Re: Smoke [ebcdic] v5.23.4-176-g61739e5 FAIL(M) os/390 24.00 (2964/)

> I haven't been able to figure out what is the problem.  Again, it says
> there is an unterminated #if preprocessor directive affecting utf8.c,
> but the apparent line number given is in the middle of a comment.  I
> have examined utf8.c and the header files that I have modified, and
> don't see such a thing.  I also hacked the compilation on my Linux box
> to pretend it was an EBCDIC machine, and it compiles.
>
> So what I've done is submitted a new branch to test, updated to latest
> blead, but with a bunch of commits backed off, to see if it works at
> this point.  If so, then we can add in some commits to eventually find
> which one is the breaking one.  If not, then we can try with even more
> commits backed off, etc, etc.
>
> If someone wants to examine this, I've attached the first part of the
> log in ASCII.  A diff of this log and the latest one that works, shows
> the first difference at line #56:
>
> ERROR CCN3198 utf8.c:99    #if, #else, #elif, #ifdef, #ifndef block must
> be ended with #endif.
>
> I don't understand why utf8.c would be getting compiled this early in
> the process, but there it is.

Yaroslav, does

make utf8.i

work in z/os to generate the preprocessed source?  If it does, Karl 
might find that useful.  Or not.