On Saturday, 30 June 2018 at 08:51:56 UTC, Joakim wrote:
> On Saturday, 30 June 2018 at 08:27:30 UTC, 鲜卑拓跋枫 wrote:
>> On Friday, 29 June 2018 at 14:52:45 UTC, Joakim wrote:
>>> On Friday, 29 June 2018 at 12:13:09 UTC, 鲜卑拓跋枫 wrote:
>>>> [...]
>>>
>>> So do people in US and Europe, the vast majority of whom watching the livestream or online videos didn't attend DConf.
>>>
>>> On Friday, 29 June 2018 at 12:30:49 UTC, Mike Parker wrote:
>>>> [...]
>>>
>>> First off, I question there's much benefit to even the key devs beyond communicating through email and video conferencing to iron things out, as Andrei indicates he does with Walter.
>>>
>>> And Jonathan only mentioned the key devs, so that does exclude. As for everybody else, see below.
>>>
>>>> [...]
>>>
>>> Then spend all your time doing those things: why waste the majority of conference time sitting through talks that you don't bother defending?
>>>
>>> Here's what a "conference" in Asia or Europe or wherever should probably look like in this day and age:
>>>
>>> - Have most talks prerecorded by the speaker on their webcam or smartphone, which produce excellent video these days with not much fiddling, and have a couple organizers work with them to get those home-brewed videos up to a certain quality level, both in content and presentation, before posting them online.
>>>
>>> - Once the videos are all up, set up weekend meetups in several cities in the region, such as Tokyo, Hong Kong, and Bangalore, where a few livestreamed talks may talk place if some speakers don't want to spend more time producing a pre-recorded talk, but most time is spent like the hackathon, discussing various existing issues from bugzilla in smaller groups or brainstorming ideas, designs, and libraries for the future.
>>>
>>> This is just off the top of my head; I'm sure I'm missing some small details here and there, as I was coming up with parts of this as I wrote it, but I estimate it'd be an order of magnitude more productive than the current conference format while being vastly cheaper in total cost to all involved. Since D is not exactly drowning in money, it makes no sense to waste it on the antiquated conference format. Some American D devs may complain that they no longer essentially get to go on a vacation to Berlin or Munich- a paid vacation if their company compensates for such tech conferences- but that's not our problem.
>>
>> Thanks for further clarification.
>> But there is still some limitation may exist, e.g., as you may note that
>> the latest Linaro Connect that held in Hong Kong add a new special "China Access" for sharing their conference resources like below:
>> http://connect.linaro.org/hkg18/resources/#1506759202543-a2113613-2111
>>
>> I noted it because I am very interested in programming on ARM, so I hope LDC
>> (https://github.com/ldc-developers/ldc) could add the support for AARCH64 as soon as possible:).
>
> Check out the ltsmaster branch of LDC from git and try it out, most tests passed for me on Ubuntu/AArch64 16.04:
>
> https://github.com/ldc-developers/ldc/issues/2153#issuecomment-384264048
>
> The few remaining exceptions are some math-related modules would need to be patched to support 128-bit floating-point real numbers, such as CustomFloat from std.numeric, std.internal.math.gammafunction, or the floating-point parser from std.conv (but only if you really need that extra precision, most of that code still works at 80-bit accuracy), though all the tests from std.math now pass. The other big issue is core.stdc.stdarg needs to be adapted for AArch64 varargs, which is what's holding back building the latest LDC 1.10 natively.

Good News!
Hope official AArch64 support will be included in their upcoming releases.