On Sun, 1 Oct 2023 11:48:07 GMT, Florian Kirmaier <fkirma...@openjdk.org> wrote:
> After I ran into Windows issues, I assumed this was a regression that > affected everyone. @FlorianKirmaier You were quite right! Thank you for finding this issue and for quickly fixing its worst effect. I introduced this error in #446 when I changed the build timestamp from its add-hoc format to the standard ISO 8601 format. Kevin caught the error in his review and wrote, "Since the : is not legal for Java version strings, it is possible (though unlikely), that some app is parsing this in a way that might run into problems. This should probably be fixed." I fixed the version string in the build file, but I failed to notice that the version string is reconstructed again at runtime using the same new ISO 8601 timestamp. Your quick fix gives me a little more time to come up with a more comprehensive fix for the bug I caused. The new test cases that reproduce the error and test its fix are found in my repository at [jgneff/version-info-test][1]. I'll be creating the bug report and submitting a pull request shortly. I added [test cases][2] to make sure this never happens again. [1]: https://github.com/jgneff/version-info-test [2]: https://github.com/jgneff/version-info-test/commit/4ebd3384bd6691f14fbb10c73bbdd7a6960b4f54 ------------- PR Comment: https://git.openjdk.org/jfx/pull/1251#issuecomment-1742146278
