Andreas Tille · Andreas Tille · Andreas Tille · Andreas Tille · Andreas Tille · Andreas Tille
--- a/.appveyor.yml
+++ b/.appveyor.yml
-# version format.
-# you can use {branch} name in version format too
-# version: 1.0.{build}-{branch}
-version: 'vers.{build}'
-
-# branches to build
-branches:
-    # Blacklist
-    except:
-      - gh-pages
-
-# Do not build on tags (GitHub and BitBucket)
-skip_tags: true
-
-# Skipping commits affecting specific files (GitHub only). More details here: /docs/appveyor-yml
-#skip_commits:
-#  files:
-#    - docs/*
-#    - '**/*.html'
-
-# We use Mingw/Msys, so use pacman for installs
-install:
-  - set HOME=.
-  - set MSYSTEM=MINGW64
-  - set PATH=C:/msys64/usr/bin;C:/msys64/mingw64/bin;%PATH%
-  - set MINGWPREFIX=x86_64-w64-mingw32
-  - "sh -lc \"pacman -S --noconfirm --needed base-devel mingw-w64-x86_64-toolchain mingw-w64-x86_64-zlib mingw-w64-x86_64-bzip2 mingw-w64-x86_64-xz mingw-w64-x86_64-curl\""
-
-build_script:
-  - set HOME=.
-  - set MSYSTEM=MINGW64
-  - set PATH=C:/msys64/usr/bin;C:/msys64/mingw64/bin;%PATH%
-  - "sh -lc \"aclocal && autoheader && autoconf && ./configure && make -j2\""
-
-#build_script:
-#  - make
-
-test_script:
-  - "sh -lc \"make test\""
--- a/INSTALL
+++ b/INSTALL
@@ -149,6 +149,12 @@ various features and specify further optional external requirements:
    by default.  It can be disabled with --disable-lzma, but be aware
    that not all CRAM files may be possible to decode.

+--with-libdeflate
+    Libdeflate is a heavily optimized library for DEFLATE-based compression
+    and decompression.  It also includes a fast crc32 implementation.
+    By default, ./configure will probe for libdeflate and use it if
+    available.  To prevent this, use --without-libdeflate.
+
 The configure script also accepts the usual options and environment variables
 for tuning installation locations and compilers: type './configure --help'
 for details.  For example,
@@ -158,6 +164,16 @@ for details.  For example,
 would specify that HTSlib is to be built with icc and installed into bin,
 lib, etc subdirectories under /opt/icc-compiled.

+If dependencies have been installed in non-standard locations (i.e. not on
+the normal include and library search paths) then the CPPFLAGS and LDFLAGS
+environment variables can be used to set the options needed to find them.
+For example, NetBSD users may use:
+
+    ./configure CPPFLAGS=-I/usr/pkg/include \
+                LDFLAGS='-L/usr/pkg/lib -Wl,-R/usr/pkg/lib'
+
+to allow compiling and linking against dependencies installed via the ports
+collection.

 Installation Locations
 ======================

--- a/Makefile
+++ b/Makefile
@@ -72,6 +72,7 @@ BUILT_TEST_PROGRAMS = \
 	test/hfile \
 	test/sam \
 	test/test_bgzf \
+	test/test_realn \
 	test/test-regidx \
 	test/test_view \
 	test/test-vcf-api \
@@ -377,6 +378,9 @@ test/sam: test/sam.o libhts.a
 test/test_bgzf: test/test_bgzf.o libhts.a
 	$(CC) $(LDFLAGS) -o $@ test/test_bgzf.o libhts.a -lz $(LIBS) -lpthread

+test/test_realn: test/test_realn.o libhts.a
+	$(CC) $(LDFLAGS) -o $@ test/test_realn.o libhts.a $(LIBS) -lpthread
+
 test/test-regidx: test/test-regidx.o libhts.a
 	$(CC) $(LDFLAGS) -o $@ test/test-regidx.o libhts.a $(LIBS) -lpthread

@@ -400,6 +404,7 @@ test/fieldarith.o: test/fieldarith.c config.h $(htslib_sam_h)
 test/hfile.o: test/hfile.c config.h $(htslib_hfile_h) $(htslib_hts_defs_h)
 test/sam.o: test/sam.c config.h $(htslib_hts_defs_h) $(htslib_sam_h) $(htslib_faidx_h) $(htslib_kstring_h)
 test/test_bgzf.o: test/test_bgzf.c $(htslib_bgzf_h) $(htslib_hfile_h)
+test/test-realn.o: test/test_realn.c $(htslib_hts_h) $(htslib_sam_h) $(htslib_faidx_h)
 test/test-regidx.o: test/test-regidx.c config.h $(htslib_regidx_h) $(hts_internal_h)
 test/test_view.o: test/test_view.c config.h $(cram_h) $(htslib_sam_h)
 test/test-vcf-api.o: test/test-vcf-api.c config.h $(htslib_hts_h) $(htslib_vcf_h) $(htslib_kstring_h) $(htslib_kseq_h)
@@ -434,7 +439,7 @@ install: libhts.a $(BUILT_PROGRAMS) $(BUILT_PLUGINS) installdirs install-$(SHLIB
 	if test -n "$(BUILT_PLUGINS)"; then $(INSTALL_PROGRAM) $(BUILT_PLUGINS) $(DESTDIR)$(plugindir); fi
 	$(INSTALL_DATA) htslib/*.h $(DESTDIR)$(includedir)/htslib
 	$(INSTALL_DATA) libhts.a $(DESTDIR)$(libdir)/libhts.a
-	$(INSTALL_MAN) htsfile.1 tabix.1 $(DESTDIR)$(man1dir)
+	$(INSTALL_MAN) bgzip.1 htsfile.1 tabix.1 $(DESTDIR)$(man1dir)
 	$(INSTALL_MAN) faidx.5 sam.5 vcf.5 $(DESTDIR)$(man5dir)

 installdirs:

--- a/NEWS
+++ b/NEWS
+Noteworthy changes in release 1.8 (3rd April 2018)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+* The URL to get sequences from the EBI reference server has been changed
+  to https://.  This is because the EBI no longer serve sequences via
+  plain HTTP - requests to the http:// endpoint just get redirected.
+  HTSlib needs to be linked against libcurl to download https:// URLs,
+  so CRAM users who want to get references from the EBI will need to
+  run configure and ensure libcurl support is enabled using the
+  --enable-libcurl option.
+
+* Added libdeflate as a build option for alternative faster compression and
+  decompression.  Results vary by CPU but compression should be twice as fast
+  and decompression faster.
+
+* It is now possible to set the compression level in bgzip.  (#675; thanks
+  to Nathan Weeks).
+
+* bgzip now gets its own manual page.
+
+* CRAM encoding now stored MD and NM tags verbatim where the reference
+  contains 'N' characters, to work around ambiguities in the SAM
+  specification (samtools #717/762).
+  Also added "store_md" and "store_nm" cram-options for forcing these
+  tags to be stored at all locations.  This is best when combined with
+  a subsequent decode_md=0 option while reading CRAM.
+
+* Multiple CRAM bug fixes, including a fix to free and the subsequent reuse of
+  references with `-T ref.fa`. (#654; reported by Chris Saunders)
+
+* CRAM multi-threading bugs fixed: don't try to call flush on reading;
+  processing of multiple range queries; problems with multi-slice containers.
+
+* Fixed crashes caused when decoding some cramtools produced CRAM files.
+
+* Fixed a couple of minor rANS issues with handling invalid data.
+
+* Fixed bug where probaln_glocal() tried to allocate far more memory than
+  needed when the query sequence was much longer than the reference.  This
+  caused crashes in samtools and bcftools mpileup when used on data with very
+  long reads. (#572, problem reported by Felix Bemm via minimap2).
+
+* sam_prop_realn() now returns -1 (the same value as for unmapped reads)
+  on reads that do not include at least one 'M', 'X' or '=' CIGAR operator,
+  and no longer adds BQ or ZQ tags.  BAQ adjustments are only made to bases
+  covered by these operators so there is no point in trying to align
+  reads that do not have them. (#572)
+
 Noteworthy changes in release 1.7 (26th January 2018)
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


