RFR: 8205199: more Linux clang compile failures

[David Holmes]

On 19/06/2018 1:00 PM, Martin Buchholz wrote:
> On Mon, Jun 18, 2018 at 7:15 PM, Kim Barrett <kim.barrett at oracle.com> wrote:
> 
>>> On Jun 18, 2018, at 9:35 PM, Martin Buchholz <martinrb at google.com>
>> wrote:
>>>
>>> Hej Gerard, here's a code review for you:
>>>
>>> 8205199: more Linux clang compile failures
>>> http://cr.openjdk.java.net/~martin/webrevs/jdk/more-linux-
>> clang-failures/
>>> https://bugs.openjdk.java.net/browse/JDK-8205199
>>
>> -----
>>   717 static bool store_context(const void* context) {
>>   718   if (memcpy(&g_stored_assertion_context, context,
>> sizeof(ucontext_t)) == NULL) {
>>
>> I don’t know what this code is supposed to be doing,  but memcpy isn’t
>> going to return
>> NULL here, any more than it would return false (as in the pre-change code).
>>
>>
> Hmmm... good point.  I guess I was so preoccupied with getting the code to
> compile I didn't stop to think if I should.
> 
> Thomas, I think this is your code from
> 
> changeset:   49653:a569cb4425f3
> user:        stuefe
> date:        2018-03-30 06:31 +0200
> 8191101: Show register content in hs-err file on assert
> Reviewed-by: adinn, clanger, simonis
> 
> Can you fix, or suggest a fix, or we can move this to an independent change.

AFAICS the memcpy can't fail and the dest is not NULL so the "if" should 
just disappear - no?

David

> 
>> ——
>>
>> The change to os_linux_x86.cpp looks good.
>>
>>