Context Navigation

Changes between Version 2 and Version 3 of Ticket #34564, comment 19

v2	v3
3	3	> I suppose the old convert_value function was there for those scenarios
4	4
5		It could have been there solely to support a subset of the cases that `empty_result_set_value` covers for aggregation and selects now. This code was introduced in f59fd15c4928caf3dfcbd50f6ab47be409a43b01 without ~~clear~~ a clear reason. We'd have to checkout the code from that time and run the suite to be sure.
	5	It could have been there solely to support a subset of the cases that `empty_result_set_value` covers for aggregation and selects now. This code was introduced in f59fd15c4928caf3dfcbd50f6ab47be409a43b01 without a clear reason. We'd have to checkout the code from that time and run the suite to be sure.
6	6
7	7	The fact you cannot reproduce against SQLite, Postgres, MySQL, or Oracle might explain why this was not caught by the test suite in the first place which would make more sense given how common such annotations are.