--- a/bcf_sr_sort.c
+++ b/bcf_sr_sort.c
@@ -627,12 +627,18 @@ int bcf_sr_sort_next(bcf_srs_t *readers, sr_sort_t *srt, const char *chr, int mi
    return nret;
 }
 void bcf_sr_sort_remove_reader(bcf_srs_t *readers, sr_sort_t *srt, int i)
+{
+    //vcf_buf is allocated only in bcf_sr_sort_next
+    //So, a call to bcf_sr_add_reader() followed immediately by bcf_sr_remove_reader()
+    //would cause the program to crash in this segment
+    if (srt->vcf_buf)
    {
        free(srt->vcf_buf[i].rec);
        if ( i+1 < srt->nsr )
            memmove(&srt->vcf_buf[i], &srt->vcf_buf[i+1], (srt->nsr - i - 1)*sizeof(vcf_buf_t));
        memset(srt->vcf_buf + srt->nsr - 1, 0, sizeof(vcf_buf_t));
    }
+}
 sr_sort_t *bcf_sr_sort_init(sr_sort_t *srt)
 {
    if ( !srt ) return calloc(1,sizeof(sr_sort_t));

--- a/bgzf.c
+++ b/bgzf.c
@@ -35,6 +35,10 @@
 #include <sys/types.h>
 #include <inttypes.h>

+#ifdef HAVE_LIBDEFLATE
+#include <libdeflate.h>
+#endif
+
 #include "htslib/hts.h"
 #include "htslib/bgzf.h"
 #include "htslib/hfile.h"
@@ -359,6 +363,64 @@ BGZF *bgzf_hopen(hFILE *hfp, const char *mode)
    return fp;
 }

+#ifdef HAVE_LIBDEFLATE
+int bgzf_compress(void *_dst, size_t *dlen, const void *src, size_t slen, int level)
+{
+    if (slen == 0) {
+        // EOF block
+        if (*dlen < 28) return -1;
+        memcpy(_dst, "\037\213\010\4\0\0\0\0\0\377\6\0\102\103\2\0\033\0\3\0\0\0\0\0\0\0\0\0", 28);
+        *dlen = 28;
+        return 0;
+    }
+
+    uint8_t *dst = (uint8_t*)_dst;
+
+    if (level == 0) {
+        // Uncompressed data
+        if (*dlen < slen+5 + BLOCK_HEADER_LENGTH + BLOCK_FOOTER_LENGTH) return -1;
+        dst[BLOCK_HEADER_LENGTH] = 1; // BFINAL=1, BTYPE=00; see RFC1951
+        u16_to_le(slen,  &dst[BLOCK_HEADER_LENGTH+1]); // length
+        u16_to_le(~slen, &dst[BLOCK_HEADER_LENGTH+3]); // ones-complement length
+        memcpy(dst + BLOCK_HEADER_LENGTH+5, src, slen);
+        *dlen = slen+5 + BLOCK_HEADER_LENGTH + BLOCK_FOOTER_LENGTH;
+
+    } else {
+        level = level > 0 ? level : 6; // libdeflate doesn't honour -1 as default
+        // NB levels go up to 12 here.
+        struct libdeflate_compressor *z = libdeflate_alloc_compressor(level);
+        if (!z) return -1;
+
+        // Raw deflate
+        size_t clen =
+            libdeflate_deflate_compress(z, src, slen,
+                                        dst + BLOCK_HEADER_LENGTH,
+                                        *dlen - BLOCK_HEADER_LENGTH - BLOCK_FOOTER_LENGTH);
+
+        if (clen <= 0) {
+            hts_log_error("Call to libdeflate_deflate_compress failed");
+            libdeflate_free_compressor(z);
+            return -1;
+        }
+
+        *dlen = clen + BLOCK_HEADER_LENGTH + BLOCK_FOOTER_LENGTH;
+
+        libdeflate_free_compressor(z);
+    }
+
+    // write the header
+    memcpy(dst, g_magic, BLOCK_HEADER_LENGTH); // the last two bytes are a place holder for the length of the block
+    packInt16(&dst[16], *dlen - 1); // write the compressed length; -1 to fit 2 bytes
+
+    // write the footer
+    uint32_t crc = libdeflate_crc32(0, src, slen);
+    packInt32((uint8_t*)&dst[*dlen - 8], crc);
+    packInt32((uint8_t*)&dst[*dlen - 4], slen);
+    return 0;
+}
+
+#else
+
 int bgzf_compress(void *_dst, size_t *dlen, const void *src, size_t slen, int level)
 {
    uint32_t crc;
@@ -395,6 +457,7 @@ int bgzf_compress(void *_dst, size_t *dlen, const void *src, size_t slen, int le
    packInt32((uint8_t*)&dst[*dlen - 4], slen);
    return 0;
 }
+#endif // HAVE_LIBDEFLATE

 static int bgzf_gzip_compress(BGZF *fp, void *_dst, size_t *dlen, const void *src, size_t slen, int level)
 {
@@ -438,6 +501,28 @@ static int deflate_block(BGZF *fp, int block_length)
    return comp_size;
 }

+#ifdef HAVE_LIBDEFLATE
+
+static int bgzf_uncompress(uint8_t *dst, size_t *dlen, const uint8_t *src, size_t slen) {
+    struct libdeflate_decompressor *z = libdeflate_alloc_decompressor();
+    if (!z) {
+        hts_log_error("Call to libdeflate_alloc_decompressor failed");
+        return -1;
+    }
+
+    int ret = libdeflate_deflate_decompress(z, src, slen, dst, *dlen, dlen);
+    libdeflate_free_decompressor(z);
+
+    if (ret != LIBDEFLATE_SUCCESS) {
+        hts_log_error("Inflate operation failed: %d", ret);
+        return -1;
+    }
+
+    return 0;
+}
+
+#else
+
 static int bgzf_uncompress(uint8_t *dst, size_t *dlen, const uint8_t *src, size_t slen) {
    z_stream zs;
    zs.zalloc = NULL;
@@ -467,6 +552,7 @@ static int bgzf_uncompress(uint8_t *dst, size_t *dlen, const uint8_t *src, size_
    *dlen = *dlen - zs.avail_out;
    return 0;
 }
+#endif // HAVE_LIBDEFLATE

 // Inflate the block in fp->compressed_block into fp->uncompressed_block
 static int inflate_block(BGZF* fp, int block_length)
@@ -482,7 +568,11 @@ static int inflate_block(BGZF* fp, int block_length)
    // Check CRC of uncompressed block matches the gzip header.
    // NB: we may wish to switch out the zlib crc32 for something more performant.
    // See PR#361 and issue#467
+#ifdef HAVE_LIBDEFLATE
+    uint32_t c1 = libdeflate_crc32(0L, (unsigned char *)fp->uncompressed_block, dlen);
+#else
    uint32_t c1 = crc32(0L, (unsigned char *)fp->uncompressed_block, dlen);
+#endif
    uint32_t c2 = le_to_u32((uint8_t *)fp->compressed_block + block_length-8);
    if (c1 != c2) {
        fp->errcode |= BGZF_ERR_CRC;
@@ -1160,7 +1250,7 @@ restart:
            pthread_cond_signal(&mt->command_c);
            pthread_mutex_unlock(&mt->command_m);
            hts_tpool_process_destroy(mt->out_queue);
-            pthread_exit(NULL);
+            return NULL;

        default:
            break;
@@ -1182,7 +1272,7 @@ restart:
        // We tear down the multi-threaded decoder and revert to the old code.
        hts_tpool_dispatch(mt->pool, mt->out_queue, bgzf_nul_func, j);
        hts_tpool_process_ref_decr(mt->out_queue);
-        pthread_exit(&j->errcode);
+        return &j->errcode;
    }

    // Dispatch an empty block so EOF is spotted.
@@ -1193,7 +1283,7 @@ restart:
    hts_tpool_dispatch(mt->pool, mt->out_queue, bgzf_nul_func, j);
    if (j->errcode != 0) {
        hts_tpool_process_destroy(mt->out_queue);
-        pthread_exit(&j->errcode);
+        return &j->errcode;
    }

