CHANGES IN VERSION 2.16
-----------------------

NEW FEATURES

    o (v 2.15.1) sortBam() gains support for sorting by tag (byTag) and using
      multiple threads (nThreads). (See
      https://github.com/Bioconductor/Rsamtools/issues/46. ; kriemo)

CHANGES IN VERSION 2.10
-----------------------

DEPRECATED AND DEFUNCT

    o (v 2.9.1) Deprecate applyPileups() in favor of pileup().

CHANGES IN VERSION 2.8
----------------------

NEW FEATURES

    o (v 2.7.2) idxstatsBam works on remote (e.g., http://) files and reports
      unmapped ('seqnames' equal to *) reads. See
      https://support.bioconductor.org/p/9136222.

CHANGES IN VERSION 2.4
----------------------

BUG FIXES

    o (v 2.3.2; from v 2.2.2) Correctly handle '*' ('unknown') RNAME
      during paired-end processing. See
      https://github.com/Bioconductor/Rsamtools/issues/16.

    o (v 2.3.5) Fix regression introduced by v 2.3.2

NEW FEATURES

    o (v 2.3.1) Don't require BAM files to have @SQ lines; allows
      parsing PacBio 'unaligned' BAM files.
      (https://github.com/Bioconductor/Rsamtools/issues/15 ; jayoung)

CHANGES IN VERSION 2.0
----------------------

SIGNIFICANT USER-VISIBLE CHANGES

    o Migrate Rsamtools to Rhtslib. See Rsamtools/migration_notes.md for
      more information about this migration.

    o Remove unused fields from BamRangeIterator

    o Remove BAM header hash init for pileup (already memoized in Rhtslib)

CHANGES IN VERSION 1.34
-----------------------

BUG FIXES

    o (v 1.34.1) indexFa,FaFile-method correctly updates the index path.

CHANGES IN VERSION 1.33
-----------------------

NEW FEATURES

    o (v 1.33.4, 1.33.7) scanBamFlag() gains isSupplementaryAlignment
      support.

BUG FIXES

    o (v 1.33.1) Do not try to grow NULL (not-yet-encountered) tags
      (https://support.bioconductor.org/p/110609/ ; Robert Bradley)

    o (v 1.33.5) Check for corrupt index 
      (https://github.com/Bioconductor/Rsamtools/issues/3 ; kjohnsen)

CHANGES IN VERSION 1.31
-----------------------

BUG FIXES

    o (v.1.31.3) pileup() examples require min_base_quality =
      10. See https://support.bioconductor.org/p/105515/#105553

CHANGES IN VERSION 1.27
-----------------------

BUG FIXES

    o qnameSuffixStart<-(), qnamePrefixEnd<-() accept 'NA' (bug report
      from Peter Hickey).

    o scanBam() accepts a single tag mixing 'Z' and 'A' format. See
      https://support.bioconductor.org/p/94553/

CHANGES IN VERSION 1.25
-----------------------

NEW FEATURES

    o *File and *FileList (e.g., TabixFile, TabixFileList)
      constructors support NA as 'index'.

    o *File and *FileList have accessor method for index.

    o asBam(), asSam() provide default desinations.

    o idxstatsBam() quickly summarizes the number of mapped and
      unmapped reads on each sequence in a BAM file.

SIGNIFICANT USER-VISIBLE CHANGES

    o index() by default returns NA rather than character(), but can be
      controled with asNA argument.

BUG FIXES

    o TabixFileList(TabixFile()) works.

    o *File constructors now check that the file argument is length 1,
      and that the index argument is length 0 or 1.

CHANGES IN VERSION 1.23
-----------------------

NEW FEATURES

    o filterBam can filter one source file into multiple destinations
      by providing a vector of destination files and a list of
      FilterRules.

    o phred2ASCIIOffset() helps translate PHRED encodings (integer or
      character) to ASCII offsets for use in pileup()

BUG FIXES

    o scanBam() fails early when param seqlevels not present in file.

    o Rsamtools.mk for Windows avoids spaces in file paths

CHANGES IN VERSION 1.21
-----------------------

SIGNIFICANT USER-VISIBLE CHANGES

    o pileup adds query_bins arg to give strand-sensitive cycle bin
      behavior; cycle_bins renamed left_bins; negative values allowed
      (including -Inf) to specify bins based on distance from
      end-of-read.

    o mapqFilter allows specification of a mapping quality filter
      threshold

    o PileupParam() now correctly follows samtools with
      min_base_quality=13, min_map_quality=0 (previously, values were
      assigned as 0 and 13, respectively)

    o Support parsing 'B' tags in bam file headers.

