Bug Fixes

• Fixes the initialization with an inital tree in Newick format (-init).
  Big thanks to @nspope for reporting and locating this bug along with
  suggesting a fix!

Improvements

• The labeling of internal nodes in the oriented forest output
  no longer changes within the same tree topology. That makes
  it easier to identify different topologies.

Bug Fixes

• Now requires that subpopulation are defined via the -I argument
  before any demographic options for all population (-M, -G or -N)
  are given (#108).
  Thanks to Jonathan Terhorst (@terhorst) for reporting this bug.
• When the multiple of the merge/admixure arguemtns (-ep, -eps and -ej)
  are used at the time, they are now executed in order in which
  they are provided on the command line (#121). This behavior now
  follows ms' implementation.
  Thanks to Scott T Small (@stsmall) for reporting this.

• Fixes a different bug that could occur when using the "-eI" option.
  The contemporaries buffer was always cleared correctly between
  repetitions, and could cause scrm to throw an error (#104).
  Thanks again to @jackkamm for reporting this bug.

Fixes a bug that could lead to false results when the "-eI" option was used to include non-contemporary samples. In that case, scrm could skip the time interval between the MRCA of the contemporary samples and the time of the "-eI" event (if any), and neglect all events during that time. We recommend to repeat all simulations which used "-eI" with the updated version (#102).

New Features

• New command line option "--transpose-segsites" that prints the segregating
  sites matrix with rows repesenting mutations instead of individuals. This
  also adds the time at which the mutation occurred as additional information
  (#85).

Improvements

• scrm now throws an error when mutation or recombination rate changes
  have invalid sequence positions (#82).
• The length of segments between recombinations is now always reported
  in non-scientific notation (#81).
• Improved the error message when lines could not coalescence because
  of missing migration or negative population growth (#87).

Bug Fixes

• Forest.coalescence_finished has not set to true after a pairwise
  coalescence event (#89). This had no effect on the command line version
  of scrm.
• Fixes scrm's tests suite on 32 bit systems (#98).

Bug Fixes

• scrm had extensive memory consumption when simulating trees in large models (#79). This is now fixed, it is however no longer possible to use "-O" and "-T" at the same time.

New Features

• It is now possible to specify the approximation exact window in number
  of recombination events (#73).
• scrm now uses a conservative approximation by default (#75).

Bug Fixes

• Critical: Position based rates changes (-sr and -st) were only
  applied to the first locus. If nrep was greater than one, all other loci
  were simulated with the rates from the end of the first one (#74).
  All simulations done with previous versions that used -sr or -st
  together with nrep greater than one should be repeated.
• In large models, the Newick trees can get larger than the character limit
  of a C++ string. In this cases, only incomplete trees were printed (#76).

• Bug fix: Fixed a cache invalidation error that could lead to runs being
  aborted after 100k recombinations (#70). Thanks to @jeromekelleher for
  report this.
• Minor: Small updates to the documentation and citation information.

Update 2015-06-04: Replaced scrm-1.5.1-x64-static.tar.gz with a binary that actually is statically linked

• New feature: Added flag "--print-model" which prints a textual representation
  of the demographic model for verification and debugging (#60).
• Bug fix: When multiple population splits and/or merges occurred at the same
  time, only one of them affected each line while the others were ignored (#61).
• New feature: Added support for partial population admixtures (-eps, #62).
• In addition to "-seed" scrm now also supports "-seeds", as ms also
  supports both (#65).

• Bug fix: Wrong population size where calculated when migration rates
  changes or population splits and merges occurred in an growth period (#56)
• Added an error message when a population size is set to 0 (#52)
• Bug fix: When -es was used at time 0, scrm ran into an endless loop (#53)