RFR(s): 8178364: Command-line flags of type double should accept integer values

Kim Barrett kim.barrett at oracle.com
Tue Apr 11 23:12:33 UTC 2017

> On Apr 10, 2017, at 7:03 AM, David Holmes <david.holmes at oracle.com> wrote:
> On 10/04/2017 5:40 PM, Per Liden wrote:
>> With the way this code is structured we only end up in
>> set_fp_numeric_flag() if the string looks like a float to begin with,
>> i.e. has the format "<flag>=<num>.<num>". If the flag looks like
>> "<flag>=<num>" we end up in set_numeric_flag(). Have a look at the call
>> sites for set_fp_numeric_flag() and set_numeric_flag() and you'll see.
> That's somewhat odd - shouldn't the type of the flag dictate how we try to parse it ??

Coming to this a little late, but I remember making that same point during the review
of the command line range checking stuff, and thought there was a bug filed for it.
But I can’t find any such bug.

More information about the hotspot-runtime-dev mailing list