tl;dr - I'm after a status report, and this thing pops up on forums often enough that I want the information in one place.

Aight guys. So here's the deal. That program I showed off at DConf this year? I have a potential customer interested in beginning usage in August.

They develop for mobile platforms. This means I'll need mobile support.

D is uniquely situated for my needs on all platforms going forward (you know, once the garbage collector is something we can talk about in terms of pride and not like the misbehaving child that won't grow up). That, of course, means it needs to compile and run on all platforms. Which realistically means code generation for two architectures these days (x64, ARM) and runtime support for non-desktop platforms (iOS, Android, consoles - although that last one is something almost everyone here can't work on).

Now, being a middleware, I'll be expecting the engine integration to supply routines like file operations and memory operations. Standard stuff.

I can probably totally avoid Phobos and avoid linking that behemoth in the runtime.

So that leaves code gen and druntime being usable.

I'll need to know before I work on my runtime library if these are usable. Otherwise, I'm going to have to write the runtime in C++ and write more code generation definitions to export the work I've already done to C++.

Do these two things work currently?

On Thursday, 25 July 2019 at 13:21:01 UTC, Ethan wrote:
> tl;dr - I'm after a status report, and this thing pops up on forums often enough that I want the information in one place.
>
> Aight guys. So here's the deal. That program I showed off at DConf this year? I have a potential customer interested in beginning usage in August.
>
> They develop for mobile platforms. This means I'll need mobile support.
>
> D is uniquely situated for my needs on all platforms going forward (you know, once the garbage collector is something we can talk about in terms of pride and not like the misbehaving child that won't grow up). That, of course, means it needs to compile and run on all platforms. Which realistically means code generation for two architectures these days (x64, ARM) and runtime support for non-desktop platforms (iOS, Android, consoles - although that last one is something almost everyone here can't work on).
>
> Now, being a middleware, I'll be expecting the engine integration to supply routines like file operations and memory operations. Standard stuff.
>
> I can probably totally avoid Phobos and avoid linking that behemoth in the runtime.
>
> So that leaves code gen and druntime being usable.
>
> I'll need to know before I work on my runtime library if these are usable. Otherwise, I'm going to have to write the runtime in C++ and write more code generation definitions to export the work I've already done to C++.
>
> Do these two things work currently?

LDC for android is working with most of druntime & phobos.   you can install ldc 1.13.0 from termux,  I am also be able to build 1.16 with some phobos & druntime small bugs.

IOS https://github.com/smolt/ldc-iphone-dev/releases only have 0.17.0-alpha1 (2.068.2 FE) LLVM 3.6.2.

If LDC core team can only cherry-pick target/march  for IOS, then I think people will be easy to  try the ABI them self.

On Thursday, 25 July 2019 at 13:49:39 UTC, Newbie2019 wrote:
> IOS https://github.com/smolt/ldc-iphone-dev/releases only have 0.17.0-alpha1 (2.068.2 FE) LLVM 3.6.2.

Sweet zombie Jesus that's dire.

Mobile game developers make the lion's share of their income from iOS.

I can deal with development builds of a compiler. But that needs to be in a significantly better state for people to even consider D on mobile a serious thing.

And, you know, in general, there's more mobile devices on the web than desktop devices.

On Thursday, 25 July 2019 at 14:08:20 UTC, Ethan wrote:
> Sweet zombie Jesus that's dire.
>
> Mobile game developers make the lion's share of their income from iOS.
>
> I can deal with development builds of a compiler. But that needs to be in a significantly better state for people to even consider D on mobile a serious thing.
>
> And, you know, in general, there's more mobile devices on the web than desktop devices.


I believe mobile hardware is get more and more powerful and necessary for everybody.  The next year Arm A77 and Apple A13  will fast then laptop.

This is a huge market and the largest number of user groups ever.   D core team should take mobile serious and provide basic support.

The current status is limit support from LDC team,  and  about zero man hours investment into IOS for so many years(last release since 2015).

Android in a better status, but we loss the key developer last year.


Consider there is consider there is about five billion people use mobile every more than 2 hours every day, but  D language developer  investment zero man hours into IOS for so many years.

I can't image how many opportunity is passed for D.

On Thursday, 25 July 2019 at 17:07:03 UTC, Newbie2019 wrote:
> On Thursday, 25 July 2019 at 14:08:20 UTC, Ethan wrote:
>
> This is a huge market and the largest number of user groups ever.
>   D core team should take mobile serious and provide basic support.
>
> Consider there is consider there is about five billion people use mobile every more than 2 hours every day, but  D language developer  investment zero man hours into IOS for so many years.
>
> I can't image how many opportunity is passed for D.

imo mobile devs is splitted by Swift and Java/Kotlin as natives and Unity for games.
it will be impossible to change something in this area.
even Go/Rust and others are relaxed about it.

On Thursday, 25 July 2019 at 17:11:56 UTC, a11e99z wrote:
> On Thursday, 25 July 2019 at 17:07:03 UTC, Newbie2019 wrote:
>> On Thursday, 25 July 2019 at 14:08:20 UTC, Ethan wrote:
>
> imo mobile devs is splitted by Swift and Java/Kotlin as natives and Unity for games.
> it will be impossible to change something in this area.
> even Go/Rust and others are relaxed about it.

more interesting is WASM cuz PWA & SPA exists that can be build with WASM that will work on mobiles too

On Thursday, 25 July 2019 at 17:11:56 UTC, a11e99z wrote:
> imo mobile devs is splitted by Swift and Java/Kotlin as natives

And that's great. But I'm not writing high-performance easy-to-maintain multi-platform code in those languages.

My software will plug in to Unity and Unreal Engine. If D can't target iOS (moreso than Android) then I straight up can't use D for the runtime.

And if I can't target iOS, then there's a very good chance I'm not going to use D for mobile gaming ever. Which in this day in age is one step removed from just straight up not using it in gaming.

On Thursday, 25 July 2019 at 17:07:03 UTC, Newbie2019 wrote:
>   D core team should take mobile serious and provide basic support.

I'm not sure what this statement means? Should they put a team of their paid developers on it? That's simply not how this project works...

> I can't image how many opportunity is passed for D.

If there's business interest because that's a lucrative market, those who profit from it need to fund the development. It's hard to see how it's an "opportunity" if you're donating your free time so companies can make money.

On Thursday, 25 July 2019 at 18:52:16 UTC, bachmeier wrote:

> If there's business interest because that's a lucrative market, those who profit from it need to fund the development. It's hard to see how it's an "opportunity" if you're donating your free time so companies can make money.


What about the people they are just want to enjoy the mobile develop with D ?  Why every one has to care about business interest ?

Look at the Github trending, most people contribute without business interest. And D trending is empty almost all the time.

On Thursday, 25 July 2019 at 14:08:20 UTC, Ethan wrote:
> On Thursday, 25 July 2019 at 13:49:39 UTC, Newbie2019 wrote:
>> IOS https://github.com/smolt/ldc-iphone-dev/releases only have 0.17.0-alpha1 (2.068.2 FE) LLVM 3.6.2.
>
> Sweet zombie Jesus that's dire.
>
> Mobile game developers make the lion's share of their income from iOS.
>
> I can deal with development builds of a compiler. But that needs to be in a significantly better state for people to even consider D on mobile a serious thing.
>
> And, you know, in general, there's more mobile devices on the web than desktop devices.

As someone also interested in iOS in LDC, maybe we can pool money to make it happen?

There are other companies interested in iOS target in LDC.