Clojure reader in Clojure
Pull request Compare This branch is 316 commits behind clojure:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.

A complete Clojure reader and an EDN-only reader, works with Clojure versions >= 1.3.0

Rationale offers all functionality of the reader from clojure-1.5.0, and more.

This means safer read/read-string, an edn-only reader, tagged-literals support, default-data-reader-fn support for every Clojure version >=1.3.0

For a list of additional features of the reader, read Differences from

Moreover, by using reader types from, if using an IndexingReader, column info is available and both line and column metadata is attached not only to lists, but to symbols, vectors and maps too, when additional debugging info is needed (note that the edn reader doesn't add any line/column metadata at all).

Note that it uses ex-info which is available on clojure.core only from clojure-1.4.0. If using clojure-1.3.0 and needing access to ex-data, use

Public API

There are four public namespaces:

Refer to docstrings in each namespace and to the project's autodoc for more documentation.

Releases and Dependency Information

Latest stable release: 0.7.4

Leiningen dependency information:

[org.clojure/tools.reader "0.7.4"]

Maven dependency information:


Example Usage

To read data structures, functions from should be used, since those are safe and don't allow any code execution at all.

Note that since no code-execution is permitted, reader literals are also disabled.

(require '[ :as edn])
;=> nil
(edn/read-string "1")
;=> 1
(edn/read-string "#inst \"2010-11-12T13:14:15.666\"")
;=> #inst "2010-11-12T13:14:15.666-00:00"
(let [my-unknown (fn [tag val] {:unknown-tag tag :value val})]
   (edn/read-string {:default my-unknown} "#foo bar"))
;=> {:unknown-tag foo, :value bar}
(edn/read-string {:readers {'foo (constantly 1)}} "#foo bar")
;=> 1

To switch from using clojure.core/read-string to in your projects, put this in your namespace declaration:

(:refer-clojure :exclude [read read-string])
(:use [ :only [read read-string]])

If (and only if) reading from a trusted source, and advanced features that need some level of code-execution during read are needed, functions from should be used.

(require '[ :as r])
;=> nil
(r/read-string "1")
;=> 1
(r/read-string "#=(+ 1 2)")
;=> 3
(binding [r/*read-eval* false]
  (r/read-string "#=(+ 1 2)))
=> ExceptionInfo #= not allowed when *read-eval* is false

To switch from using clojure.core/read-string to in your projects, put this in your namespace declaration:

(:refer-clojure :exclude [read read-string *default-data-reader-fn* *read-eval* *data-readers*])
(:use [ :only [read read-string *default-data-reader-fn* *read-eval* *data-readers*]])

Reader types example usage:

(require '[ :as t])
;=> nil
(def reader (t/string-push-back-reader "1"))
;=> #'user/reader
(t/read-char reader)
;=> \1
(t/unread reader \a)
;=> \a
(t/peek-char reader)
;=> \a
(t/read-char reader)
;=> \a
(t/read-char reader)
;=> nil

Note that the pushback buffer is of dimension 1 by default, and an exception will be thrown if trying to unread more chars than the pushback buffer dimension.

Every predefined reader type has an additional arity that allows to specify the pushback buffer dimension.

(def reader (t/string-push-back-reader "" 2))
;=> nil
(t/unread reader \a)
;=> \a
(t/unread reader \b)
;=> \b
(t/read-char reader)
;=> \b
(t/read-char reader)
;=> \a
(t/read-char reader)
;=> nil

Differences from

There are small differences from clojure.lang.LispReader:

  • read throws an ex-info for almost every exception, whereas clojure.lang.LispReader/read throws a ReaderException wrapping the causing exception.
  • read is capable of reading \x escaped chars
  • read is capable of reading Infinity +Infinity -Infinity and NaN as per #CLJ-1074
  • read is capable of reading literal tags containing periods, fixing #CLJ-1100
  • read is capable of reading the symbol / with an explicit namespace, e.g. foo//, whereas clojure.lang.LispReader/read throws an exception. Refer to #CLJ-873. Except for this special case, read throws an exception if a symbol contains more than one / character, whereas clojure.lang.LispReader/read allows them, returning a symbol with one or more / characters in its namespace name.
  • adds additional line/column info to symbols, vectors and maps when possible
  • has an additional arity with which is possible to specify the reader to read from


  • Release 0.6.0 on Feb 03, 2013
    • Initial release.
  • Release 0.6.2 on Feb 04, 2013
    • Add line/column metadata on vectors, maps and symbols
  • Release 0.6.4 on Feb 08, 2013
    • Fix Unicode char reading
    • Add *default-data-reader-fn* support
    • Add an EDN-only reader
    • Disable record literals reading when *read-eval* is bound to false
    • Made % a symbol constituent char
    • Made the EDN reader API match the clojure.edn one
  • Release 0.6.5 on Feb 09, 2013
    • Fixed reading @ ~ and `
  • Release 0.7.0 on Feb 14, 2013
    • Fixed #TRDR-1 by @jafingerhut
    • Made compatible with clojure-1.3.0
    • Decoupled from clojure.core vars (*read-eval*, *default-data-reader-fn*, *data-readers*)
    • and return nil if string is nil or empty
    • Added comprehensive docstrings
  • Release 0.7.1 on Mar 02, 2013
    • Added the syntax-quote macro to the public API
  • Release 0.7.2 on Mar 02, 2013
    • Fixed evaling read objects for clojure 1.5
  • Release 0.7.3 on Mar 08, 2013
    • AOT compile only ExceptionInfo.
  • Release 0.7.4 on Apr 19, 2013
    • Fixed docstrings position and typos
    • Fixed a bug that would have prevented future alpha versions of clojure to have column metadata
    • Made symbols containing two consecutive ':' illegal, as per
    • Made symbols containing two consecutive '/' illegal
    • Removed reflection warnings

Developer Information


Copyright © Nicola Mometto, Rich Hickey & contributors.

Licensed under the EPL. (See the file epl.html.)