v0.11.0

This release features the addition of many new primitive functions, whether added to the golang core, or the standard library which is implemented in lisp itself.

We also restored test-coverage back to 100% in all our internal packages.

In brief the changes were:

• Moved date/time functions into either "date:" or "time:" namespaces as appropriate.
  • Reported in #33, closed in #35.
• Added new file functions (file:lines), (file:read), (file:stat), etc
  • Reported in #32, closed in #36.
• Restored our test-coverage to 100%
  • Reported in #31, closed in #37.
• Added a builtin function (alias) to alias functions.
  • (set! pr print) could have done the same thing, for example, but this makes intent clearer to read.
  • Reported in #38, closed in #38.
• Added (directory:walk) to our standard-library.
  • This uses the new builtin (directory:entries) function.
  • Reported in #34, closed in #41.
• Added (random) to choose a random integer.
  • Use that to implement (random:char), and (random:item)
  • Reported in #43, closed in #44.
• Overhauled our README.md and documentation.
  • Reported in #40, closed in #42.
• 100% consistency when functions are invoked with invalid counts of arguments
  • Implemented in #45.
• Correctly pad hours, minutes, and seconds in the output of (hms)
  • Reported in #47, closed in #48.
• Allow the use of regular expressions to select which help to view.
  • Reported in #50, closed in #54.
• Allow writing to files, via (file:write ..)
  • Reported in #49, closed in #55.
• Use (alias) to define (first) and (rest), rather than using duplication.
• (range) now uses the upper bound inclusively.
  • This means (nat 10) and (seq 10) now stop at 10, not 9.
• Added a new "-v" flag to our binary, to show our version.
  • Reported in #51, closed in #56.
• Fixed the division function such that "(/ x)" is the same as 1/x.
  • reported in #52, closed in #57.
• Allow (= .. ) (numerical equality test) to accept arbitrary arguments.
  • Reported in #53, resolved in #58.
• Added (/= ..) function, for numerical inequality testing.
  • This returns true if all arguments are unique.
  • This returns false if any argument has been seen before.
  • Reported in #53, and resolved in #59.

Regressions

The only know regression is the failure to build this release upon Windows platforms, this was resolved in #60.

However I have still postponed the handling of #22:

• Half of me thinks we leave as-is.
• The other half of me thinks that we could add hacks.
  • But as noted they will spread, due to potentially arbitrarily deep nesting of scopes/environments.
• Finally the third half of me thinks we could be even more explicit, via the use of some (upvalue ..)-like solution.
  • Explicit here means that we're being very explicit the author must mentally consider scopes, and adding a new form is more explicit than a magical third-argument to (set!..)

Future Plans

• Nothing obvious springs to mind, though no doubt minor changes and additions will occur to me.

Feedback?

Something missing? A regression broke your build? Please file an issue.