[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openjdk-serviceability-dev
Subject:    Re: RFR JDK-8209605: com/sun/jdi/BreakpointWithFullGC.java fails with ZGC
From:       David Holmes <david.holmes () oracle ! com>
Date:       2018-08-22 2:29:53
Message-ID: ca401bd6-609d-12f6-eea0-c908c6a91cf8 () oracle ! com
[Download RAW message or body]

Hi Alex,

On 22/08/2018 5:36 AM, Alex Menkov wrote:
> Hi David,
> 
> Updated fix:
> http://cr.openjdk.java.net/~amenkov/sh2java/step1_zgc/webrev.01
> 
> I used "catch (Throwable)" instead suggested "catch (Error | 
> RuntimeException ex)" (we want to catch anything can be thrown)

Interesting that it worked. Throwable is a checked-exception so you 
normally can't throw a Throwable unless you declare "throws Throwable". 
But it seems our exception processing logic is getting smarter and can 
see that the only Throwable's that can be caught are in fact the 
unchecked Error and RuntimeException (which is why I suggested them) as 
no checked-exceptions are possible from the code that is called.

> Also I updated String.split to use \\R for the line-separators as you 
> suggested in review for
> JDK-8209604: [TEST] rewrite com/sun/jdi shell tests to java version - step2

Looks good! Tkanks.

David


> --alex
> 
> On 08/20/2018 18:29, David Holmes wrote:
>> Hi Alex,
>>
>> On 21/08/2018 8:24 AM, Alex Menkov wrote:
>>> Hi all,
>>>
>>> please review a fix for
>>> https://bugs.openjdk.java.net/browse/JDK-8209605
>>> webrev:
>>> http://cr.openjdk.java.net/~amenkov/sh2java/step1_zgc/webrev/
>>>
>>> The fix:
>>> - updates the fix so it works as original shell test:
>>>      - drops "-Xmx32m" argument for debuggee (for shell test default 
>>> jtreg -Xmx512m was effective);
>>
>> Okay. I don't think you need the historical commentary though - the 
>> information in the bug report is sufficient.
>>
>>>      - drops requirement for "Pause Full (System.gc())" (different GCs 
>>> use different logging, original test does not require it);
>>
>> Okay .
>>
>>> - fixes "ArrayIndexOutOfBoundsException: Index 0 out of bounds for 
>>> length 0". The cause of the issue is String.split(sep), which can 
>>> returns zero-length array;
>>
>> Okay.
>>
>>> - improve error handling - debuggee must be terminated if jdb 
>>> execution fails.
>>
>> For this part I suggest:
>>
>> catch (Error | RuntimeException ex)
>>
>> so that, for example, OOME also triggers debuggee shutdown.
>>
>> Thanks,
>> David
>>
>>>
>>> --alex
[prev in list] [next in list] [prev in thread] [next in thread]