Vladimir Sitnikov
Vladimir Sitnikov
@Sanne , I believe it is only for historical reasons like in "it was under synchronized lock previously, so we blindly converted it to the lock so keep the semantics"....
@ncasaux , Yeah. If we fix the typo, we'll need to support both variants.
> I don't think the whitespace handling is necessary, folks could always construct their strings to not have whitespace There might be an application (or library) that generates literals with...
Frankly speaking, adding more options seems fragile. The naming of `allowSpecialNumeric` looks obscure as well. In theory, we could have option like `decodeNumericas=bigdecimal|number|...`, however, a global option is hard to...
> I believe dynamically adjusting the type when encountering special return values isn't a good approach. Yeah, we should not return different types from the same API. The clients would...
It looks like it would be fine if org.postgresql.jdbc.PgResultSetMetaData#getSchemaName delegates to org.postgresql.jdbc.PgResultSetMetaData#getBaseSchemaName. > And well as receiving a call getColumnName real column name and not the column alias That seems...
Should we rather infer the schema name like Lukas suggested in https://github.com/pgjdbc/pgjdbc/issues/538#issuecomment-1346193264? Sure think we should not try parsing SQL to infer the table names. However, I think we should...
The key question is the way to handle backward compatibility. We could move asserts to a different artifact, and use it as a regular dependency. Then, at some later version...
> What if "org.testng.assertions.Assert" is introduced and "org.testng.Assert" is deprecated in the next minor release? Sure it is possible to combine the changes. I just assumed it would take some...
> In TestNG 7.13.0 we remove org.testng.Assert from the testng uber jar and be done with this activity I would ask you not doing so. It takes time for the...