#428 auto complete list

[manticore-projects]

Fixes issues with the Auto Completion when quotes schemas or table names are involved.
Added some more test cases for quoted schemas and quoted table names.

I have not been able to fix the Web Client inserting the selection into the SQL (which is broken for quoted items).

Partially fixes #428

[katzyn]

schema.quotedName can be either quoted or unquoted, you shouldn't try to quote it again.

It is actually a normalized name (how it should be inserted into SQL), so you can rename this field and introduce a new one with a real always quoted name (constructed with StringUtils.quoteIdentifier()).

[manticore-projects]

Thank you very much for feedback and advice.
I have fixed that and also added tests accordingly for identifiers like "quoted schema" and "quoted tablename".

[katzyn]

I think these historic conversions to upper case don't look right, especially for some Unicode characters. String.regionMatches(false, …) should be used instead, some helper method in StringUtils can be introduced (startsWithIgnoreCase(String, String) or something like it).

[manticore-projects]

I think these historic conversions to upper case don't look right, especially for some Unicode characters. String.regionMatches(false, …) should be used instead, some helper method in StringUtils can be introduced (startsWithIgnoreCase(String, String) or something like it).

Agreed, I just wanted to avoid changing too much.
I will introduce a StringUtils.startsWithIgnoringCase(...), thanks for the recommendation.

[manticore-projects]

I think these historic conversions to upper case don't look right, especially for some Unicode characters. String.regionMatches(false, …) should be used instead, some helper method in StringUtils can be introduced (startsWithIgnoreCase(String, String) or something like it).

Remove all toUpperCase() calls and implemented a StringUtils.startsWithIgnoringCase() dealing with all Unicode shenanigans.

[manticore-projects]

Totally unrelated, please allow me to ask: why are those test cases implemented in a special way but not using regular Junit5? Of course it does not matter in general although I experiences some disadvantages:

• no Parametrized test (e.g. as per statement)
• I can't call a test method directly, but have to fiddle with the main(String[] args) method
• IDE can't benefit from Coverage and Test reporting

No need to rock the boat, just want to understand the rational behind.

[katzyn]

1. Both Java and JUnit were significantly less advanced when these tests were written.
2. H2 is mostly tested with integration tests and unit-test frameworks aren't fully suitable for them.
3. Our tests are executed multiple times in different configurations (in-memory, networked, etc). It wasn't simple to achieve this with JUnit, but maybe its modern versions are better, I don't know.

[manticore-projects]

I have fixed the String Comparison and am using Collator now. Sorry for the delay.
Please reconsider merging it.

PS: please why do the Unit Tests fail? I don't think this was related to my PR.

[manticore-projects]

Team, any objections on merging this?

[manticore-projects]

Thanks a lot!