    // We hit EOF so can stop reading, but we may get a subsequent
@@ -1224,10 +1314,9 @@ restart:
            pthread_cond_signal(&mt->command_c);
            pthread_mutex_unlock(&mt->command_m);
            hts_tpool_process_destroy(mt->out_queue);
-            pthread_exit(NULL);
+            return NULL;
        }
    }
-    return NULL;
 }

 int bgzf_thread_pool(BGZF *fp, hts_tpool *pool, int qsize) {
@@ -1452,7 +1541,7 @@ ssize_t bgzf_block_write(BGZF *fp, const void *data, size_t length)
    uint64_t ublock_size; // amount of uncompressed data to be fed into next block
    while (remaining > 0) {
        current_block = fp->idx->moffs - fp->idx->noffs;
-        ublock_size = fp->idx->offs[current_block+1].uaddr-fp->idx->offs[current_block].uaddr;
+        ublock_size = current_block + 1 < fp->idx->moffs ? fp->idx->offs[current_block+1].uaddr-fp->idx->offs[current_block].uaddr : BGZF_MAX_BLOCK_SIZE;
        uint8_t* buffer = (uint8_t*)fp->uncompressed_block;
        int copy_length = ublock_size - fp->block_offset;
        if (copy_length > remaining) copy_length = remaining;
@@ -1462,6 +1551,7 @@ ssize_t bgzf_block_write(BGZF *fp, const void *data, size_t length)
        remaining -= copy_length;
        if (fp->block_offset == ublock_size) {
            if (lazy_flush(fp) != 0) return -1;
+            if (fp->idx->noffs > 0)
                fp->idx->noffs--;  // decrement noffs to track the blocks
        }
    }

--- a/bgzip.1
+++ b/bgzip.1
+.TH bgzip 1 "3 April 2018" "htslib-1.8" "Bioinformatics tools"
+.SH NAME
+.PP
+bgzip \- Block compression/decompression utility
+.\"
+.\" Copyright (C) 2009-2011 Broad Institute.
+.\" Copyright (C) 2018 Genome Research Limited.
+.\"
+.\" Author: Heng Li <lh3@sanger.ac.uk>
+.\"
+.\" Permission is hereby granted, free of charge, to any person obtaining a
+.\" copy of this software and associated documentation files (the "Software"),
+.\" to deal in the Software without restriction, including without limitation
+.\" the rights to use, copy, modify, merge, publish, distribute, sublicense,
+.\" and/or sell copies of the Software, and to permit persons to whom the
+.\" Software is furnished to do so, subject to the following conditions:
+.\"
+.\" The above copyright notice and this permission notice shall be included in
+.\" all copies or substantial portions of the Software.
+.\"
+.\" THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+.\" IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+.\" FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+.\" THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+.\" LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+.\" FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+.\" DEALINGS IN THE SOFTWARE.
+.\"
+.
+.\" For code blocks and examples (cf groff's Ultrix-specific man macros)
+.de EX
+
+.  in +\\$1
+.  nf
+.  ft CR
+..
+.de EE
+.  ft
+.  fi
+.  in
+
+..
+.SH SYNOPSIS
+.PP
+.B bgzip
+.RB [ -cdfhir ]
+.RB [ -b
+.IR virtualOffset ]
+.RB [ -I
+.IR index_name ]
+.RB [ -l
+.IR compression_level ]
+.RB [ -s
+.IR size ]
+.RB [ -@
+.IR threads ]
+.RI [ file ]
+.PP
+.SH DESCRIPTION
+.PP
+Bgzip compresses files in a similar manner to, and compatible with, gzip(1).
+The file is compressed into a series of small (less than 64K) 'BGZF' blocks.
+This allows indexes to be built against the compressed file and used to
+retrieve portions of the data without having to decompress the entire file. 
+
+If no files are specified on the command line, bgzip will compress (or
+decompress if the -d option is used) standard input to standard output.
+If a file is specified, it will be compressed (or decompressed with -d).
+If the -c option is used, the result will be written to standard output,
+otherwise when compressing bgzip will write to a new file with a .gz
+suffix and remove the original.  When decompressing the input file must
+have a .gz suffix, which will be removed to make the output name.  Again
+after decompression completes the input file will be removed.
+
+.SH OPTIONS
+.TP 10
+.BI "-b, --offset " INT
+Decompress to standard output from virtual file position (0-based uncompressed
+offset).
+Implies -c and -d.
+.TP
+.B "-c, --stdout"
+Write to standard output, keep original files unchanged.
+.TP
+.B "-d, --decompress"
+Decompress.
+.TP
+.B "-f, --force"
+Overwrite files without asking.
+.TP
+.B "-h, --help"
+Displays a help message.
+.TP
+.B "-i, --index"
+Create a BGZF index while compressing.
+Unless the -I option is used, this will have the name of the compressed
+file with .gzi appended to it.
+.TP
+.BI "-I, --index-name " FILE
+Index file name.
+.TP
+.BI "-l, --compress-level " INT
+Compression level to use when compressing.
+From 0 to 9, or -1 for the default level set by the compression library. [-1]
+.TP
+.B "-r, --reindex"
+Rebuild the index on an existing compressed file.
+.TP
+.B "-g, --rebgzip"
+Try to use an existing index to create a compressed file with matching
+block offsets.
+Note that this assumes that the same compression library and level are in use
+as when making the original file.
+Don't use it unless you know what you're doing.
+.TP
+.BI "-s, --size " INT
+Decompress INT bytes (uncompressed size) to standard output.
+Implies -c.
+.TP
+.BI "-@, --threads " INT
+Number of threads to use [1].
+.PP
+
+.SH BGZF FORMAT
+The BGZF format written by bgzip is described in the SAM format specification
+available from http://samtools.github.io/hts-specs/SAMv1.pdf.
+
+It makes use of a gzip feature which allows compressed files to be
+concatenated.
+The input data is divided into blocks which are no larger than 64 kilobytes
+both before and after compression (including compression headers).
+Each block is compressed into a gzip file.
+The gzip header includes an extra sub-field with identifier 'BC' and the length
+of the compressed block, including all headers.
+
+.SH GZI FORMAT
+The index format is a binary file listing pairs of compressed and
+uncompressed offsets in a BGZF file.
+Each compressed offset points to the start of a BGZF block.
+The uncompressed offset is the corresponding location in the uncompressed
+data stream.
+
+All values are stored as little-endian 64-bit unsigned integers.
+
+The file contents are:
+.EX 4
+uint64_t number_entries
+.EE
+followed by number_entries pairs of:
+.EX 4
+uint64_t compressed_offset
+uint64_t uncompressed_offset
+.EE
+
+.SH EXAMPLES
+.EX 4
+# Compress stdin to stdout
+bgzip < /usr/share/dict/words > /tmp/words.gz
+
+# Make a .gzi index
+bgzip -r /tmp/words.gz
+
+# Extract part of the data using the index
+bgzip -b 367635 -s 4 /tmp/words.gz 
+
+# Uncompress the whole file, removing the compressed copy
+bgzip -d /tmp/words.gz
+.EE
+
+.SH AUTHOR
+.PP
+The BGZF library was originally implemented by Bob Handsaker and modified
+by Heng Li for remote file access and in-memory caching.
+
+.SH SEE ALSO
+.PP
+.BR gzip (1), tabix (1)
--- a/bgzip.c
+++ b/bgzip.c
@@ -80,6 +80,7 @@ static int bgzip_main_usage(void)
    fprintf(stderr, "   -h, --help                 give this help\n");
    fprintf(stderr, "   -i, --index                compress and create BGZF index\n");
    fprintf(stderr, "   -I, --index-name FILE      name of BGZF index file [file.gz.gzi]\n");
+    fprintf(stderr, "   -l, --compress-level INT   Compression level to use when compressing; 0 to 9, or -1 for default [-1]\n");
    fprintf(stderr, "   -r, --reindex              (re)index compressed file\n");
    fprintf(stderr, "   -g, --rebgzip              use an index file to bgzip a file\n");
    fprintf(stderr, "   -s, --size INT             decompress INT bytes (uncompressed size)\n");
@@ -90,7 +91,7 @@ static int bgzip_main_usage(void)

 int main(int argc, char **argv)
 {
-    int c, compress, pstdout, is_forced, index = 0, rebgzip = 0, reindex = 0;
+    int c, compress, compress_level = -1, pstdout, is_forced, index = 0, rebgzip = 0, reindex = 0;
    BGZF *fp;
    void *buffer;
    long start, end, size;
@@ -106,6 +107,7 @@ int main(int argc, char **argv)
        {"force", no_argument, NULL, 'f'},
        {"index", no_argument, NULL, 'i'},
        {"index-name", required_argument, NULL, 'I'},
+        {"compress-level", required_argument, NULL, 'l'},
        {"reindex", no_argument, NULL, 'r'},
        {"rebgzip",no_argument,NULL,'g'},
        {"size", required_argument, NULL, 's'},
@@ -115,7 +117,7 @@ int main(int argc, char **argv)
    };

