Added asan and ubsan to CI

[vasdal]

Problem:
Discovered that asan,ub triggers on gcc
Fails:

• https://godbolt.org/z/q6K48hTaf
• https://github.com/boost-ext/sml/actions/runs/27272893452/job/80546599233

Solution:

• Added sanitizers to CI pipeline, and fixed failing fall positive issue
  Fixed: https://godbolt.org/z/noKnzbh8K

Issue: #

Reviewers:
@kris-jusiak
@PavelGuzenfeld

[PavelGuzenfeld]

@vasdal ,
Thanks for the PR. I'll reproduce it locally and get back to you.

[PavelGuzenfeld]

Reproduced on GCC 13.3.0 (what ubuntu-latest and the Godbolt g133 link use). The issue reproduces.

• No sanitizer (-O0 and -O2): fine.
• asan+ubsan -O0 (your Debug job): UBSan OOB index + real ASan SEGV at sml.hpp:2312.
• asan+ubsan -O2: fine.

I printed the actual event.id and it's 1 — a valid index into a 2-entry table. So the SML logic is fine; GCC 13.3 just emits a garbage index for the folded (this->*dispatch_table[event.id])(...) under sanitizers at -O0. 13.4 and 14.x don't repro. Your hoist-to-a-local fix is clean — full ft suite 39/39 under asan+ubsan.

The existing test/ft/actions_process.cpp already hits this path, so it segfaults on master under the new job and passes with the fix — no new test needed, it just slipped through because there was no sanitizer job before.

Two small comment nits (it lives in the header): it says "GCC-11" but everything I saw fail is GCC 13.x, and "false positive" is really a miscompilation/crash. Otherwise LGTM.

[vasdal]

Thanks @PavelGuzenfeld, fixed the comment now.
Possible to make g++ asan+ubsan required in the future?

[kris-jusiak]

Thank you @vasdal and @PavelGuzenfeld