RFR (XS): 8236177: assert(status == 0) failed: error ETIMEDOUT(60), cond_wait

[gerard ziemski]

Thank you David for the review.

On 5/3/20 4:47 AM, David Holmes wrote:
> On 3/05/2020 1:02 pm, Kim Barrett wrote:
>>> On May 1, 2020, at 4:59 PM, gerard ziemski 
>>> <gerard.ziemski at oracle.com> wrote:
>>> In the fix I introduce MAC_ONLY() macro, as per David's suggestion.
>>>
>>> bug link at https://bugs.openjdk.java.net/browse/JDK-8236177
>>> open webrev at http://cr.openjdk.java.net/~gziemski/8236177_rev3
>>> testing: Mach hs-tier1,2,3,4,5 in progress…
>>
>> I think the macro name should be MACOS_ONLY, as David suggested.
>
> Yes please.

I actually gave it some thought when I was deciding on the name of the 
macro. I chose MAC_ONLY, based on the already existing pattern, i.e.:

#ifdef LINUX
#define LINUX_ONLY(code) code

#ifdef _WINDOWS
#define WINDOWS_ONLY(code) code

#if defined(__FreeBSD__) || defined(__NetBSD__) || defined(__OpenBSD__) 
|| defined(__APPLE__)
#define BSD_ONLY(code) code

#ifdef SOLARIS
#define SOLARIS_ONLY(code) code

To note here: none of the existing platform macros include "OS".

In case of Mac OS X, I offered:

#ifdef __APPLE__
#define MAC_ONLY(code) code

but the choices, as I see them (listed in the order I personally prefer 
them), are:

#1 MAC_ONLY
#2 MACOSX_ONLY
#3 MAC_OSX_ONLY
#4 MACOS_ONLY
#5 APPLE_ONLY

I personally went with simply MAC_ONLY (again based on the existing 
pattern), but if I had to go with something else, then I'd rather use 
MACOSX_ONLY over MACOS_ONLY, since we have Mac OS X (which is the 
current designation, ex. Mac OS X 10.0) and not Mac OS (which is 
reserved for the old classic versions, ex. Mac OS 7).

What do you think?


cheers