@@ -1,13 +1,13 @@

 Package: seqbias

-Version: 0.99.1

+Version: 0.99.2

 Date: 25-12-2010

 Title: Estimation of per-position bias in high-throughput sequencing data

-Description: This package implements a model of per-position sequencing bias i

+Description: This package implements a model of per-position sequencing bias in

     high-throughput sequencing data using a simple Bayesian network, the

     structure and parameters of which are trained on a set of aligned reads

     and a reference genome sequence.

 Author: Daniel Jones <dcjones@cs.washington.edu>

-Maintainer: Dainel Jones <dcjones@cs.washington.edu>

+Maintainer: Daniel Jones <dcjones@cs.washington.edu>

 Depends: R (>= 2.13.0), GenomicRanges (>= 0.1.0),

          Biostrings (>= 2.15.0), methods 

 Suggests: Rsamtools (>= 1.3.18)

@@ -10,16 +10,16 @@ generated by GNU Autoconf 2.67.  Invocation command line was

 ## Platform. ##

 ## --------- ##

-hostname = alpaca

-uname -m = x86_64

-uname -r = 2.6.31.12-0.2-desktop

+hostname = dtp2

+uname -m = i686

+uname -r = 2.6.35-22-generic-pae

 uname -s = Linux

-uname -v = #1 SMP PREEMPT 2010-03-16 21:25:39 +0100

+uname -v = #34-Ubuntu SMP Sun Oct 10 11:03:48 UTC 2010

 /usr/bin/uname -p = unknown

 /bin/uname -X     = unknown

-/bin/arch              = x86_64

+/bin/arch              = unknown

 /usr/bin/arch -k       = unknown

 /usr/convex/getsysinfo = unknown

 /usr/bin/hostinfo      = unknown

@@ -27,18 +27,17 @@ uname -v = #1 SMP PREEMPT 2010-03-16 21:25:39 +0100

 /usr/bin/oslevel       = unknown

 /bin/universe          = unknown

-PATH: /usr/lib64/mpi/gcc/openmpi/bin

-PATH: /home/cwon2/bin

+PATH: /home/dcjones/bin

+PATH: /home/dcjones/bin/i686

+PATH: /home/dcjones/src/kent/built/bin/i386

+PATH: /home/dcjones/.cabal/bin

+PATH: /usr/local/sbin

 PATH: /usr/local/bin

+PATH: /usr/sbin

 PATH: /usr/bin

+PATH: /sbin

 PATH: /bin

-PATH: /usr/bin/X11

-PATH: /usr/X11R6/bin

 PATH: /usr/games

-PATH: /opt/kde3/bin

-PATH: /usr/lib/mit/bin

-PATH: /usr/lib/mit/sbin

-PATH: /usr/lib/qt3/bin

 ## ----------- ##

@@ -50,18 +49,18 @@ configure:2170: found /usr/bin/gcc

 configure:2181: result: gcc

 configure:2410: checking for C compiler version

 configure:2419: gcc --version >&5

-gcc (SUSE Linux) 4.4.1 [gcc-4_4-branch revision 150839]

-Copyright (C) 2009 Free Software Foundation, Inc.

+gcc (Ubuntu/Linaro 4.4.4-14ubuntu5) 4.4.5

+Copyright (C) 2010 Free Software Foundation, Inc.

 This is free software; see the source for copying conditions.  There is NO

 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

 configure:2430: $? = 0

 configure:2419: gcc -v >&5

 Using built-in specs.

-Target: x86_64-suse-linux

-Configured with: ../configure --prefix=/usr --infodir=/usr/share/info --mandir=/usr/share/man --libdir=/usr/lib64 --libexecdir=/usr/lib64 --enable-languages=c,c++,objc,fortran,obj-c++,java,ada --enable-checking=release --with-gxx-include-dir=/usr/include/c++/4.4 --enable-ssp --disable-libssp --with-bugurl=http://bugs.opensuse.org/ --with-pkgversion='SUSE Linux' --disable-libgcj --disable-libmudflap --with-slibdir=/lib64 --with-system-zlib --enable-__cxa_atexit --enable-libstdcxx-allocator=new --disable-libstdcxx-pch --enable-version-specific-runtime-libs --program-suffix=-4.4 --enable-linux-futex --without-system-libunwind --with-arch-32=i586 --with-tune=generic --build=x86_64-suse-linux

