Jan Luca Naumann: > Both reported problems seem not to be fixed in the current version yet > so you want to fix them maybe :-)
Thanx lettng me know. I didn't know that. I don't think these are bugs since they are following "Dashism" instead of "Bashism." I understand that debian has a policy to be posix-compliant shell script, but DASH (which debian recommends) supports these non-posix-compliant features. I don't know why this contradiction happes. Maybe there are too many scripts using these features in our real world? As an upstream developer, I won't fix them because it is a debian local policy. But someday if DASH stops supporting these features, I will re-consider. More simply if someone who really cares the policy sends a patch to me (which must be a generic solution for other shells too), I will probably re-consider too. IIRC, there was a similar discussion on aufs-uses ML about "test cond1 -a cond2" vs "test cond1 && test cond2". My conclusion was same, ie. it was supported by DASH. J. R. Okajima
