Previously, we added these rpaths to lime.ndll when it was built in commits c70ec9f and 333d093, but it's actually necessary only for Neko, so now I made it happen specfically after calling `nekotools boot` to create the Neko executable.
I've tested cpp and hl, and I've confirmed that the executables still launch successfully when these rpaths are omitted. It's better for their security to use fewer rpaths.
As noted commit c70ec9f, adding these rpaths is necessary due to a change in Xcode 15 where /usr/local/lib used to be available on the rpath automatically, but now it isn't, which affects the executable's ability to find the libneko dylib.
It turns out that adb isn't included GitHub Actions unless you install it specifically. However, you can technically build Android apps with Lime without adb, so we shouldn't report the error in initialize(). Most important is install(), where both adb and emulator are commonly used.
Followup to eed47e7132
The 'SDK Tools' package located in /tools/ is officially considered 'obsolete', so we should not use it by default.
We now prefer /platform-tools/adb over /tools/adb
We now prefer /emulator/emulator over /tools/emulator
If the newer replacement executables are missing, we still try to fall back to /tools/. This should allow older Android SDKs to continue to work properly.
Additionally, if neither version can be found, we report an error. For adb, we always need it, so we always report an error if it is missing. For emulator, we report an error only if we're actually going to use an emulator.
/tools/android doesn't have a newer alternative. We were running 'android list avds' to get a list of all available AVDs. However, both '/emulator/emulator -list-avds' and '/tools/emulator -list-avds' provide a simple list of AVDs separated by line breaks. So it seems that we never actually needed /tools/android. Plus, it outputs a better format that doesn't require searching every line of the output string for 'Name:', and we can just split and trim. So I completely removed /tools/android and we now use either '/emulator/emulator -list-avds' or '/tools/emulator -list-avds'.
Previously, null was propagated, which could lead to confusing errors from arithmetic operations. Now, it is ignored, and a warning is displayed that says where the invalid value can be found.
This is its only use case, so there's no need for an intermediate step of getting the library root. `haxelib path` does return the NDLL folder; the problem with the old implementation of `__findHaxelib()` was it got Lime's source folder instead.
We could fall back to `haxelib path` for very old versions of Haxelib, but most likely even people stuck on Haxe 3 have updated Haxelib.
Also, `haxelib path` gave incorrect results in the first place, so falling back wouldn't do any good.
This is where arm64 homebrew installs `libneko.dylib`. However, it may still be in `/usr/local/lib` if installed via the haxe .pkg installer, so we add both.
On ios, we build statically linked executables. This means that if lime uses mbedtls 3, hxcpp is also forced to use it which can cause crashes and other problems, as hxcpp is currently written for 2.28
It was hardcoding VS 2017, but now uses vswhere.exe to find newer Visual Studio versions too.
Newer versions of VS don't actually support UWP JavaScript, though. However, if users have UWP installed in VS, the error message will be clearer about that lack of support.
The `Asset` constructor sets `asset.id` to the asset's path, a sensible default. Of course, if the asset has a name instead of a path, we have to use that instead. Or if the user specified an ID, that's even better.
We just don't want to default to `""`, since then a bunch of assets could have the same ID.
Resolves#1758.
Originally, we forced compilation on x86, presumably because at the time HashLink lacked good 64-bit support. When this support improved, the line was changed to force x64 compilation rather than being removed, which may have been a mistake. Now that there are even more valid architectures, it just doesn't make sense to force one.
The old logic could produce inappropriate results, such as attempting to compile an x86 binary on ARM64, or compiling no binary when "64" is specified on a 32-bit machine. I'd argue that it makes sense to only check the flags when they're supported, and not to bother otherwise.
