SeedChangeRequests: name the sibling seeded requests the way the code spells them

[dfe-lance]

Two comments in SeedChangeRequests.cs pointed at a seeded request as SEED006 and
SEED008, which is how the references were written before they were generated:

(Reference: Reference(7), Status: RequestStatus.ReadyToSubmit, Pupil: p[5]), // duplicate of SEED006
(Reference: Reference(9), Status: RequestStatus.ReadyToSubmit, Pupil: p[6]), // duplicate of already-submitted SEED008

The comments were accurate — Reference(n) produces exactly CYPMD_KS4June_SEED{n:000}, so
Reference(7) is CYPMD_KS4June_SEED007 and the seeded data is unchanged. But a reader had to
map between two names for the same thing to see which row was being pointed at, and the obvious
reading was that the comment referred to something the code no longer creates.

They now name the sibling row the way the line beside them does.

Comments only. No behaviour change, and nothing about what is seeded changed.

Testing

• Seeded rows read back from the database after a rebuild: still CYPMD_KS4June_SEED001 to
  011, same statuses, same pupils
• Unit 5289 passed, integration 809 passed, full E2E 206 passed
• Full Release build of the solution clean, no warnings on changed lines

Closes #457