Cobra Differences From C#
Illustrated
# Cobra d = {'a': 1, 'b': 2} assert d['a'] == 1 and d['b'] == 2 for key, value in d print '[key] = [value]' # string interpolation
// C# var dictionary = new Dictionary<string, int> { {"a", 1}, {"b", 2} }; Assert.IsTrue(dictionary["a"] == 1 && dictionary["b"] == 2); foreach (var key in dictionary.Keys) { var value = dictionary[key]; Debug.WriteLine(string.Format("{0} == {1}", key, value)); }
credit: http://fir3pho3nixx.blogspot.com/2011/01/when-c-is-not-enough.html
Quick Reference
- Dynamic Typing and Type inference (doesnt require Type or keyword to declare locals)
- optional type declarations
- Block structured by indentation only (like Python) no squiggly braces (brackets)
- nil tracking - variables can be declared as nil assignable (or not)
- Built in language support for
- unit test code
- assertions
- Design by contract - invariants, preconditions and postconditions
- Syntax support for builtin
- Lists - [] - List of Dynamic
- Dictionaries - {:} Dictionary of <Dynamic, Dynamic>
- Sets - {,}
- Line Oriented
- # to comment line (instead of //)
- /# ... #/ for multiline comments
- Line continuation is implied in method argument lists: foo.bar(arg1,[NEWLINE]arg2)
- Use a trailing _ for explicit line continuation
- Tab or 4 Spaces to indent line (Fixed 4 space indent), Cannot mix Tabs and spaces for an indent level
- No statement terminator (goodbye to ';')
- Fractional numbers default to decimal type (rather than int)- can be overridden (to float/double) with a compiler commandline option
- Method names and properties start with lowercase. (i.e lowerCamelCase)
- Access to methods/properties in used libraries are also done via (implicitly mapped) lowerCamelCase methodnames
- In resulting binaries such as .exe and .dll, the .NET standard CamelCase is used so that Cobra libraries vend out naturally to C# and VB.NET
- Class names must be (upper) CamelCased or at least start with an upper case letter)
- instance variable optionally prefixed with _ (implicitly makes them protected) or __ (implicitly makes them private).
If so prefixed, these can be accessed directly in methods (without a leading '.') - methods (and properties) on current object invoked with leading this. or more usually (shortcut) just ..
- constructor/initializer method is called init, has no return value
- String type accessed as String
- Numeric types specified as int, uint, float/float64, float32 and decimal
- Sized int and uint types all have the same naming form
- int8, int16, int/int32, int64 instead of SByte, Int16, Int32/int, Int64
- uint8, uint16, uint/uint32, uint64 instead of Byte, UInt16, ...
- float32, float/float64 instead of float, double
- Type of numeric literals can be dictated using suffixes ( e.g. unsigned 47_u, 47_u8, signed 47_i8, 47_i64)
- Variable specification - typing, access modifiers, argument attributes specified by keyword as rather than C-like/Java-like syntax
- e.g. x as int32, instVar as string public (vs int32 x; public String instVar; )
- typecasting by keyword to as in 'x to Type' rather than '(Type)x'
- cast to and away from nilable Type.
- is shared to declare a static type
- <> instead of != for 'not equals' comparisons
- Condition negation uses keyword not ( as in not a >10 )
- boolean conditions use keywords and and or ( as in if a>3 and b<10 or b>55 )
- boolean and and or have equal precedence in Cobra whereas in C# they are differentiated. When converting code to Cobra, additional parentheses may be needed.
- nil for no object (null) value
- Supports +=, -=, etc. Does not support ++ or -- unary operators ( use += 1 and -= 1 instead).
- boolean conditions can be tested on any type. True is any of
- non false (bool Type)
- numeric not 0
- non zero char
- not nil
- blank strings and empty collections are true (non nil) - must check .length for strings and .count for collections
- Strings delimited by single or double quotes, normal range of \c expansion ( e.g \n, \t)
- Substitute expressions (values) directly in strings by enclosing with [].
- such expressions can have optional String.format specification
- 'Raw' Strings (r"a raw string") no escape sequence expansion ( vs c# @"a raw string")
- 'NoSubstitution' strings ( ns" no substitution [done] here")
- DocStrings on Programs, Classes, Interfaces (etc), instance variables, methods and properties
- Different syntax for characters c'x'
- since '[] used for declaring literal Lists, array literals are specified using @[] - comma separated values
- ref keyword for getting a reference (e.g., safe pointer) to a method, rather than calling it, so the reference can be stored or passed around. (for delegates)
- sig keyword for declaring a delegate type
- def keyword to declare a method/function
- e.g def doCalc( start as int) as int instead of int doCalc( int start)
- use keyword instead of 'using' to get access to a non default namespace
- branch... on, on, [else] statement rather than c# case statement
- Slicing syntax for sequences (Lists and Strings) [start:stop:step] as well as normal library subsequence methods (substring, range, ...)
- numeric for loop uses slice syntax. e.g. for i in 1:10:1 vs for( int i =1; i<=10; i++)
- the Cobra ternary operator has a different syntax: if(<condition>, <texpr>, <fexpr>). Example at http://cobra-language.com/docs/manual/expressions/if.html
- more as they surface