Most of the USE flags are actually described in use.local.desc but do are not for the particular package. Could probably use this list to move a few over to use.desc.
Created attachment 27803 [details] Current List
dev-libs/glib-2.2.3 - s390 not in global or local use flags that is an arch flag, so it should be considered a false positive
Apologies. I noticed "ia64" but missed "s390". I assume these wont be added to use.desc until they are officially supported?
don't really know for sure, but I'd guess so. hmm, vorbis there should be oggvorbis, and I think faad should deserve its own global flag from how things look
postgresql pgsql, aren't those the "postgres" USE flag?
xfree is false positive, it's using the USE_EXPAND functionality of portage to have INPUT_DEVICES. I've added IUSE_INPUT_DEVICES for this purpose, although I doubt portage currently has the ability to parse this or even a reason to do so, since no parallel to use.*desc exists. In fact I'm curious as to why you didn't get the same false positive with xfree-drm-4.3.0-r7.
Resolving this as later for the time being. When I can come up with a way to ensure that there are no false positives AND that no packages are missed, I'll reopen with a better quality list. Donnie, the xfree-4.3.0-r6 ebuild has the input devices in DEPEND. The script I used doesn't/can't check for the "use xxx &&" that's also commonly used. Hence, the other versions didn't match.
Closing as everything is nowdays in use.desc or metadata.xml.
