| Arnold D. Robbins | 5a18f63 | 2020-01-22 02:10:59 -0700 | [diff] [blame] | 1 | Wed Jan 22 02:10:35 MST 2020 |
| Arnold D. Robbins | df6ccd2 | 2020-01-17 14:08:59 +0200 | [diff] [blame] | 2 | ============================ |
| 3 | |
| 4 | Here are some things that it'd be nice to have volunteer |
| 5 | help on. |
| 6 | |
| 7 | 1. Rework the test suite so that it's easier to maintain |
| 8 | and see exactly which tests fail: |
| 9 | A. Extract beebe.tar into separate file and update scripts |
| 10 | B. Split apart multiple tests into separate tests with input |
| 11 | and "ok" files for comparisons. |
| 12 | |
| 13 | 2. Pull in more of the tests from gawk that only test standard features. |
| 14 | The beebe.tar file appears to be from sometime in the 1990s. |
| 15 | |
| 16 | 3. Make the One True Awk valgrind clean. In particular add a |
| 17 | a test suite target that runs valgrind on all the tests and |
| 18 | reports if there are any definite losses or any invalid reads |
| 19 | or writes (similar to gawk's test of this nature). |