Removed the last two prebuilt archives from the source tree

[yf13]

A few library files exist in the source tree now, they look like parts of the toolchain thus should be dropped?

$ find . -iname '*.a'
./ports_module/cortex_m0+/gnu/example_build/sample_threadx_module_manager/libgcc.a
./ports_module/cortex_m3/iar/example_build/stm32f2xx_library.a
./ports/arm11/gnu/example_build/libnosys.a

---

Maintainer note, added when retargeting this to dev.

Confirmed, and the repository's own rules already agree: .gitignore:23 is *.a, and it matches both remaining archives. They survive only because ignore rules do not untrack files that were committed before the rule existed. These are the last two .a files in the tree.

Neither has a consumer. sample_threadx_module_manager/ has no build script at all — just sources, a linker script and an Arm DS launch configuration — and nothing in that port mentions libgcc. stm32f2xx_library.a is referenced nowhere in the repository, and being a prebuilt vendor binary with no recorded provenance it is also the kind of file worth not shipping.

There is precedent: 6.1.10 removed a dozen committed libc.a and libgcc.a archives across arm9, arm11, cortex_a7/a8/a9/a15, m0/m3/m4/m7 and r4/r5 under "Removal of unneeded files". This PR catches two the sweep missed.

Rebased onto dev and retargeted from master. The third file in the original change, ports/arm11/gnu/example_build/libnosys.a, is already gone from dev — removed by #594 — so the rebase dropped that deletion and the remaining two apply cleanly.

One consequence worth recording separately: the 6.1.10 sweep removed those archives without updating the scripts that link them, so ports/arm9 and ports/arm11 examples now fail with cannot find libc.a / libgcc.a / libnosys.a. That is not caused by this PR and is being fixed on its own.