As documented in InternalDocs/garbage_collector.md, the garbage collectorstores flags in the least significant two bits of the _gc_prev pointerin struct PyGC_Head. Consequently, this pointer is only capable of storinga location that's aligned to a 4-byte boundary.This alignment requirement is documented but it's not actually encoded.The code only works when python happens to run on a platform that has asufficiently large minimum alignment for the structs in question.The same problem arises with PyObject pointers because the leastsignificant bits get used for PyStackRef tags.Since we know that 2 bits are needed, we also know the minimum alignmentthat's needed. Let's make that explicit, so the compiler can then makethose bits available.This patch fixes a segfault in _bootstrap_python. In 3.14.0 beta 2this fixes the "Assertion `!PyStackRef_IsTaggedInt(ref)' failed" whenbuilt with --config-pydebug.Also, making the requirements explicit improves clarity.This bug was previously investigated by Adrian Glaubitz here:https://lists.debian.org/debian-68k/2024/11/msg00020.htmlhttps://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1087600Although Adrian's patch isn't really correct (because natural alignmentis not needed), he deserves full credit for finding the root cause.

This is a common façade for the various `_Alignas` alternatives,which behave in interesting ways.The standard `alignas` errors if it would decrease the alignment;to prevent that it can be used again with the defined variable'stype.The standard `alignas` can't be used on a `struct` definition,the workaround is to use it on one of the `struct`'s members.MSVC `declspec(aligned)` only takes integer literals.MSVC `declspec(aligned)` had a bug when applied to a combinedstruct+member definition.The workaround is to separate the struct definition.Do that for `PyASCIIObject.state`.