RFR(XS) 8170981: Possible access to char array with negative index

Schmidt, Lutz lutz.schmidt at sap.com
Mon Dec 12 17:10:04 UTC 2016


thanks a lot for the review. An updated webrev (according to my comments
below) can be found here:


On 12.12.16, 17:07, "hotspot-runtime-dev on behalf of Rachel Protacio"
<hotspot-runtime-dev-bounces at openjdk.java.net on behalf of
rachel.protacio at oracle.com> wrote:

>I wonder whether it's better to declare
>    const char* spaces = "         ";
>and then set nSpaces as strlen(spaces), so it's clear that the spaces
>string will always be consistent with nSpaces?

Good idea, Rachel! I feel uncomfortable with the ³disconnected²
declarations as well. I had to use a non-constant char array because a
Œ\0¹ is stored into it.

Btw, a ³beautiful² fix would abandon the spaces array altogether and
implement columnar alignment with the help of tty->fill_to(pos). But that
would be a reengineering effort which might not take the Java9 hurdle.

>Otherwise, looks good to me.
>On 12/12/2016 9:45 AM, Schmidt, Lutz wrote:
>> Hi,
>> may I please ask for reviews and sponsorship for this small fix in
>> line flag printing?
>> Bug: https://bugs.openjdk.java.net/browse/JDK-8170981
>> webrev: http://cr.openjdk.java.net/~simonis/webrevs/2016/8170981
>> Thank you!
>> Lutz

More information about the hotspot-runtime-dev mailing list