keyNew

• start = 2021-02-14 02:55
• end = 2021-02-14 03:22
• reviewer = Stefan Hanreich stefa.nosp@m.nhan.nosp@m.i@gma.nosp@m.il.c.nosp@m.om

Signature

Key *keyNew(const char *keyname, ...)

Checklist

Doxygen

(bullet points are in order of appearance)

• [x] First line explains briefly what the function does
• [x] Simple example or snippet how to use the function
• [ ] Longer description of function containing common use cases
  
  • [ ] is 0 a valid name? returns NULL if 0 is param
• [x] Description of functions reads nicely
• [ ] @pre
  
  • [x]

    Precondition

    name must be a valid key name

• [ ] @post
  
  • [x]

    Postcondition

    returns a valid key object

• [ ] @invariant
  
  • [x] add
• [ ] @param for every parameter
  
  • [ ] add @param for ...
• [x] @return / @retval
• [ ] @since
  
  • [ ] add
• [x] @ingroup
• [x] @see

Naming

• [x] Abbreviations used in function names must be defined in the Glossary
• [x] Function names should neither be too long, nor too short
• [x] Function name should be clear and unambiguous
• [x] Abbreviations used in parameter names must be defined in the Glossary
• [x] Parameter names should neither be too long, nor too short
• [x] Parameter names should be clear and unambiguous

Compatibility

(only in PRs)

• Symbol versioning is correct for breaking changes
• ABI/API changes are forward-compatible (breaking backwards-compatibility to add additional symbols is fine)

Parameter & Return Types

• [x] Function parameters should use enum types instead of boolean types wherever sensible
• [x] Wherever possible, function parameters should be const
• [x] Wherever possible, return types should be const
• [x] Functions should have the least amount of parameters feasible

Structural Clarity

• [x] Functions should do exactly one thing
  
  • it doesn't technically - but its a convenience function
• [x] Function name has the appropriate prefix
• [x] Order of signatures in kdb.h.in is the same as Doxygen
• [x] No functions with similar purpose exist

Memory Management

• [x] Memory Management should be handled by the function wherever possible

Extensibility

• [x] Function is easily extensible, e.g., with flags
• [ ] Documentation does not impose limits, that would hinder further extensions
  
  • [x] Docs say Key *k = keyNew(0); has same effect
    as Key *k =keyNew("", KEY_END);

Tests

• [x] Function code is fully covered by tests
• [x] All possible error states are covered by tests
• [ ] All possible enum values are covered by tests
  
  • [x] KEY_META
    
  • [x] KEY_FLAGS
• [ ] No inconsistencies between tests and documentation
  
  • [x] Documentation says i can work with Key *k =keyNew("", KEY_END);
    Tests say
    k = keyNew ("", KEY_END); succeed_if (k == NULL, "should be invalid"); keyDel (k);
    
  • [x] same as above with keyNew(0)

Summary

Other Issues discovered (unrelated to function)