Activity for Alan W. Irwin
-
Alan W. Irwin committed [b79648]
Remove trailing whitespace for two new documentation files
-
Use doxygen to document the modern and legacy variants of the free_eos generic API
-
Update build, install, test, and user application link instructions
-
Move the utils subdirectory from the top-level directory to test_FreeEOS
-
Additional changes that I forgot to make part of the prior commit
-
Fix error in the formula for the radii of the Rydberg levels of H_2^+
-
Drop unnecessary "save pf" in free_eos_detailed
-
Improve commentary and change a variable name in mdh_pi.f90
-
Update mod_fermi_dirac_integrator.f90 to latest version of best
-
Implement all "best-practice changes related to the dot_product intrinsic and the logqtl_const array
-
Calculate the square of the sound speed and return it as a modern API argument
-
Implement best practices (see README.best_practices) for fjs_pi.f90
-
Comment out ifnr.eq.0 stanzas from eos_warm_step
-
Implement comprehensive special-grid tests for the GEFF and SIREFF option suites
-
Fix all but one issue found for the EOS2 option suite
-
Change where free_eos_test checks for non-zero info results from free_eos
-
Force fl to remain within the range from ln_underflow_limit to ln_overflow limit for exchange and Fermi-Dirac integral calculations
-
Fix long-standing ifsame_under bug in the maximum neutral and ionized fractions calculation in ionize
-
Move summary of best practices for Fortran 2008 from README.release to the new file README.best_practices
-
Replace equivalenced aux_scale with sum0_scale, sum2_scale, extrasum_scale, and xextrasum_scale
-
Convert the free_eos_test utility to use intrinsic iso_fortran_env module standard unit numbers
-
Optionally taint the free_eos_detailed arrays that are equivalenced to a variety of variables and arrays.
-
Implement generic free_eos subroutine
-
Follow up previous "tainting" commit with additional tests
-
Implement tainting method of detecting uninitialized elements of allocated real and integer arrays and fix the uninitialized real issues that are found
-
Implement latest version of best practices in the mod_free_eos module
-
Update the mod_excitation module to the latest version of best practices
-
Update the remaining subset (all but the fermi_dirac_ct subset that was done before) of the mod_fermi_dirac module to the latest version of best practices
-
Update test_FreeEOS/free_eos_test.stdout_standard to be consistent with the rest of the results from 2021-04-04
-
Update the fermi_dirac_ct subset of the mod_fermi_dirac module to the latest version of best practices
-
Update the mod_pi_fit module to lastest version of best practices
-
Change traditional "g", "h", "hd", and "he" internal core variable names to something less problematic
-
Add denatured valgrind tests to the comprehensive_test target
-
Fix "error stop" to "stop" regression recently introduced by commit 876889ef
-
Rename the existing mod_qmhd_calc module as mod_excitation and move excitation_sum (from mod_eos_calc) and excitation_pi, excitation_pi_pressure_free, and excitation_pi_end (from mod_pi) to that renamed module
-
Update all module routines in the mod_eos_calc module (and also excitation_pi.f90 and mod_statistical_weight_data.f90) to the latest version (see README.release) of best practices
-
Update the release notes
-
Replace all traditional forms of auxiliary variables in eos_warm_step arguments with aux forms
-
Fix plot title bugs in compare_free_eos_test.py(.in)
-
Remove some obvious PTEH pressure-ionization significance loss issues in the entropy and entropy derivative calculations
-
eos_calc: unpeel part of the remaining overflow onion
-
Eliminate Python run-time errors from the free_eos_model.py script run by the exchange_paper target
-
Rationalize the non-zero return codes returned by the free_eos library
-
Flush stderr for error stops preceded by important stderr output
-
Fix regression introduced by recent "Reduce floating-point overflow issues" commit
-
Improve error reporting for solve_linear_svd.f90
-
eos_calc: Propagate a positive return code whenever the calculated electron number density is too close to underflowing
-
Precede every "stop 'Normal termination'" Fortran statement by "flush(0)" to work around a gfortran stderr ordering issue
-
eos_bfgs and eos_calc: unpeel part of the remaining overflow onion
-
Address the remaining unused parameter issues discovered by the gfortran -Wunused-parameter option
-
On 2021-02-27 13:31-0000 oldcodebit wrote: [...] I could not find examples on such things like running a plot without a menu with a list of output options, although that menu is very useful for the beginners to test - trial display options etc. But there is no example for running a plot say x11 or cairo display mode without popping up menu. Other useful tips would be helpful such as closing plplot window and running plot on the same windows ( live plot or run time plot if possible). You have probably...
