Possible bug in jextract C header definition?

[Mark Hammons]

Hi Maurizio,

Unfortunately, it will take a little more work on my side to test the 
new panama ea.

Apparently jdk-14 does not play well with sbt.

~Mark

On 8/9/19 3:06 PM, Mark Hammons wrote:
> I'll give it a shot tonight or tomorrow. Thanks Maurizio!
>
> On 08/08/2019 23:13, Maurizio Cimadamore wrote:
>> New builds just got out of the oven :-)
>>
>> https://jdk.java.net/panama/
>>
>> Maurizio
>>
>> On 06/08/2019 19:21, Mark Hammons wrote:
>>> Great, glad to hear it, and I’m looking forward to testing the new 
>>> builds.
>>>
>>> ~Mark
>>>
>>>> On 6 Aug 2019, at 17:55, Maurizio Cimadamore 
>>>> <maurizio.cimadamore at oracle.com> wrote:
>>>>
>>>> Mark,
>>>> we suspect that both issues you are seeing could be related to some 
>>>> of the fixes that went in *after* the snapshots were last generated.
>>>>
>>>> We will do some stabilization work on panama/foreign and then 
>>>> trigger a new binary snapshot. I'm confident this will resolve your 
>>>> issues.
>>>>
>>>> Maurizio
>>>>
>>>>> On 06/08/2019 15:43, Mark Hammons wrote:
>>>>> Hi Maurizio
>>>>>
>>>>> If you're talking about the inner class issue with dotty's 
>>>>> compiler, it appears to be a small issue with the inner class 
>>>>> generation of jextract that dotty is not happy with, but the scala 
>>>>> compiler is. I am not certain that the problem lies with jextract, 
>>>>> but so far my attempts to debug it with the dotty team have 
>>>>> stalled (they appear to be busy working on other things) and I 
>>>>> have received no response from them.
>>>>>
>>>>> I would love to test the issues with the latest versions of 
>>>>> panama, and the one above was the very first one I tested 
>>>>> recently. Right now my approach to getting jextract working with 
>>>>> dotty is to hand generate bindings to the foreign api rather than 
>>>>> use jextract.
>>>>>
>>>>> ~Mark
>>>>>
>>>>>> On 06/08/2019 16:25, Maurizio Cimadamore wrote:
>>>>>> Hi Mark - that version is quite old (we should refresh it soon), 
>>>>>> and the fix for __va_list is not contained in that build.
>>>>>>
>>>>>> We would appreciate if you could help narrowing down the other 
>>>>>> issue you have reported, to make sure that that is also fixed in 
>>>>>> the latest sources. Could you help with that? More specifically, 
>>>>>> the errors you mentioned do not seem like jextract errors - 
>>>>>> what's going on?
>>>>>>
>>>>>> Maurizio
>>>>>>
>>>>>>> On 06/08/2019 15:09, Mark Hammons wrote:
>>>>>>> Hi,
>>>>>>>
>>>>>>> I do apparently have the latest version (ea-70), and it's being 
>>>>>>> mapped to __va_list_tag instead for some reason.
>>>>>>>
>>>>>>> ~Mark
>>>>>>>
>>>>>>>
>