Making StreamReader Iterable

[oahmad04]

I tried to make StreamReader iterable with the following:

extend StreamReader
    """Makes StreamReader iterable (by line)"""
    
    cue enumerate as String*
        """Yields each line in the open file"""
        
        test
            
            # Make and test a file:
            
            File.writeAllText("Source", "line 1\nline 2\nline 3")
            lines = []
            
            using file = File.openText("Source")
                for line in file, lines.append(line)
                
            assert lines == ["line 1\n", "line 2\n", "line 3"]
            
            File.delete("Source")
            
            # Make and test an empty file:
            
            File.writeAllText("Source", "")
            lines = []
            
            using file = File.openText("Source")
                for line in file, lines.append(line)
            
            assert lines == []
            
            File.delete("Source")
        
        body
        
            while true
                line = .readLine
                if line is nil, break
                yield line to !

I get the following errors:

Code: Select all

iterateFile.cobra(23): error: Cannot infer the type for "line" from the "for" loop.
iterateFile.cobra(35): error: Cannot infer the type for "line" from the "for" loop.


What did I do wrong? Thanks.

[Charles]

Part of being enumerable to a "for" loop is implementing IEnumerable. You cannot make this happen to an existing class (by using extend or any other technique).

Note that .NET 4 introduced File.readLines which basically does what you're looking for:
http://msdn.microsoft.com/en-us/library ... 3(v=vs.110).aspx

If you still wanted to put something on a reader, and you don't mind slapping a method on the reader in the for loop, then you could do something like this:

extend TextReader

    def readLines as String*
        ...

...

for line in tr.readLines, print line

Note that other possibilities for looping through a reader include characters, words and regex matches. For this reason, I would actually prefer this form over making the reader directly enumerable.

[Charles]

Also, in Cobra tests, it is customary to use StringWriter and StringReader in place of actual file I/O. The tests will run faster, have less side effects on the machine and less of a chance of failing for external reasons (permissions, etc.).

[oahmad04]

Thanks for your response. I thought it would be like Python where you just add a "def __iter__" method and it works. Now I know.

Note that other possibilities for looping through a reader include characters, words and regex matches. For this reason, I would actually prefer this form over making the reader directly enumerable.

Good point. I was missing the Python syntax. I should probably just use what's given with Cobra instead of trying to make working with files more like Python.

Note that .NET 4 introduced File.readLines which basically does what you're looking for:
http://msdn.microsoft.com/en-us/library ... 3(v=vs.110).aspx

I saw this actually. Very handy. I was just worried it wouldn't work with huge files. I don't actually have any huge files in mind. I'm not working on any projects. I'm just learning Cobra by making little namespaces I think will be useful.

[oahmad04]

I saw this actually. Very handy. I was just worried it wouldn't work with huge files. I don't actually have any huge files in mind. I'm not working on any projects. I'm just learning Cobra by making little namespaces I think will be useful.

Oh, I haven't seen this. I thought you posted a link to readAllLines. This is cool.