I'm working on a WebApi project written in F#. Here a snippet:
module MyModule
open System
let MyGuid = Guid.Parse "934F0B12-D00A-491D-862D-EE745EF3C560"
let myFunction list =
list.Get(MyGuid) // --> here MyGuid has the TypeInitializationException before list.Get is called
By debugging I can see that the MyGuid actually has an error
Changing the code followings, it works:
module MyModule
open System
let MyGuid () = Guid.Parse "934F0B12-D00A-491D-862D-EE745EF3C560"
let myFunction list =
list.Get(MyGuid())
I actually know the MyGuid of the first example is a variable and the second one a function definition, but why does the first rise the exception? I my code MyGuid is used some times. so in the first example I'd have only one instance, in the second a new instance every time MyGuid is called...
I'm not 100% sure that this is the problem here, but I've seen similar behaviour when using unit test runners sometimes. My guess is that the error happens because the top-level MyGuid variable is not initialized correctly and has the default zero value (and as a result, the lookup fails).
The way global variables are initialized in F# is tricky - if you compile code as executable, this can happen from the Main method. But if you compile code as a library, the compiler inserts an initialization checks into static constructors of the types in your library (to make sure everything is initialized before you access anything).
I think this can break if you compile your code as an executable, but then load it as a library - the entry-point is not called and so the variables are not initialized. I'm not sure how exactly WebApi loads libraries, but this could be a problem - especially if you compile the F# code as an executable.
Your workaround of turning the global variable into a function fixes this, because the function is compiled as a method and so you avoid referring to an uninitialized value. Sadly, I don't think there is a better workaround for this.
Related
I'm trying to deploy an app to production and getting a little confused by environment and application variables and what is happening at compile time vs runtime.
In my app, I have a genserver process that requires a token to operate. So I use config/releases.exs to set the token variable at runtime:
# config/releases.exs
import Config
config :my_app, :my_token, System.fetch_env!("MY_TOKEN")
Then I have a bit of code that looks a bit like this:
defmodule MyApp.SomeService do
use SomeBehaviour, token: Application.get_env(:my_app, :my_token),
other_config: :stuff
...
end
In production the genserver process (which does some http stuff) gives me 403 errors suggesting the token isn't there. So can I clarify, is the use keyword getting evaluated at compile time (in which case the application environment doest exist yet)?
If so, what is the correct way of getting runtime environment variables in to a service like this. Is it more correct to define the config in application.ex when starting the process? eg
children = [
{MyApp.SomeService, [
token: Application.get_env(:my_app, :my_token),
other_config: :stuff
]}
...
]
Supervisor.start_link(children, opts)
I may have answered my own questions here, but would be helpful to get someone who knows what they're doing confirm and point me in the right way. Thanks
elixir has two stages: compilation and runtime, both written in Elixir itself. To clearly understand what happens when one should figure out, that everything is macro and Elixir, during compilation stage, expands these macros until everything is expanded. That AST comes to runtime.
In your example, use SomeBehaviour, foo: :bar is implicitly calling SomeBehaviour.__using__/1 macro. To expand the AST, it requires the argument (keyword list) to be expanded as well. Hence, Application.get_env(:my_app, :my_token) call happens in compile time.
There are many possibilities to move it to runtime. If you are the owner of SomeBehaviour, make it accept the pair {:my_app, :my_token} and call Application.get_env/2 somewhere from inside it.
Or, as you suggested, pass it as a parameter to children; this code belongs to function body, meaning it won’t be attempted to expand during compilation stage, but would rather be passed as AST to the resulting BEAM to be executed in runtime.
I have a Clojurescript program running in the browser.
It imports a number of libraries, and then I want to allow the user to enter some small clojurescript "glue-code" that calls those libraries.
I can see (from https://cljs.github.io/api/cljs.js/eval) that you call eval with four arguments, the first being the state of the environment, which is an atom. But can I actually turn my current environment with all the functions I've required from elsewhere, into an appropriate argument to eval?
