Migrate calls to String::trim() to Guava's whitespace trimming

[wetneb]

Java 11 introduced the String::strip() method, which strips whitespace around a string, including newer Unicode whitespace characters.
We currently use the String::trim() method in a number of places. In many of those places, the string being trimmed can legitimately contain any Unicode character, so String::strip() should be used instead.

Since OpenRefine now requires Java 11+, we can migrate to String::strip() in most places. Note that String::trim() apparently removes the \u0000 character and String::strip() does not, but I am not sure if that is relevant anywhere for us.

We should rather migrate to Guava's whitespace trimming, see @ostephens's post below.

[tfmorris]

It's probably worth replacing the Guava com.google.common.base.CharMatcher calls used for whitespace stripping as well to minimize external dependencies.

Actually, that's a bad idea because Java's definition of "whitespace" doesn't match Unicode's

[sritejap]

@elroykanye Are you working on this?

[elroykanye]

Hi @sritejap , I have not had enough time to look into this. If you are interested in the issue, I can assign this to you.

[vijal-patel]

I could look into this if no one is working on it right now

[vijal-patel]

It's probably worth replacing the Guava com.google.common.base.CharMatcher calls used for whitespace stripping as well to minimize external dependencies.

wrt this, I've found 1 usage of that class for white space stripping (link)
but chaning this to use .strip() results in unit test failures for non-breaking space, figure space etc.

Shall I leave this one as is?

[ostephens]

So we have different options for stripping whitespace/space characters from a string:

• String::trim() -> strips any character whose codepoint is less than or equal to 'U+0020' (the space character).
• String::strip() -> strips any characters for which Character::IsWhitspace() == TRUE
• com.google.common.base.CharMatcher.whitespace().trimFrom() -> strips any characters defined by Unicode as whitespace

Character::IsWhitespace() specifically excludes non-breaking spaces ('\u00A0', '\u2007', '\u202F') which is why the tests are failing when you change the method there.

Since the use of the Guava method was introduced explicitly so that non-breaking spaces were included in the GREL trim command (see f29f77e) it seems like we should preserve this behaviour.

I would flag that trimming of non-breaking spaces is also being asked for explicitly by #5246 - but in a place we currently use String::strip() ... so I wonder if we should rather move to the use of the Guava method at least in that scenario for consistency and to meet this user need?

[wetneb]

Thanks @ostephens for investigating this - I have not looked much in the details but I trust your judgment and update the issue to reflect that.

[tfmorris]

I agree with @ostephens analysis. I had forgotten that Java's definition of whitespace doesn't match the Unicode Consortium's definition, making strip() unsuitable for OpenRefine's use.

https://docs.oracle.com/en/java/javase/11/docs/api/java.base/java/lang/String.html#trim()
https://docs.oracle.com/en/java/javase/11/docs/api/java.base/java/lang/String.html#strip()
https://docs.oracle.com/en/java/javase/11/docs/api/java.base/java/lang/Character.html#isWhitespace(int)