-
Reduce floating-point overflow issues
-
Update src/README.developers
-
Finish improvement of stop statements
-
Update the configured compare_free_eos_test.sh and compare_free_eos_test.py scripts
-
Substantially improve compare_free_eos_test.py(.in)
-
Convert generic stop statements to specific stop statements
-
Improve use of stop statements
-
Include free_eos_interpolated.f90 in mod_free_eos
-
Fix issues for several different document_FreeEOS/figure_utils utilities
-
free_eos_compare: Add check for bad free_eos status codes
-
Modernize form of parameter statements in src/mod_free_eos_constants.f90
-
Fix initialization bugs for pteh_pi-related routines
-
Sorry. Now that I have read your message more carefully, I suspect you simply have not built or installed the shared version of the Python libraries that you need. So to fix this current issue, I recommend that you reconfigure your Python build so that the *.so symlink and corresponding Python shared library are installed and follow up by checking that the PLplot configuration actually finds that shared Python library that you have built.
-
On 2021-01-29 00:25-0000 Walt Kailey wrote: [support-requests:#52] Plplot 15.4.0 build fails on OpenSuse 15.1 Linux with ld error on libpython3.8.a libpython3.8.a is a static library that clearly will not be compiled with the correct -fPIC flags required by the shared library version of PLplot. The linker preferentially searches for shared libraries named .so (i.e., libpython3.8.so in your case) and only falls back to the static .a form if it cannot find the *.so form. Therefore, it appears your...
-
Update modern (2020) copyright information to 2021
-
Update to modern Fortran best practices (as documented in README.release) for all module routines in the mod_free_eos_detailed module
-
Fix initialization issues for eos_calc
-
Fix an additional tranche of initialization bugs
-
Fix long-standing screwup in indexing range of the extrasum-related arguments of excitation_sum
-
More improvements to Python script for plotting (with PLplot) relative differences between old and new free_eos_test results
-
Fix a tranche of initialization bugs
-
Specify ionization data with parameter approach rather than data statement approach
-
More improvements to Python script for plotting (with PLplot) relative differences between old and new free_eos_test results
-
Improve Python script for plotting (with PLplot) relative differences between old and new free_eos_test results
-
Update utils source code to latest best practices as documented in README.release
-
Update to modern Fortran best practices (as documented in README.release) for all module routines in the mod_bfgs module
-
Provide a preliminary version of the release notes for the forthcoming 2.3.0 release of FreeEOS
-
Move to latest version of best practices in the mod_coulomb module
-
To quiet valgrind reports of memory management errors, fix inconsequential uninitalized ne_old and lambda_old issues in free_eos_detailed
-
Follow argument declaration best practices for all module routines in the mod_fermi_dirac module
-
Follow argument declaration best practices for all module routines in the mod_qmhd_calc module
-
Follow argument declaration best practices for all module routines in the mod_pi_fit module
-
Remove trailing whitespace regression that was generated by earlier commit
-
Follow argument declaration best practices for all module routines in the mod_eos_calc module
-
Follow argument declaration best practices for all module routines in the mod_coulomb module
-
Follow argument declaration and save best practices and replace entries by subroutines for all module routines in the mod_poly_sum module
-
Remove trailing blanks in comments introduced by the 2020-10-23 commit 3dfe35b4
-
Follow argument declaration and save best practices for all module routines in the mod_eos_jacobian module and propagate extrasum argument best practices to all routines
-
Work around an uninitialized regression that occurs for a particular corner case
-
Follow argument declaration and save best practices and replace entries by subroutines for all module routines in the mod_pl module
-
Follow argument declaration and save best practices in the single (deriv) module routine in the mod_free_eos_deriv module that appears in free_eos_test_main.f90 and fix master_coulomb initialization bugs for ifcoulomb == 0
-
Fix several different array and string "invalid index" issues discovered by the gfortran --check=bounds option
-
Follow argument declaration and save best practices and replace entries by subroutines for all module routines in the mod_molecular_hydrogen module
-
Factor the mod_free_eos module into the mod_free_eos, mod_free_eos_detailed, mod_eos_jacobian, and mod_eos_bfgs modules
-
Implement python script and bash wrapper for that script that plots changes between comprehensive_test results
-
Follow argument declaration and save best practices and replace entries by subroutines for all module routines in the mod_abund_process module
-
Follow argument declaration and save best practices and replace entries by subroutines for all module routines in the mod_pi module other than fjs_pi and associated entries
-
Follow argument declaration and save best practices and replace entries by subroutines for all module routines in the mod_coulomb module
-
Implement valgrind_test target and fix memory leaks which that target initially showed