    compress = 1; pstdout = 0; start = 0; size = -1; end = -1; is_forced = 0;
-    while((c  = getopt_long(argc, argv, "cdh?fb:@:s:iI:gr",loptions,NULL)) >= 0){
+    while((c  = getopt_long(argc, argv, "cdh?fb:@:s:iI:l:gr",loptions,NULL)) >= 0){
        switch(c){
        case 'd': compress = 0; break;
        case 'c': pstdout = 1; break;
@@ -124,6 +126,7 @@ int main(int argc, char **argv)
        case 'f': is_forced = 1; break;
        case 'i': index = 1; break;
        case 'I': index_fname = optarg; break;
+        case 'l': compress_level = atol(optarg); break;
        case 'g': rebgzip = 1; break;
        case 'r': reindex = 1; compress = 0; break;
        case '@': threads = atoi(optarg); break;
@@ -144,6 +147,17 @@ int main(int argc, char **argv)
    if (compress == 1) {
        struct stat sbuf;
        int f_src = fileno(stdin);
+        char out_mode[3] = "w\0";
+        char out_mode_exclusive[4] = "wx\0";
+
+        if (compress_level < -1 || compress_level > 9) {
+            fprintf(stderr, "[bgzip] Invalid compress-level: %d\n", compress_level);
+            return 1;
+        }
+        if (compress_level >= 0) {
+            out_mode[1] = compress_level + '0';
+            out_mode_exclusive[2] = compress_level + '0';
+        }

        if ( argc>optind )
        {
@@ -159,15 +173,15 @@ int main(int argc, char **argv)
            }

            if (pstdout)
-                fp = bgzf_open("-", "w");
+                fp = bgzf_open("-", out_mode);
            else
            {
                char *name = malloc(strlen(argv[optind]) + 5);
                strcpy(name, argv[optind]);
                strcat(name, ".gz");
-                fp = bgzf_open(name, is_forced? "w" : "wx");
+                fp = bgzf_open(name, is_forced? out_mode : out_mode_exclusive);
                if (fp == NULL && errno == EEXIST && confirm_overwrite(name))
-                    fp = bgzf_open(name, "w");
+                    fp = bgzf_open(name, out_mode);
                if (fp == NULL) {
                    fprintf(stderr, "[bgzip] can't create %s: %s\n", name, strerror(errno));
                    free(name);
@@ -184,7 +198,7 @@ int main(int argc, char **argv)
            return 1;
        }
        else
-            fp = bgzf_open("-", "w");
+            fp = bgzf_open("-", out_mode);

        if ( index && rebgzip )
        {

--- a/configure.ac
+++ b/configure.ac
@@ -64,9 +64,9 @@ m4_ifdef([PKG_PROG_PKG_CONFIG], [PKG_PROG_PKG_CONFIG], [PKG_CONFIG=""])

 need_crypto=no
 pc_requires=
-static_LDFLAGS=
-static_LIBS='-lz -lm'
-private_LIBS=
+static_LDFLAGS=$LDFLAGS
+static_LIBS='-lpthread -lz -lm'
+private_LIBS=$LDFLAGS

 AC_ARG_ENABLE([bz2],
  [AS_HELP_STRING([--disable-bz2],
@@ -97,6 +97,11 @@ AC_ARG_ENABLE([plugins],
  [], [enable_plugins=no])
 AC_SUBST(enable_plugins)

+AC_ARG_WITH([libdeflate],
+  [AS_HELP_STRING([--with-libdeflate],
+                  [use libdeflate for faster crc and deflate algorithms])],
+  [], [with_libdeflate=check])
+
 AC_ARG_WITH([plugin-dir],
  [AS_HELP_STRING([--with-plugin-dir=DIR],
                  [plugin installation location [LIBEXECDIR/htslib]])],
@@ -257,6 +262,26 @@ produced elsewhere unreadable) or resolve this error to build HTSlib.])
  static_LIBS="$static_LIBS -llzma"
 fi

+AS_IF([test "x$with_libdeflate" != "xno"],
+  [libdeflate=ok
+   AC_CHECK_HEADER([libdeflate.h],[],[libdeflate='missing header'],[;])
+   AC_CHECK_LIB([deflate], [libdeflate_deflate_compress],[],[libdeflate='missing library'])
+   AS_IF([test "$libdeflate" = "ok"],
+    [AC_DEFINE([HAVE_LIBDEFLATE], 1, [Define if libdeflate is available.])
+     private_LIBS="$private_LIBS -ldeflate"
+     static_LIBS="$static_LIBS -ldeflate"],
+    [AS_IF([test "x$with_libdeflate" != "xcheck"],
+       [AC_MSG_ERROR([libdeflate development files not found: $libdeflate
+
+You requested libdeflate, but do not have the required header / library
+files.  The source for libdeflate is available from
+<https://github.com/ebiggers/libdeflate>.  You may have to adjust
+search paths in CPPFLAGS and/or LDFLAGS if the header and library
+are not currently on them.
+
+Either configure with --without-libdeflate or resolve this error to build
+HTSlib.])])])])
+
 libcurl=disabled
 if test "$enable_libcurl" != no; then
  AC_CHECK_LIB([curl], [curl_easy_pause],

--- a/cram/cram_codecs.c
+++ b/cram/cram_codecs.c
@@ -311,10 +311,6 @@ static char *cram_extract_block(cram_block *b, int size) {
 * ---------------------------------------------------------------------------
 * EXTERNAL
 */
-static void cram_external_decode_reset(cram_codec *c) {
-    c->external.b = NULL;
-}
-
 int cram_external_decode_int(cram_slice *slice, cram_codec *c,
 			     cram_block *in, char *out, int *out_size) {
    int l;
@@ -322,10 +318,7 @@ int cram_external_decode_int(cram_slice *slice, cram_codec *c,
    cram_block *b;

    /* Find the external block */
-    if (!(b = c->external.b)){
    b = cram_get_block_by_id(slice, c->external.content_id);
-	c->external.b = b;
-    }
    if (!b)
        return *out_size?-1:0;

@@ -345,10 +338,7 @@ int cram_external_decode_char(cram_slice *slice, cram_codec *c,
    cram_block *b;

    /* Find the external block */
-    if (!(b = c->external.b)){
    b = cram_get_block_by_id(slice, c->external.content_id);
-	c->external.b = b;
-    }
    if (!b)
        return *out_size?-1:0;

@@ -369,10 +359,7 @@ static int cram_external_decode_block(cram_slice *slice, cram_codec *c,
    cram_block *b = NULL;

    /* Find the external block */
-    if (!(b = c->external.b)){
    b = cram_get_block_by_id(slice, c->external.content_id);
-	c->external.b = b;
-    }
    if (!b)
        return *out_size?-1:0;

@@ -416,8 +403,6 @@ cram_codec *cram_external_decode_init(char *data, int size,
        goto malformed;

    c->external.type = option;
-    c->external.b = NULL;
-    c->reset = cram_external_decode_reset;

    return c;

@@ -495,8 +480,6 @@ cram_codec *cram_external_encode_init(cram_stats *st,
 * ---------------------------------------------------------------------------
 * BETA
 */
-void cram_nop_decode_reset(cram_codec *c) {}
-
 int cram_beta_decode_int(cram_slice *slice, cram_codec *c, cram_block *in, char *out, int *out_size) {
    int32_t *out_i = (int32_t *)out;
    int i, n = *out_size;
@@ -575,8 +558,6 @@ cram_codec *cram_beta_decode_init(char *data, int size,
        return NULL;
    }

-    c->reset = cram_nop_decode_reset;
-
    return c;
 }

@@ -769,8 +750,6 @@ cram_codec *cram_subexp_decode_init(char *data, int size,
 	return NULL;
    }

-    c->reset = cram_nop_decode_reset;
-
    return c;
 }

@@ -833,8 +812,6 @@ cram_codec *cram_gamma_decode_init(char *data, int size,
    if (cp - data != size)
        goto malformed;