+Target: i686-linux-gnu

+Configured with: ../src/configure -v --with-pkgversion='Ubuntu/Linaro 4.4.4-14ubuntu5' --with-bugurl=file:///usr/share/doc/gcc-4.4/README.Bugs --enable-languages=c,c++,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.4 --enable-shared --enable-multiarch --enable-linker-build-id --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.4 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-objc-gc --enable-targets=all --disable-werror --with-arch-32=i686 --with-tune=generic --enable-checking=release --build=i686-linux-gnu --host=i686-linux-gnu --target=i686-linux-gnu

 Thread model: posix

-gcc version 4.4.1 [gcc-4_4-branch revision 150839] (SUSE Linux) 

+gcc version 4.4.5 (Ubuntu/Linaro 4.4.4-14ubuntu5) 

 configure:2430: $? = 0

 configure:2419: gcc -V >&5

 gcc: '-V' option must have argument

@@ -119,8 +118,8 @@ configure:2948: result: yes

 configure:2948: checking for pow

 configure:2948: gcc -o conftest -g -O2   conftest.c -lz  >&5

 conftest.c:35: warning: conflicting types for built-in function 'pow'

-/tmp/ccYE2dt3.o: In function `main':

-/home/cwon2/Download/seqbias/conftest.c:46: undefined reference to `pow'

+/tmp/ccxSNas5.o: In function `main':

+/home/dcjones/bio/seqbias/conftest.c:46: undefined reference to `pow'

 collect2: ld returned 1 exit status

 configure:2948: $? = 1

 configure: failed program was:

@@ -177,8 +176,8 @@ configure:2948: result: no

 configure:2948: checking for sqrt

 configure:2948: gcc -o conftest -g -O2   conftest.c -lz  >&5

 conftest.c:35: warning: conflicting types for built-in function 'sqrt'

-/tmp/ccKTGxDe.o: In function `main':

-/home/cwon2/Download/seqbias/conftest.c:46: undefined reference to `sqrt'

+/tmp/ccFZLqEa.o: In function `main':

+/home/dcjones/bio/seqbias/conftest.c:46: undefined reference to `sqrt'

 collect2: ld returned 1 exit status

 configure:2948: $? = 1

 configure: failed program was:

@@ -255,7 +254,8 @@ configure:2962: checking how to run the C preprocessor

 configure:2993: gcc -E  conftest.c

 configure:2993: $? = 0

 configure:3007: gcc -E  conftest.c

-conftest.c:16:28: error: ac_nonexistent.h: No such file or directory

+conftest.c:16: fatal error: ac_nonexistent.h: No such file or directory

+compilation terminated.

 configure:3007: $? = 1

 configure: failed program was:

 | /* confdefs.h */

@@ -278,7 +278,8 @@ configure:3032: result: gcc -E

 configure:3052: gcc -E  conftest.c

 configure:3052: $? = 0

 configure:3066: gcc -E  conftest.c

-conftest.c:16:28: error: ac_nonexistent.h: No such file or directory

+conftest.c:16: fatal error: ac_nonexistent.h: No such file or directory

+compilation terminated.

 configure:3066: $? = 1

 configure: failed program was:

 | /* confdefs.h */

@@ -298,9 +299,9 @@ configure: failed program was:

 | /* end confdefs.h.  */

 | #include <ac_nonexistent.h>

 configure:3095: checking for grep that handles long lines and -e

-configure:3153: result: /usr/bin/grep

+configure:3153: result: /bin/grep

 configure:3158: checking for egrep

-configure:3220: result: /usr/bin/grep -E

+configure:3220: result: /bin/grep -E

 configure:3225: checking for ANSI C header files

 configure:3245: gcc -c -g -O2  conftest.c >&5

 configure:3245: $? = 0

@@ -776,7 +777,7 @@ generated by GNU Autoconf 2.67.  Invocation command line was

   CONFIG_COMMANDS = 

   $ ./config.status 

-on alpaca

+on dtp2

 config.status:734: creating src/Makevars

@@ -833,8 +834,8 @@ ac_cv_header_unistd_h=yes

 ac_cv_header_zlib_h=yes

 ac_cv_lib_z_gzeof=yes

 ac_cv_objext=o

-ac_cv_path_EGREP='/usr/bin/grep -E'

