Multiline paste in jshell

Question

I was trying out jshell and couldn't find option to paste multiple line expressions. Is it even possible to paste multiple lines in jshell. Similar to what scala offers with paste mode.

Answer 1

Just in case people still end up here, a small tweak to paste in an entire code block to jshell is to wrap it around with braces {} as:

{
 int c = 2;
 int j = 4;
 int x = 5; 
 // access these only in this scope though
 System.out.println("c : " + c + ", j : " + j + ", x = " + x);
}

Sample screen:

Answer 2

So if you have code like this:

 int c = 2;
 int j = 4;
 int x = 5; 

Copy and paste into jshell, only the first two statements are processed.

But if you have code like this:

  int c = 2; int j = 4; int x = 5;

And paste into jshell:

jshell> int c = 2; int j = 4; int x = 5;
        c ==> 2
        j ==> 4
        x ==> 5 

Even more lines of code like this:

HashMap<Integer, Integer> map2 = new HashMap<>(); for (int i = 0; i < 15; ++i) { map2.put(i, i);map2.put(i, i); } System.out.println(map2);

will actually work.

Why? Me don't know.

The only way I know that copy/paste will work is via (type it in jshell) :

/edit

and you can paste as much as you want.

Answer 3

I tried it and only the first two lines are processed. Also tried with extra newlines at the end and more than three lines, and still only the first two lines were ever processed. I don't know why but I suspect it's a bug.

Answer 4

This was a bug. It has been fixed: https://bugs.openjdk.java.net/browse/JDK-8169595

Answer 5

If for some reason someone finds this because their situation forces them to use this version of the JDK with the bug in it, try this:

At the end of each line, add a \ character. As long as Jshell uses any of the usual encodings, such as ASCII or some variety of Unicode (and I would be surprised if it doesn't) and it allows escape characters (it wouldn't be much of a shell if it didn't) it will force the shell to process it as a single line.

The same is true for any shell, just as long as the conventions mentioned above are held, and in my experience, 99% of the time they are. The only anomaly that you may run into is a different character used as the escape character.

If your line doesn't end with a ; you will probably need to add a ;\ so that the line:

STATEMENT

is transformed to:

STATEMENT;\

Otherwise, if the ; is already present, just add the \.

The reason this works is because each time you enter a line such as:

COMMAND

depending on your OS, the shell sees:

COMMAND\n
# or...
COMMAND\r

The last bits are the newline (Mac and Linux) and carriage return (Windows) characters. These are special characters that tell the shell that the line breaks here. Since they are special characters, they can be escaped, or made to be interpreted as plain text, with the escape character \. This means that the shell sees:

COMMAND; \
COMMAND; \
COMMAND; \
COMMAND

...as:

COMMAND; COMMAND; COMMAND; COMMAND
```