-    c->reset = cram_nop_decode_reset;
-
    return c;

 malformed:
@@ -1031,8 +1008,6 @@ cram_codec *cram_huffman_decode_init(char *data, int size,
    if (i != ncodes)
        goto malformed;

-    h->reset = cram_nop_decode_reset;
-
    if (ncodes == 0) {
 	/* NULL huffman stream.  Ensure it returns an error if
           anything tries to use it. */
@@ -1472,11 +1447,6 @@ void cram_byte_array_len_decode_free(cram_codec *c) {
    free(c);
 }

-static void cram_byte_array_len_decode_reset(cram_codec *c) {
-    c->byte_array_len.len_codec->reset(c->byte_array_len.len_codec);
-    c->byte_array_len.val_codec->reset(c->byte_array_len.val_codec);
-}
-
 cram_codec *cram_byte_array_len_decode_init(char *data, int size,
 					    enum cram_external_type option,
 					    int version) {
@@ -1517,8 +1487,6 @@ cram_codec *cram_byte_array_len_decode_init(char *data, int size,
    if (cp - data != size)
        goto malformed;

-    c->reset = cram_byte_array_len_decode_reset;
-
    return c;

 malformed:
@@ -1617,20 +1585,13 @@ cram_codec *cram_byte_array_len_encode_init(cram_stats *st,
 * ---------------------------------------------------------------------------
 * BYTE_ARRAY_STOP
 */
-static void cram_byte_array_stop_decode_reset(cram_codec *c) {
-    c->byte_array_stop.b = NULL;
-}
-
 static int cram_byte_array_stop_decode_char(cram_slice *slice, cram_codec *c,
 					    cram_block *in, char *out,
 					    int *out_size) {
    char *cp, ch;
    cram_block *b = NULL;

-    if (!(b = c->byte_array_stop.b)){
    b = cram_get_block_by_id(slice, c->byte_array_stop.content_id);
-	c->byte_array_stop.b = b;
-    }
    if (!b)
        return *out_size?-1:0;

@@ -1668,10 +1629,7 @@ int cram_byte_array_stop_decode_block(cram_slice *slice, cram_codec *c,
    char *cp, *out_cp, *cp_end;
    char stop;

-    if (!(b = c->byte_array_stop.b)){
    b = cram_get_block_by_id(slice, c->byte_array_stop.content_id);
-	c->byte_array_stop.b = b;
-    }
    if (!b)
        return *out_size?-1:0;

@@ -1746,9 +1704,6 @@ cram_codec *cram_byte_array_stop_decode_init(char *data, int size,
    if ((char *)cp - data != size)
        goto malformed;

-    c->byte_array_stop.b = NULL;
-    c->reset = cram_byte_array_stop_decode_reset;
-
    return c;

 malformed:

--- a/cram/cram_codecs.h
+++ b/cram/cram_codecs.h
@@ -84,7 +84,6 @@ typedef struct {
 typedef struct {
    int32_t content_id;
    enum cram_external_type type;
-    cram_block *b;
 } cram_external_decoder;

 typedef struct {
@@ -95,7 +94,6 @@ typedef struct {
 typedef struct {
    unsigned char stop;
    int32_t content_id;
-    cram_block *b;
 } cram_byte_array_stop_decoder;

 typedef struct {
@@ -110,6 +108,9 @@ typedef struct {
 /*
 * A generic codec structure.
 */
+#ifdef __SUNPRO_C
+#  pragma error_messages(off, E_ANONYMOUS_UNION_DECL)
+#endif
 typedef struct cram_codec {
    enum cram_encoding codec;
    cram_block *out;
@@ -120,7 +121,6 @@ typedef struct cram_codec {
 		  char *in, int in_size);
    int (*store)(struct cram_codec *codec, cram_block *b, char *prefix,
 		 int version);
-    void (*reset)(struct cram_codec *codec); // used between slices in a container

    union {
 	cram_huffman_decoder         huffman;
@@ -138,6 +138,9 @@ typedef struct cram_codec {
 	cram_beta_decoder            e_beta;
    };
 } cram_codec;
+#ifdef __SUNPRO_C
+#  pragma error_messages(default, E_ANONYMOUS_UNION_DECL)
+#endif

 const char *cram_encoding2str(enum cram_encoding t);


--- a/cram/cram_decode.c
+++ b/cram/cram_decode.c
--- a/cram/cram_decode.h
+++ b/cram/cram_decode.h
@@ -105,6 +105,11 @@ int cram_decode_slice(cram_fd *fd, cram_container *c, cram_slice *s,
 		      SAM_hdr *hdr);


+/*
+ * Drains and frees the decode read-queue for a multi-threaded reader.
+ */
+void cram_drain_rqueue(cram_fd *fd);
+
 #ifdef __cplusplus
 }
 #endif

--- a/cram/cram_encode.c
+++ b/cram/cram_encode.c
@@ -1152,11 +1152,11 @@ static int lossy_read_names(cram_fd *fd, cram_container *c, cram_slice *s,
 	    uint64_t i64;
 	    struct {
 		int32_t e,c; // expected & observed counts.
-	    };
+	    } counts;
 	} u;

 	e = expected_template_count(b);
-	u.e = e; u.c = 1;
+	u.counts.e = e; u.counts.c = 1;

 	k = kh_put(m_s2u64, names, bam_name(b), &n);
 	if (n == -1)
@@ -1165,13 +1165,13 @@ static int lossy_read_names(cram_fd *fd, cram_container *c, cram_slice *s,
 	if (n == 0) {
 	    // not a new name
 	    u.i64 = kh_val(names, k);
-	    if (u.e != e) {
+	    if (u.counts.e != e) {
 		// different expectation or already hit the max
 		//fprintf(stderr, "Err computing no. %s recs\n", bam_name(b));
 		kh_val(names, k) = 0;
 	    } else {
-		u.c++;
-		if (u.e == u.c) {
+		u.counts.c++;
+		if (u.counts.e == u.counts.c) {
 		    // Reached expected count.
 		    kh_val(names, k) = -1;
 		} else {
@@ -2072,7 +2072,8 @@ static char *cram_encode_aux_1_0(cram_fd *fd, bam_seq_t *b, cram_container *c,
 *         NULL on failure or no rg present (FIXME)
 */
 static char *cram_encode_aux(cram_fd *fd, bam_seq_t *b, cram_container *c,
-			     cram_slice *s, cram_record *cr) {
+			     cram_slice *s, cram_record *cr,
+			     int verbatim_NM, int verbatim_MD) {
    char *aux, *orig, *rg = NULL;
    int aux_size = bam_get_l_aux(b);
    cram_block *td_b = c->comp_hdr->TD_blk;
@@ -2095,7 +2096,7 @@ static char *cram_encode_aux(cram_fd *fd, bam_seq_t *b, cram_container *c,

 	// MD:Z
 	if (aux[0] == 'M' && aux[1] == 'D' && aux[2] == 'Z') {
-	    if (cr->len && !fd->no_ref && !(cr->flags & BAM_FUNMAP)) {
+	    if (cr->len && !fd->no_ref && !(cr->flags & BAM_FUNMAP) && !verbatim_MD) {
 		while (*aux++);
 		continue;
 	    }
@@ -2103,7 +2104,7 @@ static char *cram_encode_aux(cram_fd *fd, bam_seq_t *b, cram_container *c,

 	// NM:i
 	if (aux[0] == 'N' && aux[1] == 'M') {
-	    if (cr->len && !fd->no_ref && !(cr->flags & BAM_FUNMAP)) {
+	    if (cr->len && !fd->no_ref && !(cr->flags & BAM_FUNMAP) && !verbatim_NM) {
 		switch(aux[2]) {
 		case 'A': case 'C': case 'c': aux+=4; break;
 		case 'S': case 's':           aux+=5; break;
@@ -2242,7 +2243,9 @@ static char *cram_encode_aux(cram_fd *fd, bam_seq_t *b, cram_container *c,
 	    m->codec = c;

 	    // Link to fd-global tag metrics
+	    pthread_mutex_lock(&fd->metrics_lock);
 	    m->m = k_global ? (cram_metrics *)kh_val(fd->tags_used, k_global) : NULL;
+	    pthread_mutex_unlock(&fd->metrics_lock);
 	}

 	cram_tag_map *tm = (cram_tag_map *)kh_val(c->tags_used, k);
@@ -2514,6 +2517,14 @@ static int process_one_read(cram_fd *fd, cram_container *c,
    char *cp, *rg;
    char *ref, *seq, *qual;

