RFR: 8312502: Mass migrate HotSpot attributes to the correct location [v3]

[Kim Barrett]

On Tue, 25 Jul 2023 01:41:24 GMT, Julian Waters <jwaters at openjdk.org> wrote:

> Hmm... I was doing this so that what attribute specifiers applied to would be clearer since it's helpful to have the ability to make them apply to different aspects of a method as per C++11. Could I split this up and discard some of the areas in which the changes are less agreeable? (as advised by someone who reached out to me privately)

That can be done when other attributes are being added to a declaration. For
example, various error reporting functions had the `[[noreturn]]` attribute
added by JDK-8303805. Many of them also had a trailing ATTRIBUTE_PRINTF, which
was moved to the front at the same time. There were also one or two functions
there whose trailing ATTRIBUTE_PRINTF was moved to the front even though not
noreturn, for local consistency (e.g. `warning`).  (So yes, sometimes it's
okay to normalize some outliers.  That doesn't seem like what's going on in
this PR though.)

There currently aren't other standard attributes to be dealt with.  That may
change as we adopt newer versions of the standard.  But even so, the scope of
the adoption changes is going to be limited.

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

PR Comment: https://git.openjdk.org/jdk/pull/14969#issuecomment-1648898589