On Friday, 14 November 2014 at 18:35:56 UTC, Joakim wrote:
On Friday, 14 November 2014 at 14:20:00 UTC, Chad Joan wrote:
We really should have had a toolchain compiling D into working
(release quality) Android/iOS executables about 3-5 years ago.
I agree. Well, to a large extent: the first Android smartphone
was only released 6 years ago, would've been a bit early to
jump in the year after. I picked it up last year because
nobody was working on it, after years of people asking.
This would have allowed D to scoop up a HUGE share of
deployment in a market that was very new and welcoming of
experimentation.
I don't know, I think D really has a chance to do well on
mobile, but have other alternative languages, ie not
C/C++/java, done _that_ well on Android? I know the Rust guys
are now working on it, but while languages like FreePascal are
on there, I have not heard of any big success story.
...
Native development in Android is a world of pain, given Android's
team stance on the NDK and poor tooling support.
You just have a tiny subset of APIs, the ones important for game
development, i.e. OpenGL, in memory framebuffers, sound, hardware
sensors and partial POSIX.
For everything else either you go JNI or have to bring third
party dependencies that increase the apk size.
It is ridiculous to the point there isn't access to the SQLLite,
SKIA and other C and C++ libraries from Android. One is forced to
integrate an own version.
Or Android Studio that still doesn't have a public roadmap about
C and C++ support parity to Eclipse ADT/CDT.
So NDK is only worthwhile for gaming or business code with the UI
written in Java.
Xamarin and Phonegap are the only success stories in the
traditional app market.
One has teams of developers creating platform wrappers and the
other just uses the browser via the web view, thus easier
platform integration.
I was playing around with Qt and gave up, as only the upcoming
5.4 will have UI controls with support for Android.
--
Paulo