+    // Any places with N in seq and/or reference can lead to ambiguous
+    // interpretation of the SAM NM:i tag.  So we store these verbatim
+    // to ensure valid data round-trips the same regardless of who
+    // defines it as valid.
+    // Similarly when alignments go beyond end of the reference.
+    int verbatim_NM = fd->store_nm;
+    int verbatim_MD = fd->store_md;
+
    // FIXME: multi-ref containers

    ref = c->ref;
@@ -2522,33 +2533,6 @@ static int process_one_read(cram_fd *fd, cram_container *c,

    //fprintf(stderr, "%s => %d\n", rg ? rg : "\"\"", cr->rg);

-    // Fields to resolve later
-    //cr->mate_line;    // index to another cram_record
-    //cr->mate_flags;   // MF
-    //cr->ntags;        // TC
-    cr->ntags      = 0; //cram_stats_add(c->stats[DS_TC], cr->ntags);
-    if (CRAM_MAJOR_VERS(fd->version) == 1)
-	rg = cram_encode_aux_1_0(fd, b, c, s, cr);
-    else
-	rg = cram_encode_aux(fd, b, c, s, cr);
-
-    //cr->aux_size = b->blk_size - ((char *)bam_aux(b) - (char *)&bam_ref(b));
-    //cr->aux = DSTRING_LEN(s->aux_ds);
-    //dstring_nappend(s->aux_ds, bam_aux(b), cr->aux_size);
-
-    /* Read group, identified earlier */
-    if (rg) {
-	SAM_RG *brg = sam_hdr_find_rg(fd->header, rg);
-	cr->rg = brg ? brg->id : -1;
-    } else if (CRAM_MAJOR_VERS(fd->version) == 1) {
-	SAM_RG *brg = sam_hdr_find_rg(fd->header, "UNKNOWN");
-	assert(brg);
-    } else {
-	cr->rg = -1;
-    }
-    cram_stats_add(c->stats[DS_RG], cr->rg);
-
-    
    cr->ref_id      = bam_ref(b);  cram_stats_add(c->stats[DS_RI], cr->ref_id);
    cram_stats_add(c->stats[DS_BF], fd->cram_flag_swap[cr->flags & 0xfff]);

@@ -2689,6 +2673,8 @@ static int process_one_read(cram_fd *fd, cram_container *c,
 			return -1;
 		    }
 		    for (l = 0; l < end; l++) {
+			if (rp[l] == 'N' && sp[l] == 'N')
+			    verbatim_NM = verbatim_MD = 1;
 			if (rp[l] != sp[l]) {
 			    if (!sp[l])
 				break;
@@ -2737,6 +2723,7 @@ static int process_one_read(cram_fd *fd, cram_container *c,
 			}
 		    } else {
 			/* off end of sequence or non-ref based output */
+			verbatim_NM = verbatim_MD = 1;
 			for (; l < cig_len && seq[spos]; l++, spos++) {
 			    if (cram_add_base(fd, c, s, cr, spos,
 					      seq[spos], qual[spos]))
@@ -2746,6 +2733,7 @@ static int process_one_read(cram_fd *fd, cram_container *c,
 		    apos += cig_len;
 		} else if (!cr->len) {
 		    /* Seq "*" */
+		    verbatim_NM = verbatim_MD = 1;
 		    apos += cig_len;
 		    spos += cig_len;
 		}
@@ -2832,6 +2820,24 @@ static int process_one_read(cram_fd *fd, cram_container *c,
 	fake_qual = 0;
    }

+    cr->ntags      = 0; //cram_stats_add(c->stats[DS_TC], cr->ntags);
+    if (CRAM_MAJOR_VERS(fd->version) == 1)
+	rg = cram_encode_aux_1_0(fd, b, c, s, cr);
+    else
+	rg = cram_encode_aux(fd, b, c, s, cr, verbatim_NM, verbatim_MD);
+
+    /* Read group, identified earlier */
+    if (rg) {
+	SAM_RG *brg = sam_hdr_find_rg(fd->header, rg);
+	cr->rg = brg ? brg->id : -1;
+    } else if (CRAM_MAJOR_VERS(fd->version) == 1) {
+	SAM_RG *brg = sam_hdr_find_rg(fd->header, "UNKNOWN");
+	assert(brg);
+    } else {
+	cr->rg = -1;
+    }
+    cram_stats_add(c->stats[DS_RG], cr->rg);
+
    /*
     * Append to the qual block now. We do this here as
     * cram_add_substitution() can generate BA/QS events which need to 

--- a/cram/cram_external.c
+++ b/cram/cram_external.c
@@ -142,10 +142,10 @@ static int cram_block_compression_hdr_set_DS(cram_block_compression_hdr *ch,
 	return 0;

    default:
-	return -1;
+	break;
    }

-    return 0;
+    return -1;
 }

 int cram_block_compression_hdr_set_rg(cram_block_compression_hdr *ch, int new_rg) {

--- a/cram/cram_index.c
+++ b/cram/cram_index.c
@@ -463,36 +463,57 @@ cram_index *cram_index_last(cram_fd *fd, int refid, cram_index *from) {
 * and then read from then on, skipping decoding of slices and/or
 * whole containers when they don't overlap the specified cram_range.
 *
+ * This function also updates the cram_fd range field.
+ *
 * Returns 0 on success
 *        -1 on general failure
 *        -2 on no-data (empty chromosome)
 */
 int cram_seek_to_refpos(cram_fd *fd, cram_range *r) {
+    int ret = 0;
    cram_index *e;

-    if (r->refid == HTS_IDX_NONE)
-	return -2;
+    if (r->refid == HTS_IDX_NONE) {
+	ret = -2; goto err;
+    }

    // Ideally use an index, so see if we have one.
    if ((e = cram_index_query(fd, r->refid, r->start, NULL))) {
-	if (0 != cram_seek(fd, e->offset, SEEK_SET))
-	    if (0 != cram_seek(fd, e->offset - fd->first_container, SEEK_CUR))
-		return -1;
+	if (0 != cram_seek(fd, e->offset, SEEK_SET)) {
+	    if (0 != cram_seek(fd, e->offset - fd->first_container, SEEK_CUR)) {
+		ret = -1; goto err;
+	    }
+	}
    } else {
 	// Absent from index, but this most likely means it simply has no data.
-	return -2;
+	ret = -2; goto err;
    }

+    pthread_mutex_lock(&fd->range_lock);
+    fd->range = *r;
    if (r->refid == HTS_IDX_START || r->refid == HTS_IDX_REST)
 	fd->range.refid = -2; // special case in cram_next_slice
+    pthread_mutex_unlock(&fd->range_lock);

    if (fd->ctr) {
 	cram_free_container(fd->ctr);
+	if (fd->ctr_mt && fd->ctr_mt != fd->ctr)
+	    cram_free_container(fd->ctr_mt);
 	fd->ctr = NULL;
+	fd->ctr_mt = NULL;
 	fd->ooc = 0;
+	fd->eof = 0;
    }

    return 0;
+
+ err:
+    // It's unlikely fd->range will be accessed after EOF or error,
+    // but this maintains identical behaviour to the previous code.
+    pthread_mutex_lock(&fd->range_lock);
+    fd->range = *r;
+    pthread_mutex_unlock(&fd->range_lock);
+    return ret;
 }



--- a/cram/cram_io.c
+++ b/cram/cram_io.c
@@ -69,6 +69,11 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #include <time.h>
 #include <stdint.h>

