But because only part of the tests (unbeknownst to me) get run in CI, I never realized that some tests broke as a result. This PR should fix things, and I will accompany it by a change to Cactus to make sure all HAL tests get run in CI in the future.
In https://github.com/ComparativeGenomicsToolkit/hal/commit/f7d171a09ac10b8835a520b24ae04ef20bc89000, I changed the default behaviour of
hal2mafto not leave big gaps along the reference (first row) in the presence of paralogies.But because only part of the tests (unbeknownst to me) get run in CI, I never realized that some tests broke as a result. This PR should fix things, and I will accompany it by a change to Cactus to make sure all HAL tests get run in CI in the future.