after emerging net-dns/dlint portage claims to have fixed non-posix compliant tail usage. but apparently it did not fix anything. The sed statement in fixheadtails.eclass fixes only tail +[0-9], but dlint uses a variable ($i). i'm no sed magician, so i can't come up with a fix myself - sorry. Reproducible: Always Steps to Reproduce: 1. 2. 3.
Looks like I'll have to revert to manual sed again. Thanks for reporting, I'll have a look at it shortly.
dlint-1.4.0-r2 fixes this. I've kept x86 keyword stable, and asked sparc team to test and mark stable for sparc. Thanks again, Heiko.