| 1 | | It would be very useful to be able to compare results from previous runs as I could before ... |
| | 1 | This feature writes results for each regression test run into a directory named {{{regress/results/<timestamp>}}} and creates a symlink to it called {{{regress/results/latest}}}, and the previous {{{latest/}}} is renamed to {{{last/}}}. |
| | 2 | |
| | 3 | The results directory holds, for each test: |
| | 4 | * whatever auxiliary output files LHC writes (depending on the flags), including .grin, .c, and executable files |
| | 5 | * a .lhc_out file into which lhc's stdout is redirected |
| | 6 | * .stdout and .stderr files of execution output |
| | 7 | |
| | 8 | Optionally, the .ho files for the test suite are stored in the {{{ho}}} subdirectory of the results directory. (The other choice is that they are stored in {{{results/ho}}}.) |
| | 9 | |
| | 10 | There is also top-level log.txt file containing a copy of the output of regress.pl: basically, whether each test passed or how it failed, it's compile time, and it's run time. (Haven't seen it lately, though.) |