Coverage remained the same at 99.88% when pulling 87497ab on offa:macro_fixes into 7a58bf9 on cpputest:master.

I'm a bit about the while (0) causing a "always false" warning. Why wouldn't it?

I don't hope so. The do { .. } while(0) is a well known C / C++ construct for multiline macros. Also it's the only recommended way by MISRA C to solve this problem.

Do you have some specific platforms in mind?

Not really, just a worry. But let's proceed with this.

This did break pretty much all my training examples. Now I have to add semi-colons to CHECK, LONGS_EQUAL, FAIL usages.

I wonder how many more people will have their code broken.

Do you write different code in your trainings than in the book, articles and papers? 😕

I do, and often guiding clients. With a closer look, it's not all my training examples. I was somewhat inconsistent with semi-colons on checks. I've fixed them. Now its consistent, which is good too.

The book code was consistent to start with :-)

We might want to warn the mailing list. Was the mailing list warned? I could not find it quickly. I would have been really surprised if I did not remember this thread on github.

There is a mailing list? 😲

Maybe this is a good chance to create eg. a 3.9 release to indicate some kind of API differences?

I'll warn the cpputest google group for starters.