RFR: 8319373: Serial: Refactor dirty cards scanning during Young GC [v10]

[Thomas Schatzl]

On Tue, 21 Nov 2023 14:17:30 GMT, Albert Mingkun Yang <ayang at openjdk.org> wrote:

>> I don't think I would have ever guessed that, which suggests it might not be an ideal naming choice.
>
> My concern regarding `cur_card` is that there is no contrasting `prev_` or `next_` in this context. The sole message I want to convey here is that we are iterating through cards. Using the conventional loop variable `i` plus `card` seems like the most natural approach, in my opinion. It is concise, so it doesn't introduce much overhead despite its frequent occurrence in the context. Therefore, I prefer the current name. By the way, the same style and convention are also employed in other methods in this PR. Additionally, similar naming conventions can be found in the already merged code of Parallel (`PSCardTable`) and G1 (`G1ScanHRForRegionClosure`).

I concur with Kim and would like to ask to rename these variables to use the `cur_` prefix as is much more common. I do not see a connection of `cur` with linked list prev/next pointers; in the previous recent PRs that introduced this naming I (in hindsight wrongly) did not speak up because although I did not like it, nobody else objected either.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/16492#discussion_r1401654457