On Dec 11, 2007, at 8:29 PM, Per Øyvind Karlsen wrote:
I noticed that my second patch was a bit zealotish, more correct
one is at
http://hg.rpm.org/rpm?cs=053b7741a883 , okay to adapt and commit?
Or should it stay as it is?
Make yourself happy. Names and locations of routines matter only
to those who think rpmlib needs an API/ABI, and attempting an API/ABI
when
multiple forks of rpm exist is pointless.
The reason for rpmcliImportPubkey naming was to have all CLI modes
of rpm have a common rpmcliFoo prefix. The rpmts object is at a lower
layer then the CLI routines, and there is no known usage case for
importing
pubkeys at the rpmts layer.
FWIW, one of the flaws of the rpmlib API is that most applications are
trying to get at rpmlib innards like rpmtsFoo rather than using
the far simpler rpmcliFoo API. But that's the rpmlib hysterical
accident,
taking an installer API and trying to turn it into a useful package API.
Go fer the checkin!
73 de Jeff______________________________________________________________________
RPM Package Manager http://rpm5.org
Developer Communication List rpm-devel@rpm5.org
