a proper
signal handler for various exceptions like floating point exceptions and
segmentation violations. These are then rethrown as a C++ exception so that grid
runs can recover from these. On supporting platforms a traceback at the point of
failure will also be generated.

This merge includes the following noteworthy changes:

- The commands CRASH SEGFAULT, CHASH ABORT, and CRASH GRID were added.
- A C++ exception cloudy_abort was added. This is currently not used (apart from
  the CRASH ABORT command) but should in time replace the lgAbort flag. Also the
  C++ exception bad_signal was added to handle OS exceptions like segfaults.
- The routine cdBacktrace() was added. This is intended for debugging and allows
  a non-invasive backtrace to be generated (meaning that execution will continue
  normally after cdBacktrace() exits). This is useful to figure out what code
  path was taken to get to the place where the cdBacktrace() call was inserted.
  Since it is non-invasive, cdBacktrace() can be called multiple times in a run.
- The main program has been modified to catch the new C++ exceptions. This will
  also affect all programs that call Cloudy as a subroutine.

source/cdinit.cpp:
  - Fix bug introduced by r12399.

source/hydro_tbl.cpp:
  - Fix use of undefined type uint.

source/parse_save.cpp:
  - Clarify header of the SAVE CONTINUUM BINS command.

source/iso.h:
source/hydro_bauman.cpp:
source/helike_einsta.cpp:
  - Also cache Auls calculated by H_Einstein_A_log10().

source/hydro_tbl.cpp:
  - Fix compiler warning in ndebug runs.

source/iso.h:
source/helike_einsta.cpp:
source/hydroeinsta.cpp:
source/iso_collide.cpp:
source/iso_create.cpp:
  - Delete the CachedAs array. Replacing the Bauman routines with tabulated data
    removed the strongest need for this array and has made it largely redundant.
    The implementation via CachedAs was fragile and got in the way of current
    development on the condensed branch. This change will lead to additional
    calls to he_1trans() and H_Einstein_A(). That most likely will not produce
    a significant increase in CPU time, but if it does a different solution for
    caching the data will have to be devised.

source/hydro_tbl.h:
source/hydro_tbl.cpp:
  - Revert r12357 and implement a better solution following the discussion on
    https://en.cppreference.com/w/cpp/language/static under "Constant static
    members". The constant p_stride can be odr-used as a parameter in a call
    to min(). Hence it needs a declaration at namespace scope, but without an
    initializer.

Until now we calculated photoionization cross sections and Einstein A values for
hydrogenic ions on the fly using the routines written by Robert Bauman. These
routines are very accurate and stable, but are also very slow (especially for
high values of the principal quantum number). In this commit we put these data
in external data files to circumvent the costly calculation, thus speeding up
the code significantly for large H- or He-like model atoms.

source/hydro_tbl.h:
source/hydro_tbl.cpp:
  - New class for holding tabulated data for hydrogenic ions.
source/hydro_bauman.h:
source/hydro_bauman.cpp:
  - Remove linear versions of routines for calculating photo cross sections
    or Einstein A values. These are no longer needed as these data have now
    been tabulated. Use these instead.
  - Remove routines for calculating the oscilator strengths of hydrogenic
    transitions. These were never used.
  - Vectorize the routine prodxx_log10().
source/hydroeinsta.cpp:
  - Add code for inserting Einstein A values for n,l -> n,l' transitions, but
    disable it for now as it trips SanityCheck(). Most likely the indices in
    QuantumNumbers2Index[] are no longer correct.
  - Remove hardwired Einstein A value for Lya lines.
source/parser.h:
  - Add methods for getting and setting the position in the data file.
source/prt_lines_hydro.cpp:
  - Merge collectSkipTrans() into lines_hydro(). This significantly speeds up
    the code for large hydrogenic models.
data/hydro_pcs.dat:
  - Photoionization cross sections for hydrogenic ions.
data/hydro_tpn.dat:
data/hydro_tpnl.dat:
data/hydro_tpnn.dat:
  - Transition probabilities for hydrogenic ions.

species.cpp
  - Code that converts the Stout / Chianti basename into a spectral label
    is now refactored into a function, db_basename_to_spectral().
    The new function converts, for instance, "c_2" to "C  2".
species.h
  - Expose the interface for db_basename_to_spectral().
  - Add Doxygen comments to spectral_to_chemical() and chemical_to_spectral().

source/cddefines.h:
source/cloudyconfig_vs.h:
source/configure.sh:
source/cpu.h:
source/parse_print.cpp:
source/thirdparty.cpp:
source/thirdparty.h:
  - Remove local definitions of __func__, erf(), and erfc(). These are now
    guaranteed by the C++11 standard.

source/species_pseudo_cont.cpp:
  - Add virtual destructor to fix warnings from clang++ 6.0.1.

data/stout/p/p_2/p_2.coll:
  - Update P II collision strengths to Tayal (2004).
tsuite/auto/blr_level2.in:
  - Update monitors.

