Hennadii Stepanov

As an alternative, we could explicitly set our own `SUPPRESS_ABORT_MESSAGE_BOX` environment variable in the workflow, and gate the code with this variable instead, ~removing the `#if defined(_MSC_VER)` condition~.

> > It is documented, and it is pre-defined by CI itself: > > It's not in our docs, or in our CI code, which means our CI would "work",...

This issue seems related: https://developercommunity.visualstudio.com/t/Debug-CRT-shows-abort-message-box-instea/10904782.

Added https://github.com/arun11299/cpp-subprocess/pull/117, which is upstreamed https://github.com/bitcoin/bitcoin/pull/32342.

> From what I found, the following merged upstream PRs since our initial cpp-subprocess import (see PR #28981, commit [cc8b987](https://github.com/bitcoin/bitcoin/commit/cc8b9875b104c31f0a5b5e4195a8278ec55f35f7), based on upstream [arun11299/cpp-subprocess@4025693](https://github.com/arun11299/cpp-subprocess/commit/4025693decacaceb9420efedbf4967a04cb028e7)) are not included yet in this...

> For easier review, it might make sense to add a list of the backported upstream PRs also to the PR description. Thanks! Done.

> > During my quick (and possibly not thorough) tests, I was unable to reproduce the issues mentioned in those PRs on our codebase. > > Seems odd to not...

Friendly ping @laanwj @Sjors :)

> Is this currently an issue for anything? The newest current option--`-Wno-overlength-strings`--is supported for GCC >=4.2. Only someone using an older compiler would be affected. > If the check is...

What about `ctime_tests`?