[#51213] [ruby-trunk - Bug #7645][Open] BigDecimal#== slow when compared to true/false — "mathie (Graeme Mathieson)" <mathie@...>

11 messages 2013/01/01

[#51328] [ruby-trunk - Bug #7676][Open] Comparison of Float::NAN in array behaves unexpectedly — "simonrussell (Simon Russell)" <spam+ruby@...>

11 messages 2013/01/09

[#51347] [ruby-trunk - Bug #7679][Open] IRB history is broken — "zzak (Zachary Scott)" <zachary@...>

15 messages 2013/01/10

[#51389] [ruby-trunk - Bug #7688][Open] Error hiding with rb_rescue() on Comparable#==, #coerce and others — "Eregon (Benoit Daloze)" <redmine@...>

34 messages 2013/01/11

[#51430] [ruby-trunk - Bug #7696][Open] Lazy enumerators with state can't be rewound — "marcandre (Marc-Andre Lafortune)" <ruby-core@...>

15 messages 2013/01/14

[#51437] [ruby-trunk - Bug #7698][Open] RubyGems 2.0 has an incompatibility about installation of extension libraries — "mrkn (Kenta Murata)" <muraken@...>

21 messages 2013/01/15

[#51454] [CommonRuby - Feature #7701][Open] Non-optional (required) keyword args — "headius (Charles Nutter)" <headius@...>

31 messages 2013/01/15

[#51499] [ruby-trunk - Feature #7712][Open] Add .txt extensions to all plain-text documentation files for Windows users — "postmodern (Hal Brodigan)" <postmodern.mod3@...>

9 messages 2013/01/18

[#51619] [ruby-trunk - Feature #7738][Open] Deprecate Set#+ as an alias of Set#|, use it for symmetric difference. Introduce Hash#| for Hash#reverse_merge in Rails. — "alexeymuranov (Alexey Muranov)" <redmine@...>

11 messages 2013/01/24

[#51623] [ruby-trunk - Feature #7739][Open] Define Hash#| as Hash#reverse_merge in Rails — "alexeymuranov (Alexey Muranov)" <redmine@...>

24 messages 2013/01/24

[#51714] [CommonRuby - Feature #7747][Open] Expanded API for Binding semantics — "jballanc (Joshua Ballanco)" <jballanc@...>

19 messages 2013/01/27

[#51742] [ruby-trunk - Bug #7756][Open] clang 3.2 sees through UNINITIALIZED_VAR macro, gives warning — "drbrain (Eric Hodel)" <drbrain@...7.net>

10 messages 2013/01/29

[#51763] [ruby-trunk - Bug #7758][Open] Ruby on Windows crashes when active codepage is codepage 65001 and outputting unicode character — "joshc (Josh C)" <josh.nw@...>

16 messages 2013/01/30

[ruby-core:51608] [ruby-trunk - Bug #7724] 6 bugs with Range#bsearch over floats

From: "marcandre (Marc-Andre Lafortune)" <ruby-core@...>
Date: 2013-01-23 19:38:54 UTC
List: ruby-core #51608
Issue #7724 has been updated by marcandre (Marc-Andre Lafortune).


mame (Yusuke Endoh) wrote:
> Thanks Marc-Andre!
> 
> But I'm very sorry that I'll have no enough time to pursue this issues.
> Could you please create a patch for C implementation?

Sure, I'll do it.

> Notice that the current implementation does NOT assume that the internal
> representation of double is IEEE 754.  Please do not depend on it but
> use C's constants about double, such as FLT_RADIX and DBL_MANT_DIG.

Actually, I don't think there is need to know how many bits are used for the mantissa vs exponent; I only need to know that double is represented with exponent bits then by mantissa bits. Apart from VAX (which had strange byte ordering, but we don't support), I believe this to be true.

So floatings can be seen as [exp, mantissa], thus have the same ordering as the the corresponding integers (forgetting about the sign). I.e. (double)(++(int64_t)a_float) is the smallest float strictly greater than a_float (assuming 0 <= a_float < infinity)

Do we have a list of non IEEE architectures we're hoping to support? Are there systems where sizeof(double) != sizeof(int_64_t)?

----------------------------------------
Bug #7724: 6 bugs with Range#bsearch over floats
https://bugs.ruby-lang.org/issues/7724#change-35563

Author: marcandre (Marc-Andre Lafortune)
Status: Assigned
Priority: Normal
Assignee: marcandre (Marc-Andre Lafortune)
Category: core
Target version: 2.0.0
ruby -v: r38825


Take the following code, with from, to and search all Integer or all Float:

  (from...to).bsearch{|f| f >= search}

I expected it to:
0) never yield and return nil if range is empty, i.e. from >= to ("trivial test")
1) never yield more than 65 times for floats
   or Math.log(to-from, 2)+1 for Integer ("log test")
2) return search if from <= search < to, nil otherwise ("coverage test")
3) never yield the same `f` ("uniqueness test")
4) if range is exclusive, never yield `to` nor return `to`; if range is inclusive and block returns false always yield `to` ("end of range test")
5) never yield a number < from or > to ("out of range test")

These conditions are all respected for Integers but all can be broken for Floats
Test 0, 1, 3, 4 & 5 fail even for small ordinary float values, while test 2 requires some larger floats, say 1e100 to fail.
For example bsearch can yield over 2000 times (instead of at most 65).


These tests and a correct Ruby implementation of bsearch can be found here: https://github.com/marcandre/backports/compare/marcandre:master...marcandre:bsearch
My implementation also passes the MRI tests.


-- 
http://bugs.ruby-lang.org/

In This Thread

Prev Next