rP26522 nuked passing a reference to some functions were it actually isn't used and more was set to a value before calling. In AssembleCalls the value was initialized to 0.f instead of widthRangeFrom. This causes #6451, by having a mismatch between the width checks (we miss a bufferZone in AssembleCalls)
This explains s0600204 second comment in the ticket. Since a longer string will be so long that it will be large enough to also completely fill the bufferzone.
Introduced in rP26522