RFR: 8273544: Increase test coverage for snippets [v2]

[Jonathan Gibbons]

On Fri, 19 Nov 2021 16:35:16 GMT, Jonathan Gibbons <jjg at openjdk.org> wrote:

>> What you suggested is how I had that initially. But then a test purist in me demanded a clear way of getting the link without interfering with the test logic.
>> 
>> As for text blocks: to me, indenting a text block is a complex, subjective and unsettled issue. I would postpone re-indenting them for now, primarily because of the churn it would otherwise introduce to the dependent PRs. 
>> 
>> If neither of these issues is a showstopper, I would really like to integrate this PR to unlock more urgent work on the snippets feature before RDP 1, which is on 2021/12/09 as per the Schedule at https://openjdk.java.net/projects/jdk/18/. Sometime after this PR is integrated, we can revisit this.
>
> OK, but the test purist in me dislikes the significant amount of inline overhead to address this concern.
> 
> I'll pass on this for now, because I agree with unblocking the dependent PRs, but it seems like that is a negative side-effect of dependent PRs in general.  Once we get past this, at a minimum, I would suggest moving the in-memory file manager out of this class, and then potentially reconsider the decision that it is good practice to do the separate run.

> As for text blocks: to me, indenting a text block is a complex, subjective and unsettled issue. I would postpone re-indenting them for now, primarily because of the churn it would otherwise introduce to the dependent PRs.

_complex, subjective and unsettled_?  Really? JEP 378 would indicate otherwise!   So, I disagree with the lack of indent, but (reluctantly) agree on the churn causing issues for dependent t PRs.

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

PR: https://git.openjdk.java.net/jdk/pull/6359