Umka 1.4

• New a::b syntax for imported identifiers to distinguish them from field/method access
• Separate namespace for module names
• Enumeration type
• Improved type inference for composite literals and enumeration constants
• Recursive types containing dynamic arrays or maps
• Friendly weak pointers
• %lv and %llv format specifiers for pretty-printing with printf()
• main() is optional
• New std::assert() and improved std::exitif()
• Improved disassembly output
• Bytecode optimizations
• Bug fixes

Umka 1.3.1

• for..in now allows accessing collection items by pointer
• Error handling:
  • error() is removed
  • exit() now accepts an error code returned to the OS and an optional error message
  • umkaRun() and umkaCall() now return an error code (0 on success)
  • umkaAlive() is added to check if the interpreter can still execute umkaRun() or umkaCall()
  • std.Err, std.error(), std.exitif() are added for graceful error handling
• Unicode support:
  • printf() on Windows fixed
  • Locale parameter in umkaInit() is deprecated
• Bytecode optimizations
• Cyclic import detection
• Bug fixes

Umka 1.3

• Closures compatible with conventional functions
• Variable redeclarations allowed in multiple assignments
• std.stdin(), std.stdout(), std.stderr(), std.fflush()
• Dynamic access to API via umkaGetAPI() without linking against Umka
• API change: umkaGetType() replaced with typeptr()
• Bug fixes

Umka 1.2

• Closures
• Type switch statements
• Closing braces are allowed on a new line
• Faster string implementation
• C API for creating dynamic arrays
• Bug fixes

Umka 1.1

• Ternary conditional operator: a ? b : c
• repr replaced with %v format specifier in printf, sprintf, fprintf
• Bytecode compiler optimizations
• Bug fixes

Umka 1.0

• Type inference in composite literals
• Optional initializer in new()
• All standard library modules are embedded into Umka executable
• Memory leak sanitizer
• Optimized function calls
• Optimized UTF-8 encoding/decoding
• Optimized len() for strings
• Improved std.rand()
• New Umka API documentation
• Bug fixes

Umka 0.10

• Uninitialized dynamic arrays and maps are treated as valid
• Optimized append(), insert(), delete()
• cap() for dynamic arrays, copy() for maps
• Covariant dynamic arrays
• Map key types must be comparable
• any as an alias for interface{}
• Optimized += for strings
• OS-specific suffixes (_windows or _linux) are allowed in UMI file names
• Better error messages
• Bug fixes

Umka 0.9

• Strings are now immutable
• Interface-typed map keys are no longer allowed
• New dynamic array handling functions: copy, insert
• New math functions: ceil, floor and the % operator for reals
• New date/time handling functions: std.localtime, std.gmtime, std.mktime, std.timestr
• New flow control function: exit
• Improved I/O functions: sprintf, std.fread, std.fwrite
• C API for garbage collection and map items access
• Source file and line in stack traces
• Bug fixes

Umka 0.8

• Built-in maps (map.um is deprecated)
• Module aliases
• Unused identifier/unused import warnings
• Sandbox mode (disables accessing the file system and/or UMIs)
• Bug fixes

Umka 0.7

• Variadic functions
• Default parameters of structured types
• Implicit interface-to-concrete conversions are no longer allowed
• Null strings are equivalent to empty strings
• fiber is now a pointer; ^fiber is no longer needed
• Dynamic arrays are now accessible through the C API (read only)
• valid() checks if a dynamic array, interface or function variable is not null
• Improved garbage collection for long lists
• Web-based playground
• Bug fixes