Commits on Apr 7, 2012
  1. Bump version

    committed Apr 7, 2012
  2. Avoid bombing on tab-complete attempts

    refs #32
    committed Apr 7, 2012
  3. Remove old comments

    committed Apr 7, 2012
  4. bump drawbridge and remove clj-http dep

     since drawbridge brings in a fixed version
    dakrone committed Apr 7, 2012
  5. Add windows launcher

    committed Apr 7, 2012
  6. Avoid hard fail on unrecognized signals

    Fixes #37
    committed Apr 7, 2012
Commits on Apr 6, 2012
  1. Merge pull request #43 from technomancy/clj-http-up

    Bump clj-http version for basic-auth fixes.
    committed Apr 6, 2012
Commits on Apr 5, 2012
  1. Bump clj-http version for basic-auth fixes.

    This allows us to connect to remote drawbridge repls that use basic auth.
    technomancy committed Apr 5, 2012
Commits on Apr 4, 2012
  1. don't throw exceptions when completion can't load

    (say for instance when using drawbridge to connect to an HTTP transport)
    dakrone committed Apr 4, 2012
  2. load drawbridge lazily

    cemerick committed Apr 4, 2012
  3. Use explicit metadata for lein 1.7.1

    Fixes #38
    committed Apr 4, 2012
Commits on Apr 3, 2012
  1. lazily load the clojuredocs-client

    this cuts ~2 seconds off of the startup time for reply for me
    dakrone committed Apr 3, 2012
Commits on Mar 31, 2012
  1. Correct test

    committed Mar 31, 2012
  2. Call url-connect instead of just connect.

    Accept URLs on --attach rather than just host/port.
    technomancy committed with Mar 30, 2012
  3. Make project.clj work with Leiningen 2.

    The :exclusions aren't necessary since cd-client and clojure-complete
    both depend on the same version of Clojure as reply does.
    Added a :main entry to make `lein run` a bit more convenient.
    technomancy committed with Mar 30, 2012
Commits on Mar 30, 2012
  1. Update to cd-client 0.3.4

    dakrone committed Mar 30, 2012
Commits on Mar 29, 2012
  1. Back to snapshot

    committed Mar 29, 2012
  2. Bump to alpha7

    committed Mar 29, 2012
  3. Move exit & quit to the read side

    The primary goal here is to make the client actually exit when
    the nREPL server is in a separate process, which was broken.
    This is also significantly safer, especially for scenarios
    where you have a remote nREPL server that you *don't* actually
    want to kill when you're just trying to bail out of your REPL
    session. It seems counter-intuitive to have (exit) on the
    client kill the server. We also now allow paren-less exit and
    quit - perhaps a bit magical, but this will be nice for noobs.
    The tradeoff is that this makes exit & quit pretty static.
    Redefining them is still possible, but calling (exit) or
    (quit) will quit, regardless of any new definitions on the
    execution side.  That's less than ideal, but I prefer that
    to the existing behavior.
    committed Mar 29, 2012
Commits on Mar 26, 2012
  1. Allow a custom prompt function

    closes #30
    committed Mar 26, 2012
  2. Allow starting in a different namespace

    The current mechanism uses -e (which only reads a single
    form from the arguments).
    closes #31
    committed Mar 26, 2012
Commits on Mar 24, 2012
Commits on Mar 16, 2012
Commits on Mar 13, 2012
  1. Back to snapshot

    committed Mar 13, 2012
Commits on Mar 12, 2012
  1. Send metadata to nrepl

    committed Mar 12, 2012