[#JRUBY-1151] BigDecimal to_s displays incorrect number when number has no whole part

JRuby

BigDecimal to_s displays incorrect number when number has no whole part

Created: 17/Jun/07 03:02 AM Updated: 23/Apr/08 10:02 AM

Component/s:

Core Classes/Modules

Affects Version/s:

JRuby 1.0.0

Fix Version/s:

JRuby 1.1RC2

Time Tracking:

Not Specified

File Attachments:

big_decimal_bug_fix2.patch (2 kB)
2.

bigDecimal.patch (9 kB)
3.

RubyBigDecimal.java (28 kB)
4.

testBigDecimal.rb (4 kB)

Testcase included:	yes
Patch Submitted:	Yes

Description

« Hide

The statement puts BigDecimal("0.23") prints "2.3". This is due to the way the whole and fraction part are constructed in to_s, based on the unscaled value. The unscaled value should have a leading 0 in order for the to_s function to work correctly. Included is the RubyBigDecimal.java file with a simple modification. Also included is the testBigDecimal.rb file with cases added for positive and negative numbers with 0 wholes.

All

Comments

Work Log

Change History

Sort Order:

[ Permalink | « Hide ]

Drew added a comment - 08/Oct/07 08:20 PM

I attempted to recreate this using the latest version of jruby, and it appears to already have been fixed...

$ jirb
irb(main):001:0> require 'bigdecimal'
=> true
irb(main):002:0> puts BigDecimal("0.23").to_s
0.23E0
=> nil

Can you give a test case where "2.3" will be output?

[ Permalink | « Hide ]

Charles Oliver Nutter added a comment - 22/Oct/07 12:37 AM

Already fixed at some point, and no follow-up from the reporter.

[ Permalink | « Hide ]

Charles Oliver Nutter added a comment - 06/Dec/07 12:30 AM

Reopening; apparently Rails modifies BigDecimal.to_s to be to_s("F"), which is why we couldn't reproduce this. Patch is coming from a community member.

[ Permalink | « Hide ]

Ben Curren added a comment - 06/Dec/07 12:33 AM

I created this patch off of the jruby 1.0.2 tag. This patch should be applied to the trunk as well. Please let me know if you have any questions.

[ Permalink | « Hide ]

Ben Curren added a comment - 07/Dec/07 08:40 PM

Fixed another bug in the to_s("F") method that were introduced with my last fix.

[ Permalink | « Hide ]

Charles Oliver Nutter added a comment - 12/Jan/08 12:53 AM

This needs to go in the next releases (1.0.4 and 1.1RC2)

[ Permalink | « Hide ]

Charles Oliver Nutter added a comment - 15/Jan/08 04:01 PM

Patch does not apply cleanly to trunk, so we'll have to smith it a little.

[ Permalink | « Hide ]

Robert Jay Osborne added a comment - 17/Jan/08 08:38 AM

This should patch cleanly against the trunk. I took the liberty of composing a to_s method out of the spaghetti and leaving engineeringValue and floatingPointValue as the remaining larger methods.

[ Permalink | « Hide ]

Charles Oliver Nutter added a comment - 26/Jan/08 02:09 AM

I applied this in 5751 along with a simple test:

assert_equal("0.23", BigDecimal.new("0.23").to_s("F))

If you have any additional tests we could add, from what you saw making this fix, please post them. Thank you for your help!