On Tue, 4 Oct 2022 15:48:18 GMT, Markus KARG <d...@openjdk.org> wrote:
> As the files are not zero'ed by intention (to detect differences), sparse > won't help, so next you'd eventually run into quota trouble (IIRC we already > had this a year ago). IMHO we should increase RAM a bit further (really, it > is cheap these days), or reduce that size of the tested data itself. I also pasted an example into the bug where 1Gb isn't enough. If it requires too much then the test may have to be skipped on 32-bit builds. ------------- PR: https://git.openjdk.org/jdk/pull/10524