-ac_cv_path_GREP=/usr/bin/grep

+ac_cv_path_EGREP='/bin/grep -E'

+ac_cv_path_GREP=/bin/grep

 ac_cv_prog_CPP='gcc -E'

 ac_cv_prog_ac_ct_CC=gcc

 ac_cv_prog_cc_c89=

@@ -857,9 +858,9 @@ DEFS='-DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAG

 ECHO_C=''

 ECHO_N='-n'

 ECHO_T=''

-EGREP='/usr/bin/grep -E'

+EGREP='/bin/grep -E'

 EXEEXT=''

-GREP='/usr/bin/grep'

+GREP='/bin/grep'

 LDFLAGS=''

 LIBOBJS=''

 LIBS='-lz '

@@ -872,7 +873,7 @@ PACKAGE_TARNAME=''

 PACKAGE_URL=''

 PACKAGE_VERSION=''

 PATH_SEPARATOR=':'

-SHELL='/bin/sh'

+SHELL='/bin/bash'

 ac_ct_CC='gcc'

 bindir='${exec_prefix}/bin'

 build_alias=''

@@ -1,4 +1,4 @@

-#! /bin/sh

+#! /bin/bash

 # Generated by configure.

 # Run this file to recreate the current configuration.

 # Compiler output produced by configure, useful for debugging

@@ -8,7 +8,7 @@ debug=false

 ac_cs_recheck=false

 ac_cs_silent=false

-SHELL=${CONFIG_SHELL-/bin/sh}

+SHELL=${CONFIG_SHELL-/bin/bash}

 export SHELL

 ## -------------------- ##

 ## M4sh Initialization. ##

@@ -438,7 +438,7 @@ Copyright (C) 2010 Free Software Foundation, Inc.

 This config.status script is free software; the Free Software Foundation

 gives unlimited permission to copy, distribute and modify it."

-ac_pwd='/home/cwon2/Download/seqbias'

+ac_pwd='/home/dcjones/bio/seqbias'

 srcdir='.'

 test -n "$AWK" || AWK=awk

 # The default lists apply if the user does not specify any file.

@@ -506,10 +506,10 @@ if $ac_cs_silent; then

 fi

 if $ac_cs_recheck; then

-  set X '/bin/sh' './configure'  $ac_configure_extra_args --no-create --no-recursion

+  set X '/bin/bash' './configure'  $ac_configure_extra_args --no-create --no-recursion

   shift

-  $as_echo "running CONFIG_SHELL=/bin/sh $*" >&6

-  CONFIG_SHELL='/bin/sh'

+  $as_echo "running CONFIG_SHELL=/bin/bash $*" >&6

+  CONFIG_SHELL='/bin/bash'

   export CONFIG_SHELL

   exec "$@"

 fi

@@ -592,8 +592,8 @@ echo 'BEGIN {' >"$tmp/subs1.awk" &&

 cat >>"$tmp/subs1.awk" <<\_ACAWK &&

 S["LTLIBOBJS"]=""

 S["LIBOBJS"]=""

-S["EGREP"]="/usr/bin/grep -E"

-S["GREP"]="/usr/bin/grep"

+S["EGREP"]="/bin/grep -E"

+S["GREP"]="/bin/grep"

 S["CPP"]="gcc -E"

 S["OBJEXT"]="o"

 S["EXEEXT"]=""

@@ -643,7 +643,7 @@ S["PACKAGE_VERSION"]=""

 S["PACKAGE_TARNAME"]=""

 S["PACKAGE_NAME"]=""

 S["PATH_SEPARATOR"]=":"

-S["SHELL"]="/bin/sh"

+S["SHELL"]="/bin/bash"

 _ACAWK

 cat >>"$tmp/subs1.awk" <<_ACAWK &&

   for (key in S) S_is_set[key] = 1

@@ -9,10 +9,11 @@ YAML_CPP_OBJ = aliascontent.o conversion.o emitter.o emitterstate.o \

                stream.o tag.o

 SEQBIAS_OBJ = sequencing_bias.o kmers.o miscmath.o common.o \

-			  table.o superfasthash.o logger.o seqbias.o

+			  table.o superfasthash.o logger.o seqbias.o \

+			  snprintf.o

-DFLAGS = -D_USE_KNETFILE -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE

