RFR: 8334502: gtest/GTestWrapper.java fails on armhf due to LogDecorations.iso8601_utctime_test [v3]
David Holmes
dholmes at openjdk.org
Thu Jul 18 05:24:35 UTC 2024
On Wed, 17 Jul 2024 20:00:40 GMT, Vladimir Petko <vpetko at openjdk.org> wrote:
> I was initially concerned we are narrowing to `int` to please the format specifiers. Normally, I would have expected us to work out a format specifier based on `time_t` size, and put it in `globalDefinitions.hpp`. But this looks okay as the unblocking fix.
Those two variables, like many others in that code are very obviously just small integer values. I think `time_t` use should be minimised as much as possible e.g. `seconds_per_minute` et al. should really just be int values IMO.
> correct fix should've been changing [UTC_to_local](https://github.com/vpa1977/jdk/blob/ff97727f9a7141451dcc9a2565d277368c4b5d99/src/hotspot/share/runtime/os.cpp#L180) to long
If you do that won't it just introduce different conversion issues later in the code. This whole code is really quite a mess with regards to types, but that is partly because the API and structures used are also a mess when it comes to types.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/19782#issuecomment-2235414082
More information about the hotspot-runtime-dev
mailing list