parsing these is significantly less work and more robust than parsing
C source.

Cgo first invokes gcc -E -dM on the preamble, in order to find out
about simple #defines for constants and the like. These are recorded
for later use.

Next, cgo needs to identify the kinds for each identifier. For the
identifiers C.foo, cgo generates this C program:

	<preamble>
	#line 1 "not-declared"

	paxSize     = "size"
	paxUid      = "uid"
	paxGid      = "gid"
	paxUname    = "uname"
	paxGname    = "gname"
	paxMtime    = "mtime"
	paxAtime    = "atime"
	paxCtime    = "ctime"   // Removed from later revision of PAX spec, but was valid
	paxCharset  = "charset" // Currently unused
	paxComment  = "comment" // Currently unused

	paxSchilyXattr = "SCHILY.xattr."

		b.uint16(h.ModifiedTime)
		b.uint16(h.ModifiedDate)
		b.uint32(h.CRC32)
		if h.isZip64() || h.offset >= uint32max {
			// the file needs a zip64 header. store maxint in both
			// 32 bit size fields (and offset later) to signal that the
			// zip64 extra header should be used.
			b.uint32(uint32max) // compressed size
			b.uint32(uint32max) // uncompressed size

			// append a zip64 extra block to Extra