RFR: 8348323: Corrupted timezone string in JVM crash log

[Kevin Walls]

On Fri, 24 Jan 2025 09:21:35 GMT, Kevin Walls <kevinw at openjdk.org> wrote:

>> Add a error check to "wcstombs" otherwise the "buf" prints junk or corrupted string.
>> 
>> Test performed: 
>> 1. "jcmd < jvm pid > VM.info" for different JVM versions
>
> Hi,
> It's not random corruption, as it's the same across different machines.
> We are getting timezone printing wrong, and this is I think a way of not printing the timezone.
> 
> It would be great to have more info/examples in the JBS issue.
> The one example in the JBS issue contains "-eur" so perhaps part of a timezone name in Europe...
> 
> What timezone are the machines actually set to?
> Is there a TZ variable set? like in JDK-8340465

> Thanks @kevinjwalls, One of our client's crash collection systems has captured this report; however, we currently do not have information regarding the locale or timezone of the user at the time the crash occurred.

OK thanks.  It seems like there is more to the mystery that is not yet solved. 8-)

I don't disagree that we should check wcstombs' return value.  When we look at this kind of thing in future it will be no timezone printed that means we have hit an existing problem fetching a valid timezone.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/23246#issuecomment-2613880497