docs/latex/hazy1/cmdintro.tex:
  - Further clarify the way the search path works.

the tag \rmxaa did not work, fixed with the full name of the journal

corrected hazy1, it is needed to specify iso-sequence in masers off command

explicitly state that level indices count from 1 not 0

ref to published version existed far below reference to aerXiv - now only one, but yet another alias

update reference for C17, add c17 alias

Reducied duplicated code in parse_print.cpp (Robin's solution)

source/atmdat_chianti.cpp:
source/parse_species.cpp:
  - The SPECIES "Fe+6" LEVELS command was broken in r12144. Fix this.

effects of SPHERE command on line optical depths

Fix zoning botches which were found on wolkje but not cloud9

Update (zone count) monitors for bugfixes at front face r12134:12136

radius_next() needs to be called in zone 0 to set base values

Remove monitor for blend with populations which underflow trimming limit

prep-bib.sh
  - On MacOS, set number of threads for compiling Cloudy to number
    of physical cores.
    (Hyperthreads on Mac sometimes cause the machine to freeze.)

Switch to modified ion trimming scheme as default

Move NewTrim control variable to ionbal class

scripts/db-ref-bib2json.pl -- Add comments about the error files that may be generated.

tsuite/programs/comp4/comp4.cpp:
  - Prevent warnings about grains being above the sublimation temperature.

Mop up dynamics sources below IonLow for consistency

Tweaks to revised trimming approach (not active by default)

source/parse_commands.cpp:
  - Fix optimization for the EDEN command.
docs/latex/hazy1/cont-shp.tex:
docs/latex/hazy1/miscell.tex:
docs/latex/hazy1/optimize.tex:
  - Fix documentation of the optimizer.

add xref for SET TEMPERATURE

Improve language on the 'limit' option of 'save line optical depths'.

Improve discussion of 'save line optical depth some' to discuss how the direction along which the optical depth is accumulated matters.

docs/latex/hazy1/conout.tex: Clarify that damping constant calculation in 'save line optical depth' is over the last zone, whereas the optical depth is for the whole cloud

docs/latex/hazy2/lineatomparam.tex:
  - Update description of the Voigt profile and mean optical depths.

data/ionization_potentials.dat:
  - New data file with accurate ionization potentials obtained from the Atomic
    line list and NIST.
source/atmdat.h:
source/atmdat_readin.cpp:
  - Add code to read the new data into atmdat.EIonPot[][].

source/species.cpp:
  - Fix compiler warning by g++ 7.2.0 when compiling with -DNDEBUG: writing into
    array that may be too small to hold the output.

add mix of lines to save line emissivity.  The 5199 
blend has zero emissivity, reported in ticket #415

Exercise command 'save lines, optical depths some' in tsuite

docs/latex/hazy2/lines.tex:
docs/latex/hazy2/observed.tex:
  - Fix incorrect references to the PRINT CONTINUUM command.

update citation for 2017 paper

bugfix for spherical intensity case.
Improve comment describing variable

tsuite/slow/pdr_xdr.in:
  - Fix botches in alldouble runs due to the fact that nzone has a bimodal
    distribution.

source/vectorhash.h:
source/vectorhash.cpp:
  - New very fast vectorized algorithm to obtain checksums of a large  buffer.
source/md5datafile.cpp:
  - Renamed to:
source/vh128sum.cpp:
  - Main program to obtain checksum. This has been deliberately set up such that
    it does not depend on Cloudy. This avoids bloating the binary and spurious
    rebuilds.
source/Makefile:
source/cpu.cpp:
source/cpu.h:
source/prt_comment.cpp:
  - Use the new algorithm where appropriate. The checksums in the grain opacity
    files and the SAVE TRANSMITTED CONTINUUM output are still done using the old
    MD5 algorithm for backward compatibility. The new algorithm is used to check
    if data files have been modified and so reduces the overhead from this check.
source/vectorize_math.h:
  - Enhance vpack class.
source/TestVectorize.cpp:
  - Test the new algorithm.
scripts/generate_md5sums.sh:
  - Renamed to:
scripts/generate_checksums.sh:
  - Use new checksum algorithm.
  - Optimize script for speed.
  - Sort output alphabetically to get stable sequence.
data/checksums.dat:
  - New file with checksums.
data/md5datafiles.dat:
  - Deleted.

Clarify units of the opacity() methods in the EmissionProxy classes

Use NIST energy for level 9 in Stout data for Mg10.  ADAS energy was used before.  The energy difference is about 0.02%.

lines_service.h
  - Remove SDIV() from wn2ang().
lines.cpp
  - Run wavelength through wn2ang() to account for refractive index.

this figure did not display correctly due to large white space around figure itself.

source/stars.cpp:
  - Bug-fix: Starburst99 files were parsed wrong.

