RFR: 8317951: Refactor loading of zip library to help resolve JDK-8315220 [v2]

[Markus Grönlund]

On Fri, 20 Oct 2023 06:12:47 GMT, David Holmes <dholmes at openjdk.org> wrote:

>> Realized that the usage from HeapDumperCompression should not exit the VM if there is a problem loading up the zip library. Now updated to support that usage better.
>
> Okay so if we have no control over the order in which functions are called then they all must initialize, including `close` - no?

Close seems obvious, as you pointed out earlier, to be subsequent from open. An assert is there now, to capture erroneous usages, which I think is the correct way to capture wrong callers.

The complication comes with init_params() and compress() which are used in a jcmd heap dump routine, it does not open anything, it merely wants to re-use the logic of functions in the zip library.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/16149#discussion_r1366693177