On Fri, 22 Jan 2021 07:04:47 GMT, Lin Zang <lz...@openjdk.org> wrote:
>> src/jdk.hotspot.agent/share/classes/sun/jvm/hotspot/utilities/HeapHprofBinWriter.java
>> line 475:
>>
>>> 473: if (!useSegmentedHeapDump) {
>>> 474: // Fill in final length
>>> 475: fillInHeapRecordLength();
>>
>> It's unclear to me why this code is now conditional on not using a segmented
>> heap dump.
>
> Hi Chris,
> Thanks for review. The original implementation of heap dump (when not using
> segments) is to treat all HPROF_HEAP_DUMP section as a whole segment and fill
> the size after the data are ready to be flush.
> As specified at link:
> http://hg.openjdk.java.net/jdk6/jdk6/jdk/raw-file/tip/src/share/demo/jvmti/hprof/manual.html#mozTocId848088
> .
> so here if it is not segmented heap dump, it still need call
> fillInHeapRecordLength() to fill the size slot before data flush.
ok
-------------
PR: https://git.openjdk.java.net/jdk/pull/1712
