feat: use exists for relational filters

[jacek-prisma]

ORM-395
Closes prisma/prisma#25731

This PR changes all the relational filters to use correlated queries with a WHERE clause referring to a table from the outer query instead of IN, e.g:

SELECT COUNT(*)
FROM (
  SELECT "public"."Issue"."id"
  FROM "public"."Issue"
  WHERE ("public"."Issue"."id") NOT IN (
    SELECT "t1"."A"
	FROM "public"."_IssueToUser" AS "t1"
	INNER JOIN "public"."User" AS "j1"
	ON ("j1"."id") = ("t1"."B")
	WHERE (1=1 AND "t1"."A" IS NOT NULL)
  )
) AS "sub"

becomes

SELECT COUNT(*)
FROM (
  SELECT "public"."Issue"."id"
  FROM "public"."Issue"
  WHERE NOT EXISTS (
    SELECT "t1"."A"
	FROM "public"."_IssueToUser" AS "t1"
	INNER JOIN "public"."User" AS "j1"
	ON ("j1"."id") = ("t1"."B")
	WHERE (1=1 AND "t1"."A" IS NOT NULL) AND ("public"."Issue"."id" = "t1"."A")
  )
) AS "sub"

Notes:

• I've had to change the alias generation code a little bit because the counters weren't being reused properly which led to name clashes with the correlated query. I've moved the counter into Context to make sure we always use the same counter.
• In postgres16 specifically the results seem to be ordered inconsistently for some of the EXISTS filters, I don't think that's a problem, but it did break some tests. I updated these tests with orderBy.
• The code now outputs correlated queries for relational filters, which should generally be very efficient because relational filters always query for identifier columns, but there's no guarantee that this is always going to be faster. It might be worth investigating the performance impact in different queries and databases.

[codspeed-hq]

CodSpeed Performance Report

Merging #5104 will not alter performance

_{Comparing feat/relation-exists (3d318b6) with main (4123509)}

Summary

✅ 11 untouched benchmarks

[github-actions]

WASM Query Engine file Size

Engine	This PR	Base branch	Diff
Postgres	2.097MiB	2.096MiB	885.000B
Postgres (gzip)	841.250KiB	841.363KiB	-116.000B
Mysql	2.058MiB	2.057MiB	1010.000B
Mysql (gzip)	827.479KiB	827.871KiB	-401.000B
Sqlite	1.972MiB	1.972MiB	-240.000B
Sqlite (gzip)	791.929KiB	792.450KiB	-533.000B