@todo DOC ABOUT setting up vscode and using .natvis files

GDB and ASAN (Address Sanitizer) and UBSAN (Undefined Behavior Sanitizer)

br __ubsan::ScopedReport::~ScopedReport or __ubsan::Diag::~Diag
br __asan::ReportGenericError
see https://stackoverflow.com/questions/30809022/how-can-i-break-on-ubsan-reports-in-gdb-and-continue

TSAN (thread sanitizer) on Ubuntu 24.04 (Host - not container) and later

see https://stackoverflow.com/questions/77850769/fatal-threadsanitizer-unexpected-memory-mapping-when-running-on-linux-kernels

sudo cat /proc/sys/vm/mmap_rnd_bits < /dev/null
if result > 28 - seems to default to 32 now must do
    sudo sysctl vm.mmap_rnd_bits=28
or TSAN will fail
 
NOTE - even though failure in container, and test can be done in either place, the SET must be done in the host (or special docker settings)

NOTE - though I've found no clear docs on this - setting rnd_bits=28 also appears to fix an issue on Ubuntu 23.10 (setting it in Ubuntu 24.04 host) - regtests for ASAN - where sporadically we get crash on startup of ASAN test code –LGP 2024-06-12

Valgrind (memcheck, and helgrind)

MemCheck

Stroika is tested with valgrind memcheck for each release. REVISIT to see if/how useful this is, since I cannot recall ever finding anything useful with it - just a bunch of false positives and effort to workaround stuff that was tearfully slow.

Helgrind

I spent TONS of effort over the years to study things it reported, and craft workarounds for obvious false positive warnings. I eventually gave up (as what use it migth have is presumably provided better by TSAN).

Sadly - they just don't really appear to support std c++ threading primitives (except sometimes where they fallback neatly on POSIX primitives). But things (and I have alot of them) - that use std::atomic, just flat out get mis-diagnosed, and I grew weary of the silencing and efforts to help it to understand with follows etc annotations.

So - as of Stroika v3 - no more helgrind support.