Coding style changes suggested by Peter

Update He-like triplet labels and wavelengths, in the test suite, and the data/ line lists.

source/generic_state.cpp
  - In extractLevels(), switch to using atoi() from stod()
    for string conversions to integers.
    The use of stod() was opted for with designations such
    as '1e3' in mind, which, however, are not expected to be
    used much.
  - Check against zero value in the case where only one
    level is given (e.g., '[2]').  A zero value indicates
    atoi() failure, and is in any case an unacceptable choice.
source/TestGeneric.cpp
  - Update unit test.

data/compile_tsuite_stars.in:
tsuite/README_STARS.txt:
  - Remove the last input script from the data directory and archive its contents
    in the test suite directory if we ever need to recreate the SED tables that
    are used in the test suite.

small changes in bibtex entries

source/prt.cpp
  - Add checks in t_prt_matrix::prtRates() against the number of
    available levels, when printing the arrays that enter the solver.

docs/latex/common/bibliography2.bib:
docs/latex/hazy1/miscell.tex:
  - Document the addition of the Draine (2003) rfi files.

data/LineList_He_like.dat
tsuite/auto/blr_kk81.in
tsuite/auto/hii_coolstar.in
tsuite/auto/hii_paris.in
tsuite/auto/igm_primal.in
  - Update two-photon continuum wavelengths.

Delete unwanted input scripts in the data directory

data/compile_gauntff.in:
data/opacity.in:
data/compileopac.in:
  - Obsolete. These commands are no longer supported.
data/compileatlas.in:
data/compilecostar.in:
data/compilerauch.in:
data/compilerauchold.in:
data/compilewerner.in:
  - Compiling single SED grids as shown in these scripts never worked.
data/compile_grains.in:
  - Redundant with data/compilegrains.in.
data/crash.in:
  - This script cannot work; it is also redundant with tsuite/scrash.
data/test.in:
  - This is a likely script name users may choose, so the presence of this file
    may well lead to confusion. Its presence is not needed as its functionality
    can easily be obtained with the "cloudy.exe -e test" syntax.

docs/LineLabels.txt -- Update to remove induced two-photon continuum entries

missing option on command

Update Stout references, and BibTeX file.

source/prt_lines_helium.cpp
  - Adjust comments to distinguish between the singlet and triplet
    two photon continuua of He-like ions.
docs/LineLabels.txt
  - Update labels.

Update Stout data for Al+2.  Collision strengths come from ADAS.
First 2 excited level energies come from NIST.

Generalize t_mesh method to check if energy is within the continuum mesh bounds.

source/mesh.h
  - Generalize method to operate on Energy class, instead of on a realnum variable.
    Method renamed to 'isEnergyBound()' from 'isWNbound()'.
source/atmdat_chianti.cpp
source/atmdat_lamda.cpp
source/cont_createpointers.cpp
  - Update method calls.
docs/LineLabels.txt
  - Update labels.

Only treat transitions with energies within the continuum energy mesh limits.

source/mesh.h
  - Define new method, isWNbound(), to check that a given wavenumber
    falls within the bounds of the continuum energy mesh.
source/atmdat_chianti.cpp
source/atmdat_lamda.cpp
  - Compute wavelength for transitions with energies within the mesh
    bounds.
source/cont_createpointers.cpp
  - Only setup pointers for baseline transitions with energies
    within the mesh bounds.

iso_create.cpp
  - Change comments in output of 'save lines labels' for H-like ions
    to report the principal quantum number, except for n=1 and n=2,
    for which we use spectroscopic notation.

tsuite/auto/stars_rauch_pg1159_ascii.in:
  - New test case to detect regressions from the fix in r11966. This sim also
    implements a test for reading directly from an ascii file.

update email address

docs/latex/hazy1/conout.tex:
  - Document units of normalized Xspec spectrum.

Added human-readable BibTeX code for 2017 Revista paper

Update Stout data for "S 5".  Collision strengths come from ADAS.

data/LineList_BLR.dat
data/LineList_BLR_Fe2.dat
data/LineList_NLR.dat
source/prt_lines.cpp
  - Update "S 5" wavelengths.
data/chianti/masterlist/CloudyChianti.ini
data/stout/masterlist/Stout.ini
  - Update species lists to use Stout for "S 5" by default.
data/stout/s/s_5/s_5.coll
data/stout/s/s_5/s_5.nrg
data/stout/s/s_5/s_5.tp
  - Update atomic data to use ADAS.
    NIST energies are used for the first 19 excited levels.

adjust monitor for slow drift up

update Badnell RR, DR data to results of
https://www.aanda.org/component/article?access=doi&doi=10.1051/0004-6361/201731243

source/save_fits.cpp:
  - Bug-fix: the wrong array was being saved into the FITS file.

add missing line

fix botch from clang in radegund on blr_n14_p22 sim

as per PvH suggestion, do not recommend
alias to cloudyconfig.h, rather include sys_gcc on the include path