-PKG_CFLAGS  += $(DFLAGS) -Wno-unused-result -Wno-format

+DFLAGS = -D_USE_KNETFILE -D_FILE_OFFSET_BITS=64

+PKG_CFLAGS  += $(DFLAGS)

 PKG_CPPFLAGS = $(PKG_CFLAGS)

 PKG_LIBS+=-lz 

@@ -9,10 +9,11 @@ YAML_CPP_OBJ = aliascontent.o conversion.o emitter.o emitterstate.o \

                stream.o tag.o

 SEQBIAS_OBJ = sequencing_bias.o kmers.o miscmath.o common.o \

-			  table.o superfasthash.o logger.o seqbias.o

+			  table.o superfasthash.o logger.o seqbias.o \

+			  snprintf.o

-DFLAGS = -D_USE_KNETFILE -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE

-PKG_CFLAGS  += $(DFLAGS) -Wno-unused-result -Wno-format

+DFLAGS = -D_USE_KNETFILE -D_FILE_OFFSET_BITS=64

+PKG_CFLAGS  += $(DFLAGS)

 PKG_CPPFLAGS = $(PKG_CFLAGS)

 PKG_LIBS+=@LIBS@

@@ -596,7 +596,8 @@ double conditional_likelihood( size_t n, const double* l0, const double* l1, con

     l = 0.0;

     for( i = 0; i < n; i++ ) {

-        l += (c[i] == 0 ? l0[i] : l1[i]) - logaddexp( l0[i], l1[i] );

+        l += (c[i] == 0 ? l0[i] : l1[i])

+                    - logaddexp( l0[i], l1[i] );

     }

     return l;

@@ -715,7 +716,7 @@ void train_motifs( motif& M0, motif& M1,

     while( true ) {

         round_num++;

-        log_printf( LOG_MSG, "round_num %4d (ic = %0.4e) ", round_num, ic_curr );

+        log_printf( LOG_MSG, "round %4d (ic = %0.4e) ", round_num, ic_curr );

         ic_best = -HUGE_VAL;

         j_best = i_best = 0;

@@ -730,7 +731,6 @@ void train_motifs( motif& M0, motif& M1,

             for( i = i_start; i <= j; i++ ) {

-                log_puts( LOG_MSG, "." );

                 if( M0.has_edge( i, j ) ) {

                     continue;

@@ -740,6 +740,8 @@ void train_motifs( motif& M0, motif& M1,

                     continue;

                 }

+                log_puts( LOG_MSG, "." );

+

                 /* keep track of the old parameters to avoid retraining */

                 M0.store_row(j);

                 M1.store_row(j);

@@ -917,10 +919,11 @@ void train_motifs_backwards( motif& M0, motif& M1,

             i_last = M0.num_parents(j) > 1 ? j-1 : j;

             for( i = 0; i <= i_last; i++ ) {

-                log_puts( LOG_MSG, "." );

                 if( !M0.has_edge( i, j ) ) continue;

+                log_puts( LOG_MSG, "." );

+

                 /* keep track of the old parameters to avoid retraining */

                 M0.store_row(j);

                 M1.store_row(j);

@@ -2,6 +2,7 @@

 #include "sequencing_bias.hpp"

 #include "logger.h"

 #include "common.hpp"

+#include "miscmath.hpp"

 #include <cmath>

 #include <cctype>

@@ -14,11 +15,31 @@ using namespace std;

 /* simply uniform random numbers */

-double rand_uniform( double a, double b )

+static double rand_uniform( double a, double b )

 {

     return a + b * (double)rand() / (double)RAND_MAX;

 }

+/* random gaussians (copied from GSL, to avoid dependency) */

+static double rand_gauss( double sigma )

+{

+    double x, y, r2;

+

+    do

+    {

+        /* choose x,y in uniform square (-1,-1) to (+1,+1) */

+        x = -1 + 2 * rand_uniform(0.0,1.0);

+        y = -1 + 2 * rand_uniform(0.0,1.0);

+

+        /* see if it is in the unit circle */

+        r2 = x * x + y * y;

+    }

+    while (r2 > 1.0 || r2 == 0);

