RFR: 8232211: Remove dead code from os.hpp|cpp

Claes Redestad claes.redestad at oracle.com
Wed Oct 16 10:17:09 UTC 2019

On 2019-10-16 11:59, David Holmes wrote:
> On 16/10/2019 6:46 pm, Claes Redestad wrote:
>> On 2019-10-16 04:37, David Holmes wrote:
>>>> os_version is
>>>> never read, which means initialize_os_info can go - too.
>>> Hmmmm. This is the kind of information that could be used then not 
>>> used then used again on each release, depending on whether we 
>>> introduce or remove code that is conditional on the OS version. I 
>>> guess we can always add it back ... though I'm willing to bet that if 
>>> we need this again the person making the change will reinvent the 
>>> code rather than resurrecting it. :(
>> I don't have a problem with this. At the very least I'd reinvent it into
>> a struct so that major, minor and fix version was addressable and not
>> something you need to extract from a 24-bit integer...
> The format is intended for simple comparisons as you can see in AIX/PPC 
> usage.

And if someone were to reinvent this I'm sure keeping a packed
representation internally while exposing compare_to, major, minor and
fix accessors would be an excellent way to design it.


More information about the hotspot-runtime-dev mailing list