On Wed, 2003-07-09 at 23:43, Manoj Srivastava wrote: > At this point, we are indeed supposed to preserve user > changes. User changes can be echo "" > /etc/filename, or rm > /etc/filename.
If it were clear that removal were a "change" then we would not be having this discussion. The fact that there is resistance to making policy say clearly that removal is a change (and therefore not to be undone by regenerating the removed file) indicates that some people interpret policy such that config file removal is not a "change". > Packages already have to deal with effectively removing > configuration files -- (the echo "" device). How is it so different > to also expect them to respect non existent files? (even if the > response is making the package a nop) Someone who disagreed with Manoj's position would probably argue that removal of the configuration file is different from emptying the configuration file in that it is the sort of thing that might more easily be done inadvertently. Consequently it is safer or more helpful to regenerate a missing configuration file, and not an empty one. I agree with Manoj. A package should not regenerate configuration files without asking, if it is possible for the programs in it to run without those files. At most it should print a warning message and _offer_ to regenerate the file. However, there are plausible arguments on the other side that have convinced some people. Therefore there will be no agreement on this point. Therefore, I suggest that we agree instead to document the current practice. -- Thomas Hood
