Jérémy Bobbio wrote: > Actually, it is a lot cheaper than having another udeb to track from a > d-i RM point of view... Is that really true? If bin-nmus, which require finding and prodding one of a small set of very busy people, are more maintenence cost than managing one more udeb, we're doing something wrong. It certianly can be true if the library is used by lots of packages and tends to be involved in complex testing transitions. (Agurably that's because there's something wrong, like like too-tight library deps that cause such transitions.) I don't see any reason why libaio1-udeb would be harder to manage than other simple udebs though. The reverse dependencies of libaio are very small. -- see shy jo
Attachment:
signature.asc
Description: Digital signature