Just a quick one here.
What are the benefits of using java.io.Console as opposed to using a BufferedReader wrapping an InputStreamReader for System.in?
Why would I use it?
Thanks for any advice!
Because it's code that is already written for you...no need to re-invent the wheel. Chances are, you're not going to get it any better than it already is.
You can use java.io.Console to present an interactive command-line to the user. You could do all that with System.in yourself, but you would have to implement things like noticing when the input was done, or readPassword, etc.
See java.io.Console is finally here!
One of the most popular feature
requests for J2SE in recent times has
been the request to improve console
support and provide a way to enter
passwords with echo disabled.
Developers know this feature 4050435
as it has been skulking in the Top 25
RFEs list for some time.
java.io.Console only works when you start a Java program from a command line without redirecting STDIN/STDOUT.
The main advantage I see with Console over System.in is that you have the readPassword() method, which won't echo the characters typed by the user (I couldn't find a way to do this with System.in).
You also have readLine() which will present a prompt and read a single line. You don't have to create your own LineNumberReader.
But, if you want your Java program to be able to read from STDIN when it's redirected from a file or pipe, you still have to use System.in.
Another trick I'm pretty sure you won't get with Console--I created my own input and output streams and replaced System.in/out with them. My implementation of the stream appended to a log file as well as echoing to the screen.
When I turned on my poor-man's "Debug Info", I could even have it tell me what program/line the sysout came from (It was slow though. It created an exception and examined the appropriate stack entry so it was off by default)
java.io.Console is used to take and read input from the user at runtime and output are displayed after processing the input from user.
For more and detailed information visit https://www.examsmyantra.com/article/58/java/java-io---console-input-and-output
Related
I have a task to implement a logging method that will take all program content up to logging method call and save it to file. The method should not overwrite already present info in file, but rather add new content to it.
What I need to write:
all program messages and output;
all user inputs in the way they appear on console (the program inquires it several times);
My thought on it is to create a StringBuilder object and start appending everything to it. Once logging method is invoked, ask for a file to save log to and save contents of StringBuilder to it. Then flush StringBuilder and continue to gather information. On second invocation, if the same filename is provided, just append new info gathered by StringBuilder.
However, this means that I will need to place gathering invocations all over the place where program output and user input are. Seems like not very optimal to me.
Are there any ideas on how to reach my goal differently and more optimally?
Thank you in advance.
Best regards,
Vadim
UPDATE: I actually was able to redirect system.out to gather everything to ByteArrayStream and then write to file on demand. But I still don't understand how to do it for inputstream. I don't need to redirect it, I still have to input eveyrthing from keyboard, it's just values that have to make it to logs in correct places. Am still searching for a solution.
Do not reinvent the wheel. Go for a logging framework. There is one integrated into java anyway.
Alternatively you can use log4j, and there are other such frameworks out there.
I am trying to test a command line app that waits for the user input after every step. I am able to test the app using System Rules provided by Stefan Birkner. Currently, I provide inputs from the beginning to the end which works like a charm and I can assert the final output from system log.
However, I want to test for the negative cases before the end of the app for which I give invalid inputs in the beginning to evaluate the error message. When invalid inputs are given, the console prints an error message and keeps waiting for the user to provide a valid input. How do I send Ctrl+C using as shown below:
systemInMock.provideLines(Ctrl+C);
systemInMock.provideLines accepts only strings. Is there a way to send Ctrl+C signal?
An example of my junit test is shown below:
#Test
public void testInValidMarker() throws Exception{
systemInMock.provideLines("abc","def","1");
Main.main(new String[]{});
assertTrue(systemOutRule.getLog().contains("Invalid marker, try again"));
}
Appreciate your help!
If I'm not mistaken, when you do ctrl+c, it doesn't actually get written to console. If that's true, then in no case will your program ever be given ctrl+c, so provideLines will never be in a position where it is given ctrl+c.
For proof, open up cmd and type in a program with program arguments (in my case, I use ant). If you type ant and then ctrl+c, the cursor is moved to a new line.
There are two ways you can control termination behavior:
You can use a shutdown hook (found from this previously asked question ). Doing this will allow you to handle what happens (potentially with issues).
Or you could create your own termination argument like -q or q, which would trigger an action to end the program (maybe a System.exit(1)). This way you can mock that input.
In UNIX/Linux, when you type CTRL-C, your shell intercepts it and sends the process a SIGINT signal -- see: How does Ctrl-C terminate a child process?
Therefore the System Rules project doesn't have anything to help you -- in this situation the process doesn't receive any character input.
By default, the whole JVM shuts down when it receives SIGINT. This is obviously bad news for a running test.
The SO question Signal handling using "TERM" -- may be of use.
A side effect of Java's portability is that for some OS features, it either abstracts things away until they're unrecognisable, or doesn't expose them at all.
I suspect what you're asking for can't be achieved.
If you're allowed to change the requirements slightly, you could ask the user to close with CTRL-D -- this closes stdin with EOF.
Although it's quite the overkill, you could launch a whole new JVM running your program, using ProcessBuilder. You might imagine you'd get an API to send arbitrary signals to that process. But for portability reasons, all you can do is process.destroy(), which sends SIGTERM.
