RFR: 8255119: Monitor::wait takes signed integer as timeout

[David Holmes]

On Mon, 9 Jan 2023 21:32:36 GMT, Gerard Ziemski <gziemski at openjdk.org> wrote:

> Looks like there are other APIs where we use signed values for the time, which is illegal - do we have an issue tracking that cleanup?

No, as per earlier comments in this bug report that code is fed by signed values that come from Java. We could impose a signed -> unsigned conversion at the VM boundary, but then we also have to redo all the  conversion routines to pack and unpack that time structs that we use. I don't see any point in doing all that.

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

PR: https://git.openjdk.org/jdk/pull/11898