Update :
I thought that maybe I could set the namesspace for the eval using the :ns option of the third, opts-map, argument. I set it to the namespace of my application :
:ns "fig-pat.core"
But no difference.
Looking at the console, it's definitely the case that it's trying to do the evaluation, but it's complaining that names referenced in the eval-ed code are NOT recognised :
WARNING: Use of undeclared Var /square
for example. (square is a function I'm requiring. It's visible in the application itself ie. the fig-pat.core namespace)
I then get :
SyntaxError: expected expression, got '.'[Learn More]
Which I'm assuming this the failure of eval-ed expression as a whole.
Update 2 :
I'm guessing this problem might actually be related to : How can I get the Clojurescript namespace I am in from within a clojurescript program?
(println *ns*)
is just printing nil. So maybe Clojurescript can't see its own namespace.
And therefore the :ns in eval doesn't work?
Calling eval inside a clojurescript program is part of what is called "self-hosted clojurescript".
In self-hosted clojurescript, namespaces are not available unless you implement a resolve policy. It means that have to let the browser know how to resolve the namespace e.g. loads a cljs file from a cdn.
It's not so trivial to implement namespace resolving properly.
This is explained in a cryptic way in the docstring of load-fn from cljs.js namespace.
Several tools support namespaces resolving in self-host cljs running in the browser e.g Klipse and crepl
I'm trying to use Graphics32 package. Graphics32 was compiled and installed without any issue.
When I try to execute (debug) following code under C++ Builder XE3
TBitmap32* bmp = new TBitmap32();
bmp->LoadFromFile("d:\\sample.bmp");//This calls SaveToStream instead of LoadFromFile
...
it calls another member function SaveToStream which I can trace into and step while debugging until AV rises.
I have never encountered such behavior before.
Is there any compiler directive I'm missing or some workaround to make proper function call?
Update: I use the Graphics32 source from SVN. Everything works good if I use code prior to revision 2122.
According to documentation (https://developers.google.com/closure/compiler/docs/api-tutorial3#externs), it seems the closure compiler should rename variables when no external declaration exists, including when using functions/variables from an external bit of code. The example they give is
function makeNoteDom(noteTitle, noteContent, noteContainer) {
// Create DOM structure to represent the note.
var headerElement = textDiv(noteTitle);
var contentElement = textDiv(noteContent);
...
}
where the textDiv function is declared in the global scope by a third-party lib of some sort. It says textDiv should be declared external to prevent renaming.
My question is - when I put this code or similar into the Closure Compiler without any extern declarations why is textDiv not renamed (which would break the code), as the documentation indicates?
The compiler assumes that calls to an undefined function are in fact calls to an external functions. Using the command line compiler, you can use --warning_level VERBOSE to have the compiler treat this condition as an error.
The Web Application is primarily built for demos and assumes this by default. While you can set a VERBOSE warning level, it will not change this functionality. See the Additional Web Service Options page for information on options. I've filed a bug report about this.
Due to the renaming algorithm for properties, undeclared properties will be renamed in a breaking way if that same property name isn't declared on an object in externs.
Say I have below lines any in a Java class,
System.out.println("start");
if(true)//The compiler will give compile time error if I uncomment this.
throw new RuntimeException();
System.out.println("end");
The unreachable code error message will appear if the if(true) is commented. Why Don't the compiler know that the line under if(true) will always be executed?
Is the Java compiler is designed to work like this or is it a limitation?
It's deliberate part of the design around code reachability. See section 14.21 of the JLS which has a section at the bottom about this.
The rationale for this differing treatment is to allow programmers to define "flag variables" such as:
static final boolean DEBUG = false;
and then write code such as:
if (DEBUG) { x=3; }
The idea is that it should be possible to change the value of DEBUG from false to true or from true to false and then compile the code correctly with no other changes to the program text.
It is a limitation of how far you go to check you have dead code or code which will never run.
The JVM can detect the code is dead and not compile it.