[SQLLINE-228] Set "incremental" to false by default

[liancheng]

This PR tries to address #228 by setting the default value of option incremental to false and add an extra help entry for --incremental.

[liancheng]

This PR didn't touch the docbook manual because it already declares the default value of incremental to false.

[liancheng]

@julianhyde Friendly ping :)

[snuyanzin]

From my side it looks ok. However I have a question.
@julianhyde, @liancheng do you think it would be useful to have a possibility to specify properties file to use while initialization instead of default one?
In this case it could allow to specify there incremental=false property and no other code changes in test will be required. Also it could be used as a set of default settings for tests.
I'm not suggesting do it here, just asking whether you think it would be a good refactor.

[julianhyde]

It is a shame that there are so many changes to tests, but I think @liancheng did the right thing in preserving the current behavior of tests as much as possible.

@snuyanzin I don't think it's a good idea to have a separate properties file for init. While it would solve this problem, it would confuse users, because it's not a very useful feature for users.

I think that non-incremental mode is more important than incremental mode, so in future, I think tests should be written to in the new default, non-incremental mode. (And perhaps some of the tests that were just explicitly made incremental could be converted to non-incremental.)

Why is non-incremental mode better? Because it produces smarter column widths, and uses screen real estate more efficiently. A better user experience.

I do worry a little bit that in non-incremental mode, statements that return many thousands of rows will use a lot of memory. So, how about making such statements switch to incremental mode after N rows (say N = 1000)? (Not as part of this change; perhaps a feature request for the future?)

[snuyanzin]

Thank you for your comment. I think you are right

about

I do worry a little bit that in non-incremental mode, statements that return many thousands of rows will use a lot of memory. So, how about making such statements switch to incremental mode after N rows (say N = 1000)? (Not as part of this change; perhaps a feature request for the future?)

If am not mistaking there is a very similar feature with configured --incrementalBufferRows=NUMROWS and default 1000 in beeline https://issues.apache.org/jira/browse/HIVE-14170 which could be ported to sqlline

[julianhyde]

Thanks; I have logged #251.

[julianhyde]

Merged as 30f8ddb, fixes #228.