BUG FIXES

    o segfault on range iteration introduced 1.19.35, fixed in 1.21.1

    o BamViews parallel evaluation with BatchJobs back-end requires
      named arguments

CHANGES IN VERSION 1.19
-----------------------

SIGNIFICANT USER-VISIBLE CHANGES

    o FaFile accepts a distinct index file

    o Support for cigars > 32767 characters

    o Mate pairs use pos and mpos values calculated modulo target
      length for pairing, facilitating some representations of mates
      on circular chromosomes.

    o scanBam no longer translates mapq '255' to 'NA'

BUG FIXES

    o segfault on file iteration, introduced in 1.19.35, fixed in
      1.19.44

    o scanBam correctly parses '=' and 'X'

CHANGES IN VERSION 1.17.0
-------------------------

NEW FEATURES

    o pileup visits entire file if no 'which' argument specified for
      'ScanBamParam' parameter of pileup. Buffered functionality with
      'yieldSize' available to manage memory consumption when working
      with large BAM files

    o pileup 'read_pos_breaks' parameter renamed to 'cycle_bins':
      cycle_bins allows users to differentiate pileup counts based on
      user-defined regions within a read.

    o pileup uses PileupParam and ScanBamParam instances to calculate
      pileup statistics for a BAM file; returns a data.frame with
      columns summarizing information extracted from alignments
      overlapping each genomic position

    o scanBam,BamSampler-method returns requested and actual
      yieldSize, and total reads

    o seqinfo,BamFileList-method returns the merged seqinfo of each
      BamFile; seqlevels and seqlengths behave similarly.

    o scanBamHeader accepts a 'what' argument to control input of the
      targets and / or text portion of the header, and is much faster
      for BAM files with many rnames.

SIGNIFICANT USER-VISIBLE CHANGES

    o rename PileupParam class and constructor -> ApplyPileupsParam

    o seqinfo,BamFile-method orders levels as they occur in the file,
      reverting a change introduced in Rsamtools version 1.15.28
      (version 1.17.16).

BUG FIXES

    o scanBam(BamSampler(), param=param) with a 'which' argument no
      longer mangles element names, and respects yield size

    o applyPileups checks that seqlevels are identical across files

    o scanFa documentation incorrectly indicated that end coordinates
      beyond the range of the sequence would be truncated; they are an
      error.

    o applyPileups would fail on cigars with insertion followed by
      reference skip, e.g., 2I1024N98M (bug report of Dan Gatti).

CHANGES IN VERSION 1.15.0
-------------------------

NEW FEATURES

    o asSam converts BAM files to SAM files

    o razip, bgzip re-compress directly from .gz files

    o yieldReduce through a BAM or other file, applying a MAP function
      to each chunk and reducing the result to it's final representation

SIGNIFICANT USER-VISIBLE CHANGES

    o bgzip default extension changed to '.bgz'

    o seqinfo,BamFile-method attempts to return seqnames in 'natural'
      order, e.g., chr1, chr2, ...

    o yieldSize now works on BAM files queried with ranges. Successive
      ranges are input until the total number of records first equals
      or exceeds yieldSize..

    o scanFa supports DNA, RNA, and AAStringSet return objects

BUG FIXES

    o scanFa returns correct sequence at the very end of files

    o razip compresses small files

    o applyPileups no longer crashes in the absence of an index file

CHANGES IN VERSION 1.14.0
-------------------------

NEW FEATURES

    o seqinfo(FaFile) returns available information on sequences and
      lengths on Fa (indexed fasta) files.

    o filterBam accepts FilterRules call-backs for arbitrary
      filtering.

    o add isIncomplete,BamFile-method to test for end-of-file

    o add count.mapped.reads to summarizeOverlaps,*,BamFileList-method;
      set to TRUE to collect read and nucleotide counts via countBam.