+

+    /* Box-Muller transform */

+    return sigma * y * sqrt (-2.0 * log (r2) / r2);

+}

+

 /* pseudocount used when sampling foreground and background nucleotide * frequencies */

 const double sequencing_bias::pseudocount = 1;

@@ -172,14 +193,14 @@ void sequencing_bias::build( const char* ref_fn,

         failf( "Can't open bam file '%s'.\n", reads_fn );

     }

+    /* find the first n unique reads by hashing */

     table T;

-    hash_reads( &T, reads_f, n );

-    n = T.m;

+    hash_reads( &T, reads_f, 0 );

-    /* resort the remaining (1-q)*n by position */

-    log_puts( LOG_MSG, "sorting by position ... " );

+    /* sort by position */

+    log_puts( LOG_MSG, "shuffling ... " );

     struct hashed_value** S;

-    table_sort_by_position( &T, &S );

+    table_sort_by_seq_rand( &T, &S );

     log_puts( LOG_MSG, "done.\n" );

@@ -196,13 +217,12 @@ void sequencing_bias::build( const char* ref_fn,

     std::deque<sequence*> training_seqs;

-    /* sample this many background sequence for each foreground sequence */

-    const int bg_count = 1;

+    /* background sampling */

+    const size_t bg_samples = 1; // make this many samples for each read

+    int bg_sample_num;           // keep track of the number of samples made

+    struct read_pos bg;          // background position being considered

-    int j;

-    pos bg_pos;

-    pos bg_offset;

-

+    int b;

     char*          seqname   = NULL;

     int            seqlen    = 0;

     int            curr_tid  = -1;

@@ -213,7 +233,7 @@ void sequencing_bias::build( const char* ref_fn,

     local_seq[L+R+1] = '\0';

-    for( i = 0; i < n; i++ ) {

+    for( i = 0; i < n && i < T.m; i++ ) {

         /* Load/switch sequences (chromosomes) as they are encountered in the

          * read stream. The idea here is to avoid thrashing by loading a large

@@ -241,7 +261,8 @@ void sequencing_bias::build( const char* ref_fn,

         }

         if( seq == NULL ) continue;

-        

+

+

         /* add a foreground sequence */

         if( S[i]->pos.strand ) {

             if( S[i]->pos.pos < R ) continue;

@@ -258,31 +279,29 @@ void sequencing_bias::build( const char* ref_fn,

         /* add a background sequence */

         /* adjust the current read position randomly, and sample */

-        for( j = 0; j < bg_count; j++ ) {

+        for( bg_sample_num = 0; bg_sample_num < bg_samples; bg_sample_num++ ) {

-            /* make things a bit more robust by sampling the background from a

-             * random offset */

-            bg_offset = (pos)rand_uniform( 100.0, 200.0 );

-            if( rand_uniform( -1.0, 1.0 ) < 0.0 ) bg_offset = -bg_offset;

+            /* attempt to sample a position near the current read, with no reads

+             * itself. */

+            memcpy( (void*)&bg, (void*)&S[i]->pos, sizeof(struct read_pos) );

-            bg_pos = S[i]->pos.pos + bg_offset;

+            bg.pos = S[i]->pos.pos + (pos)ceil( rand_gauss( 10 ) );

-            if( S[i]->pos.strand ) {

-                if( bg_pos < R ) continue;

-                memcpy( local_seq, seq + bg_pos - R, (L+1+R)*sizeof(char) );

+            if( bg.strand ) {

+                if( bg.pos < R ) continue;

+                memcpy( local_seq, seq + bg.pos - R, (L+1+R)*sizeof(char) );

                 seqrc( local_seq, L+1+R );

             }

             else {

-                if( bg_pos < L ) continue;

-                memcpy( local_seq, seq + (bg_pos-L), (L+1+R)*sizeof(char) );

+                if( bg.pos < L ) continue;

+                memcpy( local_seq, seq + (bg.pos-L), (L+1+R)*sizeof(char) );

             }

             training_seqs.push_back( new sequence( local_seq, 0 ) );

         }

     }

-

-    size_t max_k = 5;

+    size_t max_k = 3;

     size_t max_dep_dist = 5;

     M0 = new motif( L+1+R, max_k, 0 );

     M1 = new motif( L+1+R, max_k, 1 );

new file mode 100644

@@ -0,0 +1,1025 @@

+/*

+ * snprintf.c - a portable implementation of snprintf

+ *

+ * AUTHOR

+ *   Mark Martinec <mark.martinec@ijs.si>, April 1999.

