New Features
After Release 1.3, we will collect all changes in a CHANGELOG file in the repository with each commit that contains the changes. However, before that we will collect a list here.
Headlines:
- Functionality
- New full sensitivity local aligner stellar (Blast speed for genome sized data)
- New split read mapper splazers
- New journaled string data structure
- New random function generator
- Completely refurbished score module.
- New fragment store for 2nd generation sequencing:
- Represents multi-read alignments and genome annotations
- Supports SAM, Amos, GFF, GTF, UCSC Genome Browser formats
- Can be displayed as a visual SVG file
- New read simulator mason
- Development support
- New test system. Templates for applications and tests are available.
- Harmonized naming conventions.
- Documentation and Distribution
- Added a new WIKI based tutorial
- The documentation can now be built in less than one minute.
- Changed license to 3-clause BSD license.
- Nightly documentation builds can be found at www.seqan.de/dddoc/html_devel/
- Current changes in the trunk can now be found in the file projects/library/CHANGELOG and on the wiki at ChangeLog.
Changelog
- Various identifier have been renamed (list of renamed identifiers).
- SeqAn is now built nightly and errors are reported to our CDash Dashboard.
- We dropped Support for Visual Studio 7, GCC 3.x and added support for LLVM and more current Visual Studio and GCC versions. The full list now is:
- On Windows: Visual Studio 8, Visual Studio 9, and Visual Studio 10, MinGW
- On Mac Os X and Linux: GCC 4.1-4.5 as well as the current LLVM trunk version.
- Experimental seeds2 module replacing module seeds in the future.
- Overhauled test system.
- Documentation building is much faster now.
- Added journaled string feature in module string_journaled.
- Added swiss army knife (sak) tool for slicing and dicing DNA files (FASTA, FASTQ).
- Added RNA alphabet
- Implemented local version of the SWIFT filter additionally to semi-global version (index/find_swift.h).
- Added banded local alignment
- Fixed assignments between infixes and suffixes (#320).
- Patterns provide a copy constructor now (#318).
- Fixed StringSet documentation: Default specialization (#275).
- Fixed problem in suffix array implementation with max repeat calculation (#61).
- Fixed _MakeSigned metafunction to work with unsigned long long.
- Changing Infix Segment constructor to construct an infix of the whole string if no begin and end positions are given (#288).
- Listing headers to include for each symbol in the documentation (#25).
- Added much more documentation, fixing some issues (#326, #330).
- Documentation works now correctly in Opera.
- Fixed bug of insertValue with String of CharStrings (#273).
- ModifiedStrings can now be appended to std::string objects (#251).
- Implementing std::swap for StringSet and String (#296).
- Fixing bug in assignSource and globalAlignment.
- Fixing issue with missing needle in MyersUkkonen (#319).
- Indices can be built in StringSets of Segments now (#191).
- HMM algorithms use the new random module now (#351).
- Fixing collision with begin and end in C++0x (#374).
- operator- is now implemented for iterators of a string and infixes of a string (#317).
- Unused reference counting feature of Holder is removed (#381).
- Documented haystack, host, needle more clearly (#371).
- atEnd() now works for StringSet iterators (#84).
- The command line parser now accepts floating point arguments (#295).
- Pdf<Uniform> now works correctly with float types (#502).
- Line endings are set to native now via the svn:eol-style property (#386).
- Non-external nested typedefs are suffixed with an underscore now to prevent accidental access (#289).
- Naming of acronyms is more consistent now (#73).
- Internal symbols for tag definitions do not violate the C++ standard any more (#290).
- fill has been renamed to resize (#77).
- added std::vector adaptions (#381)
Deprecations
- supremumValue, infimumValue are deprecated
- The find_motif module has known issues and will be replaced in a future version.
- The seeds module has known issues. There is an experimental new version seeds2 available in SeqAn trunk that will eventually replace seeds in the next release.
- Currently, all non-container values (e.g. int, char*, even custom structs) are interpreted to be containers of length 1. This feature is rarely, if ever, used and will not be available in future releases.
