Right now udev handles binaries' stripping by itself, instead of leaving that up to portage, so when using splitdebug the .debug files generated are stripped and thus useless. By passing STRIP=true to the make line on src_compile, udev does not strip the binaries and leaves that up to portage. HTH, Diego
I'm sorry, but what do you want me to do with this? Is this an informational thing? Or do you require something to be changed in the ebuild for udev?
Well if it's not a requirement that udev strips them by itself for some reason, the ebuild should pass STRIP=true to the make as I said to leave to portage handling binaries stripping (that way splitdebug will work). If instead the strip has to be done by udev's makefile (but I can't see why), the ebuild should RESTRICT=strip.
Great, care to offer up a patch to the udev ebuild that shows what you are suggesting?
Created attachment 82056 [details, diff] Ebuild patch
Attached (it's literally what I said before :P) I'm using this for a few days now and system is fine.
adding an empty: STRIPCMD= should do the trick without magically invoking /bin/true. I'm currently thinking if we should better make the non-stripping of binaries the default bahavior in the Makefile, because we don't provide a functional "end user install" anyway from the upstream udev tree.
089 release doesn't strip anything. marking as fixed.