ghostscript already was a pretty slow build, but not *that* slow.

Is that all the emerge logs we can get? It looks like there are hours of time in which there's no output from the ghostscript build (but 'Still building...'). It's possible it's truly hung, and so there's no more stdout, but I just wanted to confirm.

At this point, those logs are all we've got to go on.

I'd wager that it was hung and there was no further stdout.

Any possible way https://chromium-review.googlesource.com/759082 could be involved?  I can't see how, but it's the only thing anywhere near this that has changed recently AFAIK.

Looks like there could be some kind of circular expansion?  This message keeps appearing.

ghostscript-gpl-9.19-r9: #define min_int (-1 << (ARCH_SIZEOF_INT * 8 - 1))
ghostscript-gpl-9.19-r9:                  ~~ ^
ghostscript-gpl-9.19-r9: ./base/gsfunc4.c:342:42: warning: shifting a negative signed value is undefined [-Wshift-negative-value]
ghostscript-gpl-9.19-r9:             if (prod < min_int || prod > max_int)
ghostscript-gpl-9.19-r9:                                          ^~~~~~~
ghostscript-gpl-9.19-r9: ./base/std.h:66:19: note: expanded from macro 'max_int'
ghostscript-gpl-9.19-r9: #define max_int (~min_int)
ghostscript-gpl-9.19-r9:                   ^~~~~~~
ghostscript-gpl-9.19-r9: ./base/std.h:65:21: note: expanded from macro 'min_int'
ghostscript-gpl-9.19-r9: #define min_int (-1 << (ARCH_SIZEOF_INT * 8 - 1))
ghostscript-gpl-9.19-r9:                  ~~ ^
ghostscript-gpl-9.19-r9: ./base/gsfunc4.c:353:33: warning: shifting a negative signed value is undefined [-Wshift-negative-value]
ghostscript-gpl-9.19-r9:             if (vsp->value.i == min_int)
ghostscript-gpl-9.19-r9:                                 ^~~~~~~
ghostscript-gpl-9.19-r9: ./base/std.h:65:21: note: expanded from macro 'min_int'
ghostscript-gpl-9.19-r9: #define min_int (-1 << (ARCH_SIZEOF_INT * 8 - 1))
ghostscript-gpl-9.19-r9:                  ~~ ^

Do you see any kind of loop in that? I think this is just an instance of a macro (or set of macros -- min_int, max_int, etc.) that is used in a lot of places, and clang doesn't like it. Makes the logs extremely noisy, but I don't see anything inherently wrong in that.

Is this still happening? 
and, if so, is the priority correct?

I haven't seen this in months so I'm going to guess it just flaked.