= Proposal topic =
When doing some research for systemd I came across what seems to be a packaging bug which seems to be either fault in the guidelines or consensus has not been reach on which one to use /etc/init.d or /etc/rc.d/init.d/
Tools such as repoquery and yum are unaware that these are the "same".
Running repoquery against /etc/init.d/ ( LFS spec ) yields 45 results Running repoquery against /etc/rc.d/init.d/ yields 488 results
Regardless of which one is correct there can be only one so we are dealing with 45 minor packaging bugs or 488 minor packaging bugs.
When/If consciousness has been reached and documented, I'm happy to file bug against the packages in question refer to the documents and get this sorted out with the maintainers.
But there can be only one...
This is an FPC matter rather than FESCo. It's also already been decided:
https://fedoraproject.org/wiki/Packaging:SysVInitScript#Initscripts_on_the_filesystem
%_initddir (%_initrddir on EPEL4 and 5) must be used which puts things in /etc/rc.d/init.d. As you noted, yum, rpm, and repoquery do not understand that the paths refer to the same file which leads to the rationale in the packaging guidelines:
""" This split made it more difficult to use yum install against the initscript filename and path. """
Log in to comment on this ticket.