RFR: 8304015: G1: Metaspace-induced GCs should not trigger maximal compaction

[Ivan Walulya]

On Mon, 13 Mar 2023 11:37:36 GMT, Albert Mingkun Yang <ayang at openjdk.org> wrote:

> I think it's better to keep the var so that the semantic of the literal boolean is clear in this context.

Decided to go with using the comment /* do_maximal_compation */ as done in many other cases where we want to be clear but also avoid declaring variables. 

> (Slightly off the PR -- it's not obvious that metaspace related GC enters this path. Why isn't the decision on what GC cycle (young/old/clear-soft-or-not) to trigger made inside GC code?)

I think this is a design issue that we need to discuss along with other concerns around call paths to full-gc.

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

PR: https://git.openjdk.org/jdk/pull/12987