[77778f8] | 1 | This is a changelog for the dev branch of FLEXPART (see flexpart.eu) |
---|
| 2 | Created by Petra Seibert, 8 June 2018 to document and explain in detail why I changed what. |
---|
| 3 | |
---|
[1a8fbee] | 4 | * makefile, 2018-06-08 |
---|
| 5 | ======== |
---|
[77778f8] | 6 | |
---|
| 7 | 1) Add a GPL3+ License statement, author statement, version information. Prudent to do that, not harmful. |
---|
| 8 | |
---|
| 9 | 2) There are already two sections, one for gfortran4 and one for gfortran5. |
---|
| 10 | Both have NILU-specific paths. So let us have at least one generic and put that in front. |
---|
| 11 | |
---|
| 12 | 2) Compiler options. |
---|
| 13 | 2.1) For the regular compile, we don't want -g (debug symols) |
---|
| 14 | 2.2) We don't need -m64, on a 64bit platform this is default |
---|
| 15 | 2.3) Remove $(USER) as it is never defined |
---|
| 16 | 2.4) Remove -Warray-bounds, this is default. |
---|
| 17 | 2.5) Add -fbacktrace in the commented-out part, it is important for debugging |
---|
| 18 | 2.6) Change comment for VPATH, explain what this directive does |
---|
| 19 | 2.7) Added some explanation of options as comment. Maybe expand later. |
---|
| 20 | |
---|
| 21 | Notes and questions: |
---|
| 22 | 1. Would be nice to have a debug version also for serial code. |
---|
| 23 | 2. Would be nice to have "as fast as possible" option (thought it might be machine-dependent ...) |
---|
| 24 | 3. What is the purpose of -Warray-bounds ? Is it relevant for Fortran? I think if there is a compile-time array-bound violation this would be a hard error? |
---|
| 25 | |
---|
[1a8fbee] | 26 | * FLEXPART.f90 |
---|
| 27 | ======== |
---|
[77778f8] | 28 | |
---|
| 29 | Update version string! |
---|
| 30 | |
---|
[1a8fbee] | 31 | * readcommand.f90 |
---|
| 32 | =========== |
---|
[77778f8] | 33 | |
---|
| 34 | Correct misleading error message (replace "open" by "write to") |
---|
| 35 | Improve error messages: combine pathname+filename, remove unaligned closing ### |
---|
| 36 | Let STOP say in which subroutine we are stopping |
---|
| 37 | Some minor changes |
---|
[1a8fbee] | 38 | |
---|
| 39 | * verttransform_ecmwf.f90 PS 2018-06-13 |
---|
| 40 | =================== |
---|
| 41 | |
---|
| 42 | 1) Remove some of commented out testing stuff |
---|
| 43 | 2) Fix indenting in the init if block |
---|
| 44 | 3) Code layout improvement |
---|
| 45 | 4) Change name of loops to represent the index |
---|
| 46 | 5) Fix ticket:140 Introduce a new way of establishing the reference position |
---|
| 47 | for the vertical grid. Also correct a minor bug in the calculation |
---|
| 48 | of height (array assignment where it was not intended) |
---|
| 49 | 6) Add back the SAVE attribute to INIT, just to be sure |
---|
| 50 | 7) Bring some more code under the IF (INIT) block |
---|
| 51 | 8) make some annotations |
---|
[f251e57] | 52 | [minor modification PS 2018-06-25] |
---|
[1a8fbee] | 53 | |
---|
| 54 | * com_mod.f90 PS 2018-06-13 |
---|
| 55 | ======= |
---|
| 56 | |
---|
| 57 | correct the comment for nmixz |
---|
| 58 | |
---|
| 59 | * verttransform_nest.f90 PS 2018-06-17 |
---|
| 60 | =================== |
---|
| 61 | 1) insert proper boundaries for implied loops in array expressions |
---|
| 62 | (fixes bug annoted by ESO in 2016) |
---|
| 63 | 2) Code layout improvement |
---|
| 64 | 3) Change name of loops to represent the index |
---|
| 65 | 4) make some annotations |
---|
[f251e57] | 66 | |
---|
| 67 | * verttransform_ecmwf.f90 PS 2018-06-25 |
---|
| 68 | =================== |
---|
| 69 | |
---|
| 70 | 1) Some code layout improvement in the first part |
---|
| 71 | 2) Fix ticket:140 Introduce a new way of establishing the reference position |
---|
| 72 | for the vertical grid. Also correct a minor bug in the calculation |
---|
| 73 | of height (array assignment where it was not intended) |
---|
| 74 | 3) Add back the SAVE attribute to INIT, just to be sure |
---|