RFR: 8298728: Cells in VirtualFlow jump after resizing [v2]

[Johan Vos]

On Wed, 21 Dec 2022 23:43:03 GMT, Andy Goryachev <angorya at openjdk.org> wrote:

> 1. I see the same issue with page up/page down described earlier (see the video clip attached to the jira ticket https://bugs.openjdk.org/secure/attachment/102048/does%20not%20page%20back%20to%20the%20same%20spot.mov )
> 2. I am still not clear what exactly the problem is.  The jira ticket contains conditionals ("if we want..."), rather than a typical `steps to reproduce -> expected to see -> actual behavior` steps.
> 
> Could you comment on #1 and #2 please?

#1: I had a look at that video and I don't think this is a bug. Nowhere in the spec or javadoc it is mentioned that a page-up followed by a page-down should give the same position. This would imho actually be a bad requirement, as in case sizes are changing (which happens to be the case in most real-world scenario's I've seen so far) this would conflict with other requirements. One might argue that "in case the size of individual cells didn't change", the scroll-up/scroll-down should result in a same "offset" as before the operation but that is not a requirement so far. Could be discussed in another issue.

#2. I don't see what is missing. It seems to me both the ticket and the new test in the PR show the scenario of what goes wrong before the patch, and what is expected (in the assert). I can copy it here again, but I'm not understanding your question I think.

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

PR: https://git.openjdk.org/jfx/pull/974