Changelog for dataordlist0.4.7.0
Version 0.4.7.0: (20141119)
foldt
tweaked to be made slightly lazier, thus also affectingmergeAll
,mergeAllBy
,unionAll
, andunionAllBy
. Thanks goes to Will Ness for providing the following example that previously entered into an infinite nonproductive loop:
let ps=3:minus[5,7..] (unionAll [[p*p,p*(p+2)..]p<ps]) in take 100 (2:ps)
Note that the _de facto_ semantics of `foldt` has been modified
slightly: instead of computing the sum of a list, it computes
the sum of a list plus zero. Of course, if zero is a true
identity of plus, then this change is of no semantic consequence.
Version 0.4.6.1: (20140419)

Added strictness annotations inside of sortOn and nubSortOn

Tweaked documentation of nubSort

Conditionally reexport sortOn from Data.List when available. (See GHC commit 44512e3c)
Version 0.4.6: (20140215)

Generalized type signature of
isectBy
,minusBy
, andminusBy'
, leaving the definition unchanged. 
Made the documentation of
union
more precise.
Version 0.4.5: (20120312)

New function,
minus'

Exported
foldt
andfoldt'

Documentation improvements

Reverted the implementation of
mergeAll
andunionAll
to version 0.4.2 because the simplified implementation would force the head of the inner list appearing at the next highest power of 2 well before it was necessary to do so.
Version 0.4.4: (20101224)

Simplified the implementation of
mergeAll
andunionAll
based on comments from Will Ness.http://www.haskell.org/pipermail/haskellcafe/2010December/087587.html
Version 0.4.3: (20100302)

Improved the implementation of
nubSort
, mirroring the improvements made toData.List.sort
in GHC6.13.20091224 and first released in GHC7.0.1. Thanks to Gwern Branwen for calling the change to my attention. Instead of initially breaking the input list into singletons before the merge process, the improved version breaks the input list into monotonic runs.http://www.haskell.org/pipermail/libraries/2010March/013066.html

Minor formatting improvements in the Haddock documentation.
Version 0.4.2: (20100218)

Fixed nonproductive loop in
unionAll
when applied to an infinite list of lists. Thanks to Omar Antolín Camarena for reporting the bug and Heinrich Apfelmus for some useful comments.http://www.haskell.org/pipermail/haskellcafe/2010February/073403.html http://www.haskell.org/pipermail/haskellcafe/2010February/073437.html

Added regression test to test suite.
Version 0.4.1: (20100217)

Simplified the implementation of
mergeAll
andunionAll
thanks to some pointers by Heinrich Apfelmus. This introduced an infinite nonproductive loop intounionAll
, which was fixed in 0.4.2 without giving up the simplifications.http://www.haskell.org/pipermail/haskellcafe/2010February/073356.html

Minor documentation fixes
Version 0.4: (20100215)

The "CHANGES" file was added to document the changes between releases.

Documentation Improvements

A rough first pass at a test suite

The functions
mergeAll
andunionAll
were added. They operate on a possibly infinite list of possibly infinite ordered lists; assuming the heads of the lists are ordered.Thanks goes to Omar Antolín Camarena, Heinrich Apfelmus, and Dave Bayer.
Omar Antolín Camarena suggested the addition, located the article used as the basis for the implementation, and was quite helpful with testing and debugging.
Heinrich Apfelmus wrote his "Implicit Heaps" article, where he simplified an algorithm by Dave Bayer. It is this article that forms the basis of our implementation.
http://apfelmus.nfshost.com/articles/implicitheaps.html
Dave Bayer posted his 'venturi' implementation to the haskellcafe mailing list on 2007 Jul 22. It also appears as "BayerPrimes.hs" inside of Melissa O'Neill's "haskellprimes.zip":
http://www.haskell.org/pipermail/haskellcafe/2007July/029391.html http://www.cs.hmc.edu/~oneill/code/haskellprimes.zip
Version 0.2: (20100207)

The module name was changed from
Data.OrdList
toData.List.Ordered

Fixed bugs in
insertSetBy
,insertBagBy
, andnub
. The insertion functions assumed reversed lists, whilenub
failed to remove duplicates.Thanks to Topi Karvonen for reporting the first issue!

Changed semantics of
insertSetBy
slightly: the new version replaces an element if it is already there. If the old semantics turns out to be important, a new function can be added at a later date. 
Changed semantics of
nubBy
: the new version negates the binary relation, so thatnew_nubBy f == old_nubBy (not . f)
. It is now in better keeping with the spirit of the rest of the library, and mades the bug innub
more obvious. 
Better documentation, I hope. At the very least, the process of documenting
nubBy
revealed the bug innub
.
Version 0.0.1: (20090710)
 The initial release, sadly, did not contain the source file
Version 0.0: (20090710)
 Initial Release