add compensational definition for OPENSSL_MSTR in macros.h

[DDvO]

The recent commit f386632 broke include/opensll/cmp_util.h when the trace API is enabled.
This PR adds in that header file a compensational declaration.

[levitte]

It would probably be a better idea to add them in includ/openssl/macros.h and give them a more generic name (i.e the original names)

[DDvO]

all right, done

[mattcaswell]

These two macros were two of the "newly added but undocumented" macros that got into the 3.0 codebase before we set doc-nits to start complaining about undocumented stuff. See #9095. f386632 co-incidentally fixed that by removing the macros - so this PR introduces it again. doc-nits doesn't complain because they weren't also removed form missingmacros.txt. But we should probably document these macros.

[DDvO]

No problem to add some documentation for these two macros, but where to do this?

I would have put it wherever similar macros are documented, but also (most of) the other macros in macros.h, including OPENSSL_API_LEVEL, OPENSSL_FILE and OPENSSL_FUNC are not (yet) documented either.

[levitte]

Do these have to be public?

[DDvO]

Currently the two macros are used only in the public header include/openssl/cmp_util.h, namely in the implementation of the OSSL_CMP_LOG_START macro. So, for technical reasons, the macro is already public, but we could #undef it after use.

On the other hand, the CCP string composition trick is a generally useful macro (and can be found online as C programming hint under various names). So we could also export it for the benefit of others.

Which way do you prefer?
If we go for exporting it, in which file should I document it?

[levitte]

Nah, if we use it publicaly, let's expose if publicaly.

[DDvO]

Fine, and where to document these macros?

[DDvO]

Since I got no response on where to document these two macros,
I've introduced for this purpose doc/man3/OPENSSL_FILE.pod,
where I also documented OPENSSL_FILE, OPENSSL_LINE, and OPENSSL_FUNC.

[levitte]

Uhmmmm, this PR is suddenly a minor gazillion commits. What happened?

[DDvO]

Hmm, it looks like rebase went somewhat wrong. Fixed.

[DDvO]

Recent Travis CI hang resulted in false positive. This time it worked.

[DDvO]

Can someone please give a second approval on this and merge it soon?
Our CMP contribution chunk 6 depends on this.

[mattcaswell]

Can someone please give a second approval on this and merge it soon?

Merge happens (at the earliest) 24 hours after "approval: done" is reached.

[DDvO]

Thanks for the swift (re-)action!

[mspncp]

ping: ready to merge

[mattcaswell]

I pushed this. Note: I merged the last 2 commits into 1.