    o add summarizeOverlaps,*,character-method to count simple file
      paths

    o add sequenceLayer() and stackStringsFromBam()

    o add 'with.which_label' arg to readGAlignmentsFromBam(),
      readGappedReadsFromBam(), readGAlignmentPairsFromBam(), and
      readGAlignmentsListFromBam()

SIGNIFICANT USER-VISIBLE CHANGES

    o rename:
        readBamGappedAlignments() -> readGAlignmentsFromBam() 
        readBamGappedReads() -> readGappedReadsFromBam()
        readBamGappedAlignmentPairs() -> readGAlignmentPairsFromBam()
        readBamGAlignmentsList() -> readGAlignmentsListFromBam()
        makeGappedAlignmentPairs() -> makeGAlignmentPairs()

    o speedup findMateAlignment()

DEPRECATED AND DEFUNCT

    o deprecate readBamGappedAlignments(), readBamGappedReads(),
      readBamGappedAlignmentPairs(), readBamGAlignmentsList(), and
      makeGappedAlignmentPairs()

BUG FIXES

    o scanVcfHeader tolerates records without ID fields, and with
      fields named similar to ID.

    o close razip files only once.

    o report tabix input errors

CHANGES IN VERSION 1.12.0
-------------------------

NEW FEATURES

    o BamSampler draws a random sample from BAM file records, obeying
    any restriction by ScanBamParam().

    o Add argument 'obeyQname' to BamFile. Used with qname-sorted
    Bam files only.

    o Add readBamGAlignmentsList function for reading qname-sorted
    Bam files into a GAlignmentsList object. 

USER-VISIBLE CHANGES

    o bamPath and bamIndicies applied to BamViews returns named
      vectors.

    o 'yieldSize' argument in BamFile represents the number of
    unique qnames when 'obeyQname=TRUE'. 

BUG FIXES

    o completely free razip, bgzip files when done.

    o sortBam, indexBam fail gracefully on non-BAM input.

    o headerTabix on an open TabixFile no longer reads the first
    record

    o scanBcfHeader provides informative error message when header
    line ('#CHROM POS ...') is missing

CHANGES IN VERSION 1.10.0
-------------------------

NEW FEATURES

    o BamFile and TabixFile accept argument yieldSize; repeated calls
      to scanBam and scanTabix return successive yieldSize chunks of
      the file. readBamGappedAlignments, VariantAnnotation::readVcf
      automatically gain support for yield'ing through files.

    o Add getDumpedAlignments(), countDumpedAlignments(), and
      flushDumpedAlignments() low-level utilities for manipulating
      alignments dumped by findMateAlignment().

    o Add quickBamCounts() utility for classifying the records in a BAM file
      according to a set of predefined groups (based on the flag bits) and
      for counting the nb of records in each group.

SIGNIFICANT USER-VISIBLE CHANGES

    o scanBamFlag isValidVendorRead deprecated in favor of
      isNotPassingQualityControls

    o Rename makeGappedAlignmentPairs() arg 'keep.colnames' -> 'use.mcols'.

BUG FIXES

    o close razip, bgzip files when done

    o bamReverseComplement<- failed to return the updated object

    o scanBcfHeader works on remote files

    o allow asBam to work without warnings on header-only SAM files

    o some bug fixes and and small performance improvements to
      findMateAlignment()

    o fix bug in readBamGappedAlignmentPairs() where fields and tags
      specified by the user were not propagated to the returned
      GappedAlignmentPairs object

CHANGES IN VERSION 1.8.0
------------------------

NEW FEATURES

    o Add readBamGappedAlignmentPairs() (plus related utilities
      findMateAlignment() and makeGappedAlignmentPairs()) to read a BAM
      file into a GappedAlignmentPairs object.

SIGNIFICANT USER-VISIBLE CHANGES

    o update samtools to github commit
      dc27682f70713a70d4f31bca652cf78e00757da2

    o Add 'bitnames' arg to bamFlagAsBitMatrix() utility.

    o By default readBamGappedAlignments() and readBamGappedReads() don't
      drop PCR or optical duplicates anymore.

