[foreign-abi] RFR 8227718: Add support for SystemABI

[Maurizio Cimadamore]

On 19/07/2019 10:49, Jorn Vernee wrote:
> Hi,
>
> I took a look at this earlier already (I had an older copy of the 
> branch, so patch applied cleanly :) ), but the new version also passes 
> all the tests on my machine.
>
> As you said, there are some high-level issues, like varargs functions 
> not working on Windows (at least not if you want to pass float 
> arguments), but also other things like the current 
> MemorySegmentImpl.everything() not working for pointers outside of the 
> range of [0, Long.MAX_VALUE). It also seems like a good time to 
> re-examine whether we want to keep the default library.
Yep - these are all valid points; I have internally a list of the things 
that need _more work_, I'll polish it up and send it up here. To be 
clear: I'm in no way proposing that this patch is 100% of the story - 
more 20% (well, maybe a bit more :-), but you get the idea). But I'm 
also confident that we will find a way to make that work.
>
> But, overall this looks good as a first version. The only minor nit I 
> have is to maybe break up the MemoryLayout::ofAddress implementations 
> across multiple lines.

Sure - I'll fix that and push

Thanks
Maurizio


>
> Jorn
>
> [1] : https://openjdk.java.net/jeps/357
>
> On 2019-07-19 08:05, Nick Gasson wrote:
>> Hi Maurizio,
>>
>>> Here's a revised webrev:
>>>
>>> http://cr.openjdk.java.net/~mcimadamore/panama/8227718_v3/
>>>
>>
>> I tested the new webrev on an Arm system, all the jdk_foreign tests 
>> pass!
>>
>> Thanks,
>> Nick