(@ (rnrs base) log) optional second argument

(@ (rnrs base) log) optional second argument

[Ian Price]

[-- Attachment #1: Type: text/plain, Size: 401 bytes --]

Hello guilers,

Guile does not implement the optional second argument for log. I have
attached a simple patch that implements it using the formula you learn
in high school.

I have left the behaviour for two special cases alone, as there is no
clear right answer among the current R6RS implementations. If you have
strong opinions on this behaviour, feel free to write something better.

Regards,
Ian

[-- Attachment #2: 0001-Added-optional-second-arg-to-R6RS-log-function.patch --]
[-- Type: application/octet-stream, Size: 2192 bytes --]

From 733d8f68226db82fb1dda0b492ed4ab74ba779dd Mon Sep 17 00:00:00 2001
From: Ian Price <ianprice90@googlemail.com>
Date: Wed, 6 Apr 2011 01:53:38 +0100
Subject: [PATCH] Added optional second arg to R6RS log function

* module/rnrs/base.scm (log): now takes a base argument, using the
  change of base formula for logs.
* test-suite/tests/r6rs-base.test ("log (2nd arg)"): Add test cases.
---
 module/rnrs/base.scm            |    9 +++++++++
 test-suite/tests/r6rs-base.test |   16 ++++++++++++++++
 2 files changed, 25 insertions(+), 0 deletions(-)

diff --git a/module/rnrs/base.scm b/module/rnrs/base.scm
index 2f5a218..be25b6e 100644
--- a/module/rnrs/base.scm
+++ b/module/rnrs/base.scm
@@ -74,6 +74,7 @@
 
 	  syntax-rules identifier-syntax)
   (import (rename (except (guile) error raise)
+                  (log log-internal)
                   (euclidean-quotient div)
                   (euclidean-remainder mod)
                   (euclidean/ div-and-mod)
@@ -85,6 +86,14 @@
                   (inexact->exact exact))
           (srfi srfi-11))
 
+ (define log
+   (case-lambda
+     ((n)
+      (log-internal n))
+     ((n base)
+      (/ (log n)
+         (log base)))))
+
  (define (boolean=? . bools)
    (define (boolean=?-internal lst last)
      (or (null? lst)
diff --git a/test-suite/tests/r6rs-base.test b/test-suite/tests/r6rs-base.test
index 1509b04..dfddf7c 100644
--- a/test-suite/tests/r6rs-base.test
+++ b/test-suite/tests/r6rs-base.test
@@ -21,6 +21,22 @@
   :use-module ((rnrs base) :version (6))
   :use-module (test-suite lib))
 
+
+;; numbers are considered =? if their difference is less than a set
+;; tolerance
+(define (=? alpha beta)
+  (< (abs (- alpha beta)) 1e-10))
+
+(with-test-prefix "log (2nd arg)"
+  (pass-if "log positive-base" (=? (log 8 2) 3))
+  (pass-if "log negative-base" (=? (real-part (log 256 -4))
+                                   0.6519359443))
+  (pass-if "log base-one" (= (log 10 1) +inf.0))
+  (pass-if "log base-zero"
+    (catch #t
+      (lambda () (log 10 0) #f)
+      (lambda args #t))))
+
 (with-test-prefix "boolean=?"
   (pass-if "boolean=? null" (boolean=?))
   (pass-if "boolean=? unary" (boolean=? #f))
-- 
1.7.3.4

Re: (@ (rnrs base) log) optional second argument

[Andy Wingo]

Hi Ian,

On Wed 06 Apr 2011 03:11, Ian Price <ianprice90@googlemail.com> writes:

> Guile does not implement the optional second argument for log. I have
> attached a simple patch that implements it using the formula you learn
> in high school.

Thanks, applied.

We're brushing the point at which you should assign copyright to the
FSF.  Are you OK with that?  Please let me know and I'll kick off the
process.

Thanks,

Andy
-- 
http://wingolog.org/

Re: (@ (rnrs base) log) optional second argument

[Ian Price]

Andy Wingo <wingo@pobox.com> writes:
>
> We're brushing the point at which you should assign copyright to the
> FSF.  Are you OK with that?  Please let me know and I'll kick off the
> process.

Yeah sure. Just tell me what you need.

Regards,
Ian