I searched the web for how to enforce srictfp in Scala but could not find any hint of it. There are some people complaining about it, but real solutions cannot be found. There is a bugtracker entry about it which is almost two years old. As it seems there is no elegant fix for it on the way I'm looking for workarounds.
My current idea is to set the appropiate method flag ACC_STRICT in the generated bytecode by myself somehow but I have no idea what would be the best solution to do so. A Scala Compiler Plugin comes to mind or just hacking flags in a hex editor. Maybe someone faced the same challenge and can tell me his or her solution?
You could add a post-processor in your build process that would add the strictfp modifier to the generated class (i.e. setting the ACC_STRICT flag as you say).
You can implement such a post-processor using Javassist for example. This could look like this:
CtClass clazz = ClassPool.getDefault().makeClass(
new FileInputStream("old/HelloWorld.class"));
CtMethod method = clazz.getDeclaredMethod("testMethod");
method.setModifiers(method.getModifiers() | Modifier.STRICT);
clazz.detach();
clazz.toBytecode(new DataOutputStream(new FileOutputStream(
"new/HelloWorld.class")));
You would then have to find a way to configure which classes/method need to be modified this way.
Scala has a strictfp annotation now:
#strictfp
def f() = …
Related
Cache2k looks like a very promising caching implementation. Unfortunately the documentation is very limited, which is why I need some help with the following issue. I am using the latest version 0.26-BETA.
According to the documentation the cache is supposed to be created like this:
Cache<String,String> c =
CacheBuilder.newCache(String.class, String.class).build();
String val = c.peek("something");
assertNull(val);
c.put("something", "hello");
val = c.get("something");
assertNotNull(val);
c.destroy();
Unfortunately at least 2 of these methods are deprecated, including the CacheBuilder class itself. I therefore tried creating the cache like this:
org.cache2k.Cache<String, Object> newCache = Cache2kBuilder.of(String.class, Object.class)
.name(cacheKey.toString())
.entryCapacity(100000)
.eternal(true)
.build();
This however throws the "java.lang.UnsupportedOperationException: loader not set" exception.
The question therefore is: how am I supposed to build the cache so that I do not get this exception?
EDIT:
This gives me the same exception:
org.cache2k.Cache<Object, Object> newCache =
CacheBuilder.newCache(Object.class, Object.class)
.eternal(true)
.build();
EDIT #2:
Just one more note: When I copy&paste the code from the wiki page I get an error - as can be seen in the image below.
With what jdk version are you testing? I'll try just removing the <> that are causing the problem for now.
Thanks very much in advance!
Michael
Cache2k looks like a very promising caching implementation.
Thanks :)
According to the documentation the cache is supposed to be created like this
There are new interfaces in place. The deprecated one is still there to support users of old cache2k versions. That will get cleared up in the next weeks. Sorry for the confusion.
Please take a look here for the latest getting started information:
https://github.com/cache2k/cache2k/blob/master/doc/src/docs/asciidoc/user-guide/sections/_start.adoc
This however throws the "java.lang.UnsupportedOperationException: loader not set" exception.
The question therefore is: how am I supposed to build the cache so that I do not get this exception?
Short answer: Either use cache.peek() or wait for 0.27, since then it is working with cache.get() transparently.
Longer answer: In our own applications I use cache.get() only when a loader is defined and cache.peek() when no loader is defined or when I want to inspect the cache only. Reserving cache.get() only for the read through usage, seemed like a good idea. However, I reasoned that it might be a caveat for new users, so I change that behavior and align it to other cache solutions.
Answer to Edit 2:
For an untyped cache use the factory method Cache2kBuilder.forUnkownTypes(). Constructing the anonymous class is only needed for specific types.
With intellij idea, how do I find out what makes a variable be visible?
An example of when it is hard:
Suppose you look at class A, and you see a variable something. If you jump to source you see that it's defined in trait X. But you don't extend trait X directly. What do you extend, then, that makes this variable visible? If you have a deeply nested hierarchy, tracking can be hard.
Any recommendations or solutions?
EDIT: Please vote for the feature if you're interested: http://youtrack.jetbrains.com/issue/IDEA-124369
I don't think that IntelliJ IDEA has any shortcut for "finding what makes a variable visible".
However you can determine it using the "Find Usages" option (Alt + F7). For example:
import java.nio._
object TempObj extends App {
def func = 2
val p = file.Paths.get("some-path")
func
}
So Find Usages on "file", tells you that its from the Package "file" (in heading of the new Tab it also shows the complete package name, ex: Find Usages of java.nio.file in Project Files).
Whereas Find Usages on func will tell you that its a Method (And the Tab heading now says: Find Usages of func() in Project and Libraries)
So now in way you can determine, what exactly makes the variable visible. This also works for imports since it shows the package from which it is imported and you can then look for import of that packages.
I know of two almost-solutions to this problem.
Go-to-declaration, as you mentioned, solves this problem in the case of local variables.
More generally, the "find usages" feature gives you a neat little breakdown by type and file of different uses of the variable. From this you can see if it's involved in a static import.
It's not perfect, but with a moment's thought these two are generally sufficient to figure out what you want.
