d3dd17d9

By: ValentinHirschi <valentin.hirschi@gmail.com>

stream lower-sector CFF component products

Feed the existing ordered component-product expansion lazily into its final
emitter instead of retaining each complete Cartesian intermediate layer with
duplicated momentum maps. Product order, coefficients, source ownership and
numerical assertions are unchanged. The final output and one-component fanout
can still be large; the GL16 memory benefit remains to be measured.

Synchronize the architecture notes with the already implemented unconditional
nonzero M parameter contract and label earlier validation as historical.

Independent ordering/ownership reviews, strict release checking, dev-optim
checking, Clippy and formatting pass. The existing focused build is running;
fresh GL16/full scalar and literal full-suite validation remain pending.
The prior GL16 attempt stopped at the actual 30 GB guard without a PASS.
This is a draft validation candidate, not the ALL_GREEN milestone.

User prompt:
Yes, continue as planned, it seems like it should be ok memory wise, so just let it run until LU scalar matrix is complete, and then really finalise the ALL_GREEN PR as mentioned earlier.

Related user instruction:
Ok continue as planned now, trying to finish this damn scalar LU matrix, and only looking for improvement to the code if you can find relatively non-invasive cheap wins. The core objective is getting to ALL_GREEN commit and push merge-ready state of the corresponding PR. Then stop yourself and deem the goal achieved (if there does not seem to be any gross oversight w.r.t what is in it).

Timeline