RFR (S): 8237777 "Dumping core ..." is shown despite claiming that "# No core dump will be written."
David Holmes
david.holmes at oracle.com
Sat Apr 25 04:44:45 UTC 2020
On 25/04/2020 4:02 am, gerard ziemski wrote:
>
>
> On 4/23/20 7:55 PM, David Holmes wrote:
>
> [trimmed]
>
>> Given that the "Dumping core ..." is only shown in non-product builds
>> I think we are way over-thinking/engineering this. We have:
>>
>> if (dump_core) {
>> ...
>> #ifndef PRODUCT
>> ...
>> out.print_raw_cr("Dumping core ...");
>> #endif
>> ::abort(); // dump core
>> }
>>
>> so lets just change one line:
>>
>> out.print_raw_cr("Requesting a core dump if possible ...");
>>
>> otherwise you need one boolean to control whether we call abort vs
>> exit, and a second boolean to control what we print.
>>
>> Or change that one line to:
>>
>> out.print_raw_cr("Calling ::abort() ...");
>
> Making sure we do not mess up the core creation logic is paramount, so I
> like your suggestion to keep it simple, but perhaps we can take it one
> small step further and just not bother with that message at all?
>
> Maybe we should simply remove the line:
>
> out.print_raw_cr("Dumping core ...");
>
> completely? After all, we print lots of core dumping related info anyhow.
We've been printing
Current thread is xxx
Dumping core ...
to the default output stream in non-product builds since day one of this
code - literally. Why only non-product and why only for when dump_core
is true I have no idea. The whole thing seems somewhat superfluous. I'd
suggest either tweaking the second line, or removing it all, rather than
just deleting the second line (which gives some additional context to
the first).
Cheers,
David
>
> cheers
>
More information about the hotspot-runtime-dev
mailing list