Junio C Hamano [off-list ref] writes:
... How hard it would be to construct a test
case that fails if filter computed as v1 is marketed as v2?
There may be a more effective way for future-proofing, besides
ensuring the test coverage.
Although this series added a way to see which version the on-disk
data is using using the version field, I do not think it touched the
"struct bloom_filter" and "struct bloom_key" that represent the
in-core data. If we had a member in these structures that
get_or_compute_bloom_filter() can fill in from the on-disk structure
or the version it used when it computed the filter anew, would it
become easier to catch the case where we try to add a version 2
computed key to a filter that was read from version 1 on-disk
structure, presumably at add_key_to_filter()?
Thanks.