db: genericize rangeDel{Cache,Frags} into keySpan{Cache,Frags}

[jbowens]

Rename the memtable's rangeDelCache and rangeDelFrags types to
keySpanCache and keySpanFrags so they may be used for caching range
key fragments too.

In addition to the pure renames, I made a small refactor to decouple the
dependency on memTable.tombstones by adding a pointer to the skiplist
in a keySpanCache field.

[cockroach-teamcity]

This change is 

[sumeerbhola]

Reviewed 1 of 1 files at r1, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @jbowens)

---

mem_table.go, line 371 at r1 (raw file):

		return nil
	}
	return frags.get(c.skl, c.cmp, c.formatKey)

Can't the code in this method extract these from m like the callee used to do, and then call frags.get. I am trying to understand why we need the member variables in keySpanCache.

Or is this just the first step in the refactor and get will eventually not take a memTable parameter. Should it be removed now itself, since it is unused?

[jbowens]

Reviewable status: 0 of 1 files reviewed, 1 unresolved discussion (waiting on @sumeerbhola)

---

mem_table.go, line 371 at r1 (raw file):

Previously, sumeerbhola wrote…

Can't the code in this method extract these from m like the callee used to do, and then call frags.get. I am trying to understand why we need the member variables in keySpanCache.

Or is this just the first step in the refactor and get will eventually not take a memTable parameter. Should it be removed now itself, since it is unused?

Oops, yeah, I forgot to remove the memTable parameter when extracting this change from another branch. Thanks :)

[sumeerbhola]

Reviewed 1 of 1 files at r2, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @sumeerbhola)

[jbowens]

TFTR!

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @sumeerbhola)