+#ifdef HAVE_LIBDEFLATE
+#include <libdeflate.h>
+#define crc32(a,b,c) libdeflate_crc32((a),(b),(c))
+#endif
+
 #include "cram/cram.h"
 #include "cram/os.h"
 #include "htslib/hts.h"
@@ -2036,7 +2041,7 @@ static int cram_populate_ref(cram_fd *fd, int id, ref_entry *r) {
 	 * If we have no ref path, we use the EBI server.
 	 * However to avoid spamming it we require a local ref cache too.
 	 */
-	ref_path = "http://www.ebi.ac.uk:80/ena/cram/md5/%s";
+	ref_path = "https://www.ebi.ac.uk/ena/cram/md5/%s";
 	if (!local_cache || *local_cache == '\0') {
 	    const char *extra;
 	    const char *base = get_cache_basedir(&extra);
@@ -2257,7 +2262,7 @@ static void cram_ref_decr_locked(refs_t *r, int id) {
 		RP("%d FREE REF %d (%p)\n", gettid(),
 		   r->last_id, r->ref_id[r->last_id]->seq);
 		ref_entry_free_seq(r->ref_id[r->last_id]);
-		r->ref_id[r->last_id]->length = 0;
+		if (r->ref_id[r->last_id]->is_md5) r->ref_id[r->last_id]->length = 0;
 	    }
 	}
 	r->last_id = id;
@@ -2741,13 +2746,23 @@ void cram_free_container(cram_container *c) {
    if (c->comp_hdr_block)
 	cram_free_block(c->comp_hdr_block);

+    // Free the slices; filled out by encoder only
    if (c->slices) {
-	for (i = 0; i < c->max_slice; i++)
+        for (i = 0; i < c->max_slice; i++) {
            if (c->slices[i])
                cram_free_slice(c->slices[i]);
+            if (c->slices[i] == c->slice)
+                c->slice = NULL;
+        }
        free(c->slices);
    }

+    // Free the current slice; set by both encoder & decoder
+    if (c->slice) {
+	cram_free_slice(c->slice);
+	c->slice = NULL;
+    }
+
    for (id = DS_RN; id < DS_TN; id++)
 	if (c->stats[id]) cram_stats_free(c->stats[id]);

@@ -2880,6 +2895,7 @@ cram_container *cram_read_container(cram_fd *fd) {

    c->offset = rd;
    c->slices = NULL;
+    c->slice = NULL;
    c->curr_slice = 0;
    c->max_slice = c->num_landmarks;
    c->slice_rec = 0;
@@ -3102,7 +3118,11 @@ void *cram_flush_thread(void *arg) {
 static int cram_flush_result(cram_fd *fd) {
    int i, ret = 0;
    hts_tpool_result *r;
+    cram_container *lc = NULL;

+    // NB: we can have one result per slice, not per container,
+    // so we need to free the container only after all slices
+    // within it have been freed.  (Automatic via reference counting.)
    while ((r = hts_tpool_next_result(fd->rqueue))) {
 	cram_job *j = (cram_job *)hts_tpool_result_data(r);
 	cram_container *c;
@@ -3119,23 +3139,49 @@ static int cram_flush_result(cram_fd *fd) {
 	    if (0 != cram_flush_container2(fd, c))
 		return -1;

-	/* Free the container */
+	// Free the slices; filled out by encoder only
+	if (c->slices) {
 	    for (i = 0; i < c->max_slice; i++) {
-	    if (c->slices && c->slices[i]) {
+		if (c->slices[i])
 		    cram_free_slice(c->slices[i]);
+		if (c->slices[i] == c->slice)
+		    c->slice = NULL;
 		c->slices[i] = NULL;
 	    }
 	}

+	// Free the current slice; set by both encoder & decoder
+	if (c->slice) {
+	    cram_free_slice(c->slice);
 	    c->slice = NULL;
+	}
 	c->curr_slice = 0;

-	cram_free_container(c);
+	// Our jobs will be in order, so we free the last
+	// container when our job has switched to a new one.
+	if (c != lc) {
+	    if (lc) {
+		if (fd->ctr == lc)
+		    fd->ctr = NULL;
+		if (fd->ctr_mt == lc)
+		    fd->ctr_mt = NULL;
+		cram_free_container(lc);
+	    }
+	    lc = c;
+	}

+	if (fd->mode == 'w')
 	    ret |= hflush(fd->fp) == 0 ? 0 : -1;

 	hts_tpool_delete_result(r, 1);
    }
+    if (lc) {
+	if (fd->ctr == lc)
+	    fd->ctr = NULL;
+	if (fd->ctr_mt == lc)
+	    fd->ctr_mt = NULL;
+	cram_free_container(lc);
+    }

    return ret;
 }
@@ -3447,15 +3493,17 @@ cram_slice *cram_read_slice(cram_fd *fd) {
 		min_id = s->block[i]->content_id;
 	}
    }
-    if (min_id >= 0 && max_id < 1024) {
-	if (!(s->block_by_id = calloc(1024, sizeof(s->block[0]))))
+
+    if (!(s->block_by_id = calloc(512, sizeof(s->block[0]))))
 	goto err;

    for (i = 0; i < n; i++) {
 	if (s->block[i]->content_type != EXTERNAL)
 	    continue;
-	    s->block_by_id[s->block[i]->content_id] = s->block[i];
-	}
+	int v = s->block[i]->content_id;
+	if (v < 0 || v >= 256)
+	    v = 256 + (v > 0 ? v % 251 : (-v) % 251);
+	s->block_by_id[v] = s->block[i];
    }

    /* Initialise encoding/decoding tables */
@@ -3473,6 +3521,7 @@ cram_slice *cram_read_slice(cram_fd *fd) {
    s->crecs = NULL;

    s->last_apos = s->hdr->ref_seq_start;
+    s->decode_md = fd->decode_md;

    return s;

@@ -3664,7 +3713,7 @@ static void full_path(char *out, char *in) {
    size_t in_l = strlen(in);
    if (*in == '/' ||
 	// Windows paths
-	(in_l > 3 && toupper(*in) >= 'A'  && toupper(*in) <= 'Z' &&
+	(in_l > 3 && toupper_c(*in) >= 'A'  && toupper_c(*in) <= 'Z' &&
 	 in[1] == ':' && (in[2] == '/' || in[2] == '\\'))) {
 	strncpy(out, in, PATH_MAX);
 	out[PATH_MAX-1] = 0;
@@ -4068,6 +4117,7 @@ cram_fd *cram_dopen(hFILE *fp, const char *filename, const char *mode) {
    fd->record_counter = 0;

    fd->ctr = NULL;
+    fd->ctr_mt = NULL;
    fd->refs  = refs_create();
    if (!fd->refs)
 	goto err;
@@ -4088,6 +4138,8 @@ cram_fd *cram_dopen(hFILE *fp, const char *filename, const char *mode) {
    fd->multi_seq = -1;
    fd->unsorted   = 0;
    fd->shared_ref = 0;
+    fd->store_md = 0;
+    fd->store_nm = 0;

    fd->index       = NULL;
    fd->own_pool    = 0;
@@ -4133,6 +4185,8 @@ int cram_seek(cram_fd *fd, off_t offset, int whence) {

    fd->ooc = 0;

+    cram_drain_rqueue(fd);
+
    if (hseek(fd->fp, offset, whence) >= 0) {
        return 0;
    }
@@ -4193,18 +4247,22 @@ int cram_close(cram_fd *fd) {
 	    return -1;
    }

+    if (fd->mode != 'w')
+	cram_drain_rqueue(fd);
+
    if (fd->pool && fd->eof >= 0) {
 	hts_tpool_process_flush(fd->rqueue);

 	if (0 != cram_flush_result(fd))
 	    return -1;

+	if (fd->mode == 'w')
+	    fd->ctr = NULL; // prevent double freeing
+
 	pthread_mutex_destroy(&fd->metrics_lock);
 	pthread_mutex_destroy(&fd->ref_lock);
 	pthread_mutex_destroy(&fd->bam_list_lock);

-	fd->ctr = NULL; // prevent double freeing
-
 	//fprintf(stderr, "CRAM: destroy queue %p\n", fd->rqueue);

 	hts_tpool_process_destroy(fd->rqueue);
@@ -4259,6 +4317,9 @@ int cram_close(cram_fd *fd) {
    if (fd->ctr)
 	cram_free_container(fd->ctr);

+    if (fd->ctr_mt && fd->ctr_mt != fd->ctr)
+	cram_free_container(fd->ctr_mt);
+
    if (fd->refs)
 	refs_free(fd->refs);
    if (fd->ref_free)
@@ -4402,9 +4463,14 @@ int cram_set_voption(cram_fd *fd, enum hts_fmt_option opt, va_list args) {
 	}
 	break;

