Cristian Morales Vega

I was looking at the specific warnings, and I am surprised lines ending with `[-Wformat-overflow=]`, for which the regex will match `-Wformat-overflow=` to the id capture group... actually end up...

My employer has to approve, but I will create a PR with the gcc rules updated for 14.2.0 and a python script to create them "soon". This issue can be...

> Yes right, not supported. There is always a long discussion about whether compiler warnings should be included or not, or whether they should be checked in a CI/CD system...

> Also for some reason I can't get the coroutine tests working on 1.87 Without even looking at those tests, could it be this (from the ASIO changelog) > Removed...

I certainly don't see the issue. I have [reported it to gcc](https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120915). Let's see what they say.

Not saying Boost.JSON should do it. But FWIW you can have `__attribute__ ((no_sanitize ("pointer-subtract")))` in `remain()`, and `__attribute__ ((no_sanitize ("pointer-compare")))` in `operator bool()`, and gcc stops complaining. Unfortunately, it's different...