Rewrite more functions

[systay]

This PR adds support for SCHEMA() and FOUND_ROWS(). SQL_CALC_FOUND_ROWS is being implemented in a separate PR

[morgo]

There is a bug in SQL_CALC_FOUND_ROWS. Here is the test case:

DROP TABLE IF EXISTS t1;
CREATE TABLE t1 (id INT NOT NULL PRIMARY KEY auto_increment);
INSERT INTO t1 VALUES (NULL),(NULL),(NULL),(NULL),(NULL);

SELECT SQL_CALC_FOUND_ROWS * FROM t1 LIMIT 1;
SELECT FOUND_ROWS();

SELECT SQL_CALC_FOUND_ROWS * FROM t1 LIMIT 2;
SELECT FOUND_ROWS();

The first FOUND_ROWS() returns 1, second returns 2. Both would be correct if the SQL_CALC_FOUND_ROWS was not present. Because it is, both calls are expected to return 5.

[systay]

There is a bug in SQL_CALC_FOUND_ROWS.

Well, I've only added parsing support so far. I have yet to add the bugs to the planner. ;)

[sougou]

I don't see the tabletserver part where we have to turn around, request found_rows() and return it to vtgate.
That still needs to be implemented, right?

Everything else looks good.

[harshit-gangal]

do you not need locking around updating the foundRows?

[harshit-gangal]

nit: add comment for being in the plan like others. Also need to add json:"-" if this does not need json representation.

[harshit-gangal]

LGTM

[gitmko0]

I'm having problem with SQL_CALC_FOUND_ROWS not being Rewrite as well. Is this resolved?
When will it be resolved?

SELECT SQL_CALC_FOUND_ROWS t_posts.ID FROM t_posts WHERE 1=1 AND t_posts.post_type = 'post' AND (t_posts.post_status = 'publish') ORDER BY t_posts.post_date DESC LIMIT 0, 10

New Edit : i found out this was not updated in the vitess.io repo.
this github one is working now. sorry for this comment.