Summary

integer? returns true for special flonum types on 64-bits machines

Metadata

• Id: c5ebc2837bf0ace0ba51bdc5d24be435fcb4d3ec
• Trac id: 791
• Type: defect
• Reporter: sjamaan
• Owner: ckeen
• Cc:
• Status: closed
• Component: core libraries
• Estimated difficulty:
• Resolution: fixed
• Priority: major
• Milestone: 4.9.0
• Version: 4.7.x
• Changetime: 2012-09-24 21:47:48 UTC
• Created: 2012-02-21 19:21:45 UTC
• Keywords: flonums, integers

Description

The C_i_integerp function doesn't first unpack flonum values before checking if they represent infinity or NaN. On x86-64, this means that (integer? +nan.0) => #t and (integer? +inf.0) => #t.

A suggested initial patch is at http://paste.call-cc.org/paste?id=2d5b17c595ae79bacbe36c52d4613ccd3e0373a2, but Christian said he'd work on a proper patch which includes changes to the testsuite to catch these errors (for the other number predicates too).

Changes and comments

[2012-03-04 12:02:38 UTC] ckeen wrote:

Patch has been submitted to chicken-hackers

[2012-03-04 12:03:54 UTC] ckeen changed status from new to accepted

[2012-03-18 17:10:25 UTC] ckeen changed status from accepted to closed

[2012-03-18 17:10:25 UTC] ckeen set resolution to fixed

[2012-03-18 17:10:25 UTC] ckeen wrote:

Fixed with patch ae91f816d0a3911d71a37e22435f83ade9c77d69 to master

[2012-09-24 21:47:48 UTC] felix changed milestone from 4.8.0 to 4.9.0

[2012-09-24 21:47:48 UTC] felix wrote:

Milestone 4.8.0 deleted