8229280: Test failures on several linux hosts after JDK-8181493

[Brian Burkhalter]

Ping ...

> On Aug 9, 2019, at 5:05 PM, Brian Burkhalter <brian.burkhalter at oracle.com> wrote:
> 
> Here is a potential fix [1] for the issue [2] provoked by the fix for [3]. The change is not to require strict equality in timestamps except for some file system types which are known to support nanosecond precision and are explicitly identified. For non-strict timestamp comparisons, equality is defined by the timestamp read back from the file having a number of trailing zeros which is a multiple of 3, and having a prefix which is equal to a prefix of the same length of the nanosecond precision timestamp.
> 
> If there are more file system types which should be added to the list of strictly tested types, please let me know.
> 
> Thanks,
> 
> Brian
> 
> [1] http://cr.openjdk.java.net/~bpb/8229280/webrev.00/ <http://cr.openjdk.java.net/~bpb/8229280/webrev.00/>
> [2] https://bugs.openjdk.java.net/browse/JDK-8229280 <https://bugs.openjdk.java.net/browse/JDK-8229280>
> [3] https://bugs.openjdk.java.net/browse/JDK-8181493 <https://bugs.openjdk.java.net/browse/JDK-8181493>
> 
> 
>> On Aug 8, 2019, at 8:52 AM, Brian Burkhalter <brian.burkhalter at oracle.com <mailto:brian.burkhalter at oracle.com>> wrote:
>> 
>> HI Christoph,
>> 
>> I don’t know of a programmatic way to check whether nanosecond precision is supported. Do you happen to know the file system type (and os.name property)? Right now the test has this exclusion
>>   53         if (System.getProperty("os.name").toLowerCase().startsWith("mac") &&
>>   54             store.name().equalsIgnoreCase("hfs")) {
>>   55             System.err.println
>>   56                 ("HFS on macOS does not have nsec timestamps: skipping test");
>>   57             return;
>>   58         }
>> Thanks,
>> 
>> Brian
>> 
>>> On Aug 8, 2019, at 7:56 AM, Langer, Christoph <christoph.langer at sap.com <mailto:christoph.langer at sap.com>> wrote:
>>> 
>>> it seems that we‘re encountering this behavior on old Suse SLES 11.3 and 11.4 machines, both ppc64 and linuxx86_64.
>>>  
>>> Maybe these can somehow be excluded from the test? Or the test can detect upfront when nanotime isn’t supported?
>> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.java.net/pipermail/nio-dev/attachments/20190826/62d8546a/attachment.html>