Use ctrl+b or F4 to jump to source code. Alternatively you can use ctrl+shift+a to get option/action. You can find shortcuts at http://gaerfield.github.io/ide-shortcuts/ as well. Hope it will help.
From what I understood you want to see the code that creates an Object you use, for instance Mystery someMystery;.
That gives you two options to populate someMystery:
someMystery = ... where ... is your code to populate
someMystery and if that is the case you should follow
that code (with ctrl+B as far as you need to) to the point where it
actually creates the Mystery object.
Use CDI to populate that object instance for you, in which case you should look into the CDI mechanism in order to see in what way the object instance is populated.
In either way IMO there is no way to know for sure if the someMystery instance is of some more concrete class than Mystery, because it is decided in runtime, not in compile time, so your next bet would be to run the program in debug and see what object goes into someMystery, although you are not guaranteed to get the same type of object every time.
PS. My answer is based entirely on my java understanding of the topic, can't say if it is valid for scala also.
This might not be exactly the answer you were hoping to get.
However, quoting yourself,
If you have a deeply nested hierarchy, tracking can be hard.
Have you considered using composition over inheritance? Perhaps this would remove the need for the feature you are looking for.
Deeply nested hierarchy doesn't sound good. I understand your pain about that.
When you override vals or defs there is a little circle next to the line number that shows where it is from even when it is from nested hierarchy. Hovering over vals with the command key down also shows you a little tooltip where it is from.
Does this help?
https://youtu.be/r3D9axSlBo8
if you want class, field or method to be visible, you need to implement them as public. If it was your question.
I messed up in my code - I create lots of short-lifetime objects and use MEF to resolve dependencies. As anyone with some MEF experience knows, MEF holds onto a reference on these guys. The result is the short lifetime turns into LOOONNNGGG lifetime. So, this is easy enough to fix - use SatisfysImportsOnce on CompositionContainer.
But I'd like to make this x-check part of my unit tests now so that I don't accidently add an incorrect part into the container (or perhaps I missed one in the code search I just did). Sooo... how to do this?
My code, a bit abstracted, looks something like this:
_catalog = new AggregateCatalog();
_container = new CompositionContainer(_catalog);
_batch = new CompositionBatch();
I then add parts to a batch, and call _container.Compose(_batch). I thought perhaps something like the following would work to get a count, but it always returns zero:
int nParts1 = _container.Parts.Count();
int nParts2 = _container.Catalogs.Sum(c => c.Parts.Count());
However, both of those return zero. How can I get this count for tests? It is ok if perf isn't fantastic - this is part of the test, not the running app. Many thanks in advance!
There's not a simple way to do this. You'll need to go digging around in the internal implementation of MEF to find this information. Since the source code is available (mef.codeplex.com) this is quite possible, but still not too simple.
...and why has the package this misleading name (I assumed it had something to do with JavaME or mobile/smart phones)?
I found no references on the internet about scala.mobile.Code or scala.mobile.Location at all nor did I manage to do anything with those classes except getting ClassCastExcetions or NoSuchMethodErrors.
Actually there is not even a single test against scala.mobile in the Scala's test tree which could help understanding that code.
The classes really smell like they were forgotten in the source tree a long time ago and got accidentally released since that.
Maybe I just missed something about them?
Update:
scala.mobile was removed in Scala 2.9.
I just checked the source code.
When Scala changed the name mangling of class files a few years ago and it seems people forgot to update these classes accordingly.
So my answer would be:
At least Location has no purpose, because it is not possible to get anything sensible out of it (except exceptions) and Code without Location is severely limited. It works though if you pass the class literal to Code directly:
import scala.mobile._
val c = new Code(classOf[scala.collection.mutable.StringBuilder])
c.apply[StringBuilder, String]("append")("Foo")
c.apply[String]("toString")() // returns "Foo"
c.apply[Int]("length")() // returns 3
Looks like yet-another implementation in the standard library of reflection-slightly-nicer.
The description of Location pretty much explains what that is about:
The class Location provides a create method to instantiate objects
from a network location by specifying the URL address of the jar/class file.
It might be used by remote actors. Maybe.
As for why it has this misleading name? Well, back in 2004 smart phones had really low penetration, so maybe the association wasn't all that strong.
First of all, Hello everyone(/world) !
I am making an Eclipse RCP app' who needs to inspect some .properties located in some plugins who's Id I know at runtime.
I'd like to know if there is an easy way to read them, for example I know that in my plugin *org.anyname.myplugins.mypluginthatrocks I have a file named myawsomeproperties.properties*, how can I open it and read it easyly with a sort of getter/setter system like String getInProperties(String fileNameOrPath,String myPropertieId)
void setInProperties(String fileNameOrPath,String myPropertieId,String myPropertieValue)orString getInProperties(File file,String myPropertieId)
void setInProperties(File file,String myPropertieId,String myPropertieValue)or any other mean in that type if you get what I mean ...
Thanks in advance ^^
Well, if these are juste plain old properties, there is the good old ResourceBundle that will do the trick.
But, if you would like to do it, IMHO, a little bit cleaner, and in a more Eclipse oriented mind, I would suggest that you create a class that you use the NLS class, as explained here.