pureboot is at git.blackmark.me/avr/pureboot now, with its own history: the
files it kept in pureboot/ sit at the top level there, its CMakeLists is the
merged whole of that unit and the build around it, and the v1..v8 tags moved
with it - each one checks out and compiles to exactly the .text it compiled to
here. The loader that repo builds is byte-identical to the one this commit
removes, verified before the removal rather than after.
Nothing is rewritten on this side. The history and the tags are untouched, so
every commit before this one still has pureboot in it and still builds it;
this is one commit that stops carrying it forward.
What goes with it: the four pureboot files, the thirteen pb*.py protocol
drivers and their four host-side unit tests, pbapp and the pureboot simavr
runner, check_pi.py, pbhw.py, pbrig.py, sizes.py, and the Studio project.
check_unit.cmake goes too - it had no caller left once the unit tests moved.
What is left is the four TinySafeBoot tiers, and the build shrinks to fit
them: 757 lines of CMakeLists to 137, and the preset matrix from 37 chips to
one, because the tiers reimplement an ATmega328P-only protocol and every other
chip in that list was there for pureboot. check.sh loses its size-table pass -
the table it checked was pureboot's README - and the Studio solution loses the
project that is now in the other repo.
Gate green: nine tests, four tiers at their section sizes and each one's
protocol suite against the simulator.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
master carries bootloader.atsln, so main does too. Two projects, because a
.cppproj is one binary at one flag set and this build has hundreds: the stock
328P pureboot loader (USART0 at 115200 on a 16 MHz crystal), and the tsb_asm
tier that occupies the same 512-byte section master's own tsb project targeted.
Both come out byte-identical to the Ninja build — 404 B and 510 B of .text —
in both configurations.
Debug keeps -Os and adds only -gdwarf-4. A loader's section is a correctness
bound, and -Og builds this source to 590 B: the link at 0x7e00 accepts that
without a diagnostic, 78 bytes past flash end, where rcall/rjmp wrap modulo
flash size and the image dies just after activation. Debug info costs no flash,
so the optimisation level stays where correctness needs it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
