Merged PRs
dolt
go-mysql-server
- 3034: Allow
drop trigger ...when trigger is invalid
We should ignore parser errors when trying to drop triggers that are invalid.
fixes: #9359 - 3032: Generalize types in
CASE,IF, andIFNULL
Fixes #9321
Generalize types inCASE,IF, andIFNULLas described by MySQL docs - 3031: Fix 7998 json escape double quotes
Key write now uses same recursive call towriteMarshalledValue()as value underjson_encode.go
Fixes: #7998 - 3030: Distinguish nils in firstBuffer
Fixes #9035
InAggregationBuffer.firstBuffer, the buffer should not update if the first row has already been written. However, there was no way of distinguishing between anilbecause the buffer is empty and anilbecause the value in that column of the first row isnil.
When you have two rows(NULL, 1)and(1, NULL), the first row gets written into[]firstBuffer.valas{nil, 1}. Because there's no way to tell whether the first value is supposed to benil, it ends up getting overwritten by the1in the second row. Because the second value is notnil,Updateskips it. As a result, we end up with a row{1, 1}that matches neither of the original rows and evaluates toTRUEwhen grouped byc0 = c1, even though the original rows evaluate toNULL.
This isn't an issue when the rows are(NULL, 1), (NULL, 1)or(1, NULL), (1, NULL)because thenilvalue remains anilvalue whenUpdateis called on the second row.
I fixed this by adding awrittenNilflag tofirstBufferto indicate that thenilvalue is meant to be there and should not be overwritten. - 3029: Fix 9347 panic case statements info schema
information_schemaupdated to respect enum type
Eval()underenum.gois more flexible with strings now
Fixes: #9347 - 3028: add mysql dialect tag to new
group bytest
newgroup bytest added in dolthub/go-mysql-server#3025 is not Postgres/Doltgres compatible - 3024: Adding skipped tests for
UPDATE ... JOINbugs - 2994: fix NewServer call
this is a followup to dolthub/go-mysql-server#2989