-    case CRAM_OPT_RANGE:
-	fd->range = *va_arg(args, cram_range *);
-	return cram_seek_to_refpos(fd, &fd->range);
+    case CRAM_OPT_RANGE: {
+	int r = cram_seek_to_refpos(fd, va_arg(args, cram_range *));
+	pthread_mutex_lock(&fd->range_lock);
+	if (fd->range.refid != -2)
+	    fd->required_fields |= SAM_POS;
+	pthread_mutex_unlock(&fd->range_lock);
+	return r;
+    }

    case CRAM_OPT_REFERENCE:
 	return cram_load_reference(fd, va_arg(args, char *));
@@ -4443,6 +4509,7 @@ int cram_set_voption(cram_fd *fd, enum hts_fmt_option opt, va_list args) {
 	    fd->rqueue = hts_tpool_process_init(fd->pool, nthreads*2, 0);
 	    pthread_mutex_init(&fd->metrics_lock, NULL);
 	    pthread_mutex_init(&fd->ref_lock, NULL);
+	    pthread_mutex_init(&fd->range_lock, NULL);
 	    pthread_mutex_init(&fd->bam_list_lock, NULL);
 	    fd->shared_ref = 1;
 	    fd->own_pool = 1;
@@ -4459,6 +4526,7 @@ int cram_set_voption(cram_fd *fd, enum hts_fmt_option opt, va_list args) {
 						0);
 	    pthread_mutex_init(&fd->metrics_lock, NULL);
 	    pthread_mutex_init(&fd->ref_lock, NULL);
+	    pthread_mutex_init(&fd->range_lock, NULL);
 	    pthread_mutex_init(&fd->bam_list_lock, NULL);
 	}
 	fd->shared_ref = 1; // Needed to avoid clobbering ref between threads
@@ -4472,6 +4540,16 @@ int cram_set_voption(cram_fd *fd, enum hts_fmt_option opt, va_list args) {

    case CRAM_OPT_REQUIRED_FIELDS:
 	fd->required_fields = va_arg(args, int);
+	if (fd->range.refid != -2)
+	    fd->required_fields |= SAM_POS;
+	break;
+
+    case CRAM_OPT_STORE_MD:
+	fd->store_md = va_arg(args, int);
+	break;
+
+    case CRAM_OPT_STORE_NM:
+	fd->store_nm = va_arg(args, int);
 	break;

    case HTS_OPT_COMPRESSION_LEVEL:

--- a/cram/cram_io.h
+++ b/cram/cram_io.h
@@ -475,9 +475,17 @@ char *cram_content_type2str(enum cram_content_type t);
 */

 static inline cram_block *cram_get_block_by_id(cram_slice *slice, int id) {
-    if (slice->block_by_id && id >= 0 && id < 1024) {
+  //fprintf(stderr, "%d\t%p\n", id, slice->block_by_id);
+    if (slice->block_by_id && id >= 0 && id < 256) {
        return slice->block_by_id[id];
    } else {
+        int v = 256 + (id > 0 ? id % 251 : (-id) % 251);
+	if (slice->block_by_id &&
+	    slice->block_by_id[v] &&
+	    slice->block_by_id[v]->content_id == id)
+	    return slice->block_by_id[v];
+
+	// Otherwise a linear search in case of collision
 	int i;
        for (i = 0; i < slice->hdr->num_blocks; i++) {
 	    cram_block *b = slice->block[i];

--- a/cram/cram_structs.h
+++ b/cram/cram_structs.h
@@ -296,6 +296,7 @@ typedef struct cram_block_compression_hdr {
    int AP_delta;
    // indexed by ref-base and subst. code
    char substitution_matrix[5][4];
+    int no_ref;

    // TD Dictionary as a concatenated block
    cram_block *TD_blk;          // Tag Dictionary
@@ -312,8 +313,6 @@ typedef struct cram_block_compression_hdr {

    char *uncomp; // A single block of uncompressed data
    size_t uncomp_size, uncomp_alloc;
-
-    unsigned int data_series; // See cram_fields enum below
 } cram_block_compression_hdr;

 typedef struct cram_map {
@@ -381,6 +380,7 @@ typedef struct cram_container {

    /* For construction purposes */
    int max_slice, curr_slice;   // maximum number of slices
+    int curr_slice_mt;           // Curr_slice when reading ahead (via threads)
    int max_rec, curr_rec;       // current and max recs per slice
    int max_c_rec, curr_c_rec;   // current and max recs per container
    int slice_rec;               // rec no. for start of this slice
@@ -469,8 +469,7 @@ typedef struct cram_record {
 * A feature is a base difference, used for the sequence reference encoding.
 * (We generate these internally when writing CRAM.)
 */
-typedef struct cram_feature {
-    union {
+typedef union cram_feature {
    struct {
        int pos;
        int code;
@@ -530,7 +529,6 @@ typedef struct cram_feature {
        int code;
        int len;
    } H;
-    };
 } cram_feature;

 /*
@@ -589,6 +587,12 @@ typedef struct cram_slice {
    // For going from BAM to CRAM; an array of auxiliary blocks per type
    int naux_block;
    cram_block **aux_block;
+
+    unsigned int data_series; // See cram_fields enum
+    int decode_md;
+
+    int max_rec, curr_rec;       // current and max recs per slice
+    int slice_num;               // To be copied into c->curr_slice in decode
 } cram_slice;

 /*-----------------------------------------------------------------------------
@@ -685,9 +689,12 @@ typedef struct cram_fd {
    //cram_block_compression_hdr *comp_hdr;
    //cram_block_slice_hdr       *slice_hdr;

-    // Current container being processed.
+    // Current container being processed
    cram_container *ctr;

+    // Current container used for decoder threads
+    cram_container *ctr_mt;
+
    // positions for encoding or decoding
    int first_base, last_base;

@@ -717,6 +724,8 @@ typedef struct cram_fd {
    int use_lzma;
    int shared_ref;
    unsigned int required_fields;
+    int store_md;
+    int store_nm;
    cram_range range;

    // lookup tables, stored here so we can be trivially multi-threaded
@@ -741,6 +750,7 @@ typedef struct cram_fd {
    hts_tpool_process *rqueue;
    pthread_mutex_t metrics_lock;
    pthread_mutex_t ref_lock;
+    pthread_mutex_t range_lock;
    spare_bams *bl;
    pthread_mutex_t bam_list_lock;
    void *job_pending;
@@ -834,7 +844,7 @@ enum cram_fields {

 /* Internal only */
 #define CRAM_FLAG_STATS_ADDED          (1<<30)
-#define CRAM_FLAG_DISCARD_NAME         (1<<31)
+#define CRAM_FLAG_DISCARD_NAME         (1U<<31)

 #ifdef __cplusplus
 }

--- a/cram/open_trace_file.c
+++ b/cram/open_trace_file.c
@@ -120,10 +120,13 @@ char *tokenise_search_path(char *searchpath) {
 	if (path_sep == ':') {
 	    if ((i == 0 || (i > 0 && searchpath[i-1] == ':')) &&
 		(!strncmp(&searchpath[i], "http:",     5) ||
+		 !strncmp(&searchpath[i], "https:",    6) ||
 		 !strncmp(&searchpath[i], "ftp:",      4) ||
 		 !strncmp(&searchpath[i], "|http:",    6) ||
+		 !strncmp(&searchpath[i], "|https:",   7) ||
 		 !strncmp(&searchpath[i], "|ftp:",     5) ||
 		 !strncmp(&searchpath[i], "URL=http:", 9) ||
+		 !strncmp(&searchpath[i], "URL=https:",10)||
 		 !strncmp(&searchpath[i], "URL=ftp:",  8))) {
 		do {
 		    newsearch[j++] = searchpath[i];
@@ -342,6 +345,7 @@ mFILE *open_path_mfile(char *file, char *path, char *relative_to) {
 		return fp;
 	    }
 	} else if (!strncmp(ele2, "http:", 5) ||
+		   !strncmp(ele2, "https:", 6) ||
 		   !strncmp(ele2, "ftp:", 4)) {
 	    if ((fp = find_file_url(file, ele2))) {
 		free(newsearch);
@@ -394,6 +398,7 @@ char *find_path(char *file, char *path) {

 	if (!strncmp(ele2, "URL=", 4) ||
 	    !strncmp(ele2, "http:", 5) ||
+	    !strncmp(ele2, "https:", 6) ||
 	    !strncmp(ele2, "ftp:", 4)) {
 	    continue;
 	} else {