rpmlib: Support for creating/destroying RPM databases about crates HOT 2 CLOSED

Creating and destroying rpmdb's is not all that useful because there isn't the usual SQL schema instantiation involved with Berkeley DB.

Watch out for ancient hysterical baggage: there are 3-5 lurking interfaces that seem to be able to "open" a database. Note that none of those interfaces takes the necessary directory path as an argument. The directory path is set in a %_dbpath macro.

Similarly, the "initialize" operation is a remnant from a pre-BDB implementation. I'm pretty sure that BDB will create the necessary files as a side effect of, say, trying to create a key ring from the pubkeys in an rpmdb.

BDB should use a DB_CONFIG file to set parameters these days.

Note that there are two other rpmdb back ends that you need to worry about. No one has seriously considered what API's would be useful with alternative backends, so expect some pain if attempting bindings.

from crates.

Creating and destroying rpmdb's is not all that useful because there isn't the usual SQL schema instantiation involved with Berkeley DB.

Watch out for ancient hysterical baggage: there are 3-5 lurking interfaces that seem to be able to "open" a database. Note that none of those interfaces takes the necessary directory path as an argument. The directory path is set in a %_dbpath macro.

Similarly, the "initialize" operation is a remnant from a pre-BDB implementation. I'm pretty sure that BDB will create the necessary files as a side effect of, say, trying to create a key ring from the pubkeys in an rpmdb.

BDB should use a DB_CONFIG file to set parameters these days.

Note that there are two other rpmdb back ends that you need to worry about. No one has seriously considered what API's would be useful with alternative backends, so expect some pain if attempting bindings.

from crates.