Tried this as a comment, but it didn't read right. It's not exactly an "Answer" though.
So Java is really bad at console input. It reads an entire line at once and you can't do anything about it--there is no way to trap special characters or even see any of the input before the user hits return. Also I think a ctrl-d will close your input session--(Add that test to your use case if you don't use any other suggestion here because it can put you into a state you didn't expect!)
Three suggestions:
The simplest: If you can use a GUI and aren't really looking for an ongoing input/response REPL the simplest answer is usually to use JOptionPane to throw up a quick dialog. It's a one-line solution to get some user input, but not so good for an ongoing command-driven system.
If you can't use swing (If you are running headless) then you may have little choice, but you can use the JLine library. That will give you a lot more flexibility. This is how Groovysh does it's REPL. It will let you see each character as it is typed and do things like completion where a user might type part of a file name and hit tab and you put the rest in for him.
If you don't want to use JLine but want a REPL feel there is also a more complex GUI solution--create a swing console window. A trivial solution would just be a text input box to allow typing and a text area to display results, but there are certainly libraries out there with more complete console solutions.
The point here is that using Java standard input alone is just not a good solution for anything beyond a trivial/personal script--and even then I avoid it. Perhaps not the answer you asked for, but maybe it's the one you need :)
I get that this isn't possible to do with normal java, although if there are any libraries out this it would be very useful.
Essentially, I'm designing a console app and running into an issue that when output happens while something is typed in the input line, that input text will move up and appear before the line that just got output. Is it possible to fix this in some form so that the text you are inputting that stays at the bottom?
EX:
I'm typing something as input into my commandline app, and then the program prints something WHILE I'm typing - this causes what was originally on the input line to be scrolled up with whatever the output text was. When you are trying to type something in this can obviously be detrimental. I know it's possible to prevent this.. (Other programs have done it... EX: Minecraft Server)
(If I need to be more descriptive I can.)
You could use the help of threads. One that listens to user input, the other process the actual output. This problem is similar to basic race condition problems when multiple threads attempt to read and write to a shared resource.
Your shared resource is that console. You need to keep the Input/Output operations synchronized. Have a look at race condition.
I have two threads - one awaits for input and the other is printing the debugging info.
However only one console window, so I can't type 'exit' (or whatever to stop the process), because System.out.println constantly prints the stuff. Can I have two separate console windows for each?
P.S. I wouldn't want to use Swing just for this purpose - there must be a way.
The only way I could think of would be to have two difference processes and a link betweeen the two processes. But I don't have a clue as to how to do that. Perhaps your best bet is to use the JOptionPane class.
While you stated that you don't want to use Swing, I believe that JOptionPane would be the best option for you. Simply using JOptionPane.showInputDialog is a fast way to solve your issue. Here's a link to JOptionPane's JavaDoc.
If you really can't use Swing, there's always the option to press Ctrl + C to stop the process.
A final option would be to buffer the output and only write it after the input. After you receive input, you would flush the buffer and then deal with the input. In this manner, you would prevent the application from closing before the buffer is flushed. There are two ways to do this:
You can use a BufferedWriter with a very large size (maybe 100,000?) and store this as a static variable. Instead of calling System.out.println(), you could call MyClass.out.println()
You could override System using System.setOut(). You would create your own PrintWriter that would take any input and send it to a LinkedList (or your own LinkedList designed for chars, if you choose). I suggest you use a linked list because appending is O(1) for a linked list while appending is O(n) for an array list.
Edit:
As for hmjd's suggestion (file writing), you would do that like this:
System.setOut(new FileWriter(new File(myFileName)));
Log to a file then go into another window and tail the file (in unix/mac use "tail -f filename", in another os--install unix/cygwin!)
This keeps your log separate from your console and makes it persistent as well.
There are a lot of logging utilities out there that will help with this and will even help a bit more by telling you what file a given line is coming from.
Your question is similar to this one, so I think the answer is the same. However, maybe this question might be right for you.
What is the difference between the "standard streams" (System.out, System.in, System.err) and the "Console" (System.console()) in Java?
Every process (not just Java programs) have three streams: in, out, and error. In Java access to these has been simplified to System.in, System.out, and System.err. These are used for reading from or writing to the command line. For example, if you ever had a command line program that asked for input, which you typed and then pressed enter, that input went to standard in.
There are two out streams (out and err), because they report different things. For example, you might want to save the error output but not care about the generic output a program prints out. Or you might want to suppress the standard output so only errors are printed. Or you have a program with all kinds of problems and you want to suppress the errors so you can see what it's trying to do and not 10 million stack traces.
So, the short answer to your question is that the standard streams are ways of writing to and reading from the console.
In short, a console is the thing that might capture the streams and display them on a screen. Every program has access to the input/output streams, but not every program has access to a console.
Some programs lose access to the console as their streams are initially (or eventually redirected). Other programs are launched in configurations where they never had access to a console. In the latter case, everything written to System.out might as well have been written to /dev/null.
While this applies to Java, the Console (and the "other side" of the streams) are really operating system concepts.