So from your output it seems:
virStorageFileMetadataNew
virStorageFileGetMetadataFromFD
virStorageSourceCopy
virStorageSourceUpdateCapacity
virStorageSourceNewFromBackingAbsolute
virStorageSourceNewFromBacking
virStorageFileGetMetadataRecurse
virStorageFileGetBackingStoreStr
or IOW everything in src/util/virstoragefile.c other than I think
virStorageSourceNewFromBackingRelative, although there is one pile of
compiler output without an "In function '%s':" before the compiler
error
output "inlining failed in call to" before it which I assume is the
*Relative call. Dropping those means dropping a bunch of patches before
too since the only reason to change was to do this AUTOPTR magic.
As a side note I find it really odd that the order listed in the
compiler output doesn't exactly follow the order in the source. In
particular starting with virStorageSourceUpdateCapacity (line 3832) is
listed in the output before virStorageSourceNewFromBackingAbsolute (line
3613).
That could be the effect of make's --jobs.
Still it makes me wonder what's different in src/util from others that
causes src/util to throw up, while src/conf, src/qemu, tests/, etc.
don't throw up. Reading m4/extern-inline.m4 doesn't help me.
The error message is not really helpful in figuring ^this out, I'm puzzled too.
Erik