BUG FIXES

    o readBamGappedAlignments handles empty 'tag' fields

    o scanTabix would omit variants overlapping range ends

    o scanFa would segfault on empty files or empty ids

CHANGES IN VERSION 1.6.0
------------------------

NEW FEATURES

    o TabixFile, indexTabix, scanTabix, yieldTabix index (sorted,
    compressed) and parse tabix-indexed files

    o readBamGappedReads(), bamFlagAsBitMatrix(), bamFlagAND()

    o Add use.names and param args to readBamGappedAlignments(); dropped
    which and ... args.

    o PileupFiles, PileupParam, applyPileup for visiting several BAM
    files and calculating pile-ups on each.

    o Provide a zlib for Windows, as R does not currently do this

    o BamFileList, BcfFileList, TabixFileList, FaFileList clases
    extend IRanges::SimpleList, for managings lists of file references 

    o razfFa creates random access compressed fasta files.

    o count and scanBam support input of larger numbers of records; 
    countBam nucleotide count is now numeric() and subject to rounding
    error when large.

    o Update to samtools 0.1.17

    o asBcf and indexBcf coerces VCF files to BCF, and indexes BCF

    o Update to samtools 0.1.18

    o scanVcf parses VCF files; use scanVcf,connection,missing-method
    to stream, scanVcf,TabixFile,*-method to select subsets. Use
    unpackVcf to expand INFO and GENO fields.

SIGNIFICANT USER-VISIBLE CHANGES

    o ScanBamParam argument 'what' defaults to character(0) (nothing)
    rather than scanBamWhat() (everything)

    o bamFlag returns a user-friendly description of flags by default

BUG FIXES

    o scanBam (and readBamGappedAlignments) called with an invalid or
    character(0) index no longer segfaults.

    o scanBcfHeader parses values with embedded commas or =

    o scanFa fails, rather than returns incorrect sequences, when file
    is compressed and file positions are not accessed sequentially

    o scanBcf parses VCF files with no genotype information.

    o scanBam called with the first range having no reads returned
    invalid results for subsequent ranges; introduced in svn r57138

    o scanBamFlag isPrimaryRead changed to isNotPrimaryRead,
    correctly reflecting the meaning of the flag.

CHANGES IN VERSION 1.4.0
------------------------

NEW FEATURES

    o BamFile class allows bam files to be open across calls to
    scanBam and friends. This can be helpful when wanting to avoid
    repeated loading of the index, for instance.

    o BcfFile, scanBcf, scanBcfHeader to parse bcftools' .vcf and .bcf
    files. Note that this implements bcftools notions of vcf and bcf,
    and are not fully compliant with vcf-4.0.

    o asBam converts SAM files to (indexed) BAM files

    o FaFile, indexFa, scanIndexFa, scanFa index and parse (indexed)
    fasta files.

BUG FIXES

    o scanBamFlag isValidVendorRead had reversed TRUE/FALSE logic

    o Attempts to read too many records caught more gracefully.

    o samtools output to fprintf() or calls to exit() are handled more
    gracefully

CHANGES IN VERSION 1.2.0
------------------------

NEW FEATURES

    o Update to samtools 0.1.8

    o Update to samtools svn rev 750 (Mon, 27 Sep 2010)

    o sortBam sorts a BAM file

BUG FIXES

    o Attempts to parse non-existent local files now generate an error

    o Reads whose last nucleotide overlaps the first of a range are
    now scanned / counted.

    o scanning / counting reads late in large Windows files is fast

    o scanBam tag fields of type 'A' parsed correctly

CHANGES IN VERSION 1.0.0
-----------------------

SIGNIFICANT USER-VISIBLE CHANGES

    o scanBam returns minus-strand reads in the manner presented in
    the BAM file, i.e., as though on the positive strand. This occurs
    in revision 0.1.34

    o readBamGappedAlignments replaces readBAMasGappedAlignments

NEW FEATURES

    o ScanBamParam() accepts 'tag' argument for parsing optional fields

    o BamViews can be used with scanBam, countBam,
    readBamGappedAlignments

BUG FIXES

    o No changes classified as 'bug fixes' (package under active
    development)