+ *

+ *   Copyright 1999, Mark Martinec. All rights reserved.

+ *

+ * TERMS AND CONDITIONS

+ *   This program is free software; you can redistribute it and/or modify

+ *   it under the terms of the "Frontier Artistic License" which comes

+ *   with this Kit.

+ *

+ *   This program is distributed in the hope that it will be useful,

+ *   but WITHOUT ANY WARRANTY; without even the implied warranty

+ *   of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

+ *   See the Frontier Artistic License for more details.

+ *

+ *   You should have received a copy of the Frontier Artistic License

+ *   with this Kit in the file named LICENSE.txt .

+ *   If not, I'll be glad to provide one.

+ *

+ * FEATURES

+ * - careful adherence to specs regarding flags, field width and precision;

+ * - good performance for large string handling (large format, large

+ *   argument or large paddings). Performance is similar to system's sprintf

+ *   and in several cases significantly better (make sure you compile with

+ *   optimizations turned on, tell the compiler the code is strict ANSI

+ *   if necessary to give it more freedom for optimizations);

+ * - return value semantics per ISO/IEC 9899:1999 ("ISO C99");

+ * - written in standard ISO/ANSI C - requires an ANSI C compiler.

+ *

+ * SUPPORTED CONVERSION SPECIFIERS AND DATA TYPES

+ *

+ * This snprintf only supports the following conversion specifiers:

+ * s, c, d, u, o, x, X, p  (and synonyms: i, D, U, O - see below)

+ * with flags: '-', '+', ' ', '0' and '#'.

+ * An asterisk is supported for field width as well as precision.

+ *

+ * Length modifiers 'h' (short int), 'l' (long int),

+ * and 'll' (long long int) are supported.

+ * NOTE:

+ *   If macro SNPRINTF_LONGLONG_SUPPORT is not defined (default) the

+ *   length modifier 'll' is recognized but treated the same as 'l',

+ *   which may cause argument value truncation! Defining

+ *   SNPRINTF_LONGLONG_SUPPORT requires that your system's sprintf also

+ *   handles length modifier 'll'.  long long int is a language extension

+ *   which may not be portable.

+ *

+ * Conversion of numeric data (conversion specifiers d, u, o, x, X, p)

+ * with length modifiers (none or h, l, ll) is left to the system routine

+ * sprintf, but all handling of flags, field width and precision as well as

+ * c and s conversions is done very carefully by this portable routine.

+ * If a string precision (truncation) is specified (e.g. %.8s) it is

+ * guaranteed the string beyond the specified precision will not be referenced.

+ *

+ * Length modifiers h, l and ll are ignored for c and s conversions (data

+ * types wint_t and wchar_t are not supported).

+ *

+ * The following common synonyms for conversion characters are supported:

+ *   - i is a synonym for d

+ *   - D is a synonym for ld, explicit length modifiers are ignored

+ *   - U is a synonym for lu, explicit length modifiers are ignored

+ *   - O is a synonym for lo, explicit length modifiers are ignored

+ * The D, O and U conversion characters are nonstandard, they are supported

+ * for backward compatibility only, and should not be used for new code.

+ *

+ * The following is specifically NOT supported:

+ *   - flag ' (thousands' grouping character) is recognized but ignored

+ *   - numeric conversion specifiers: f, e, E, g, G and synonym F,

+ *     as well as the new a and A conversion specifiers

+ *   - length modifier 'L' (long double) and 'q' (quad - use 'll' instead)

+ *   - wide character/string conversions: lc, ls, and nonstandard

+ *     synonyms C and S

+ *   - writeback of converted string length: conversion character n

+ *   - the n$ specification for direct reference to n-th argument

+ *   - locales

+ *

+ * It is permitted for str_m to be zero, and it is permitted to specify NULL

+ * pointer for resulting string argument if str_m is zero (as per ISO C99).

+ *

+ * The return value is the number of characters which would be generated

+ * for the given input, excluding the trailing null. If this value

+ * is greater or equal to str_m, not all characters from the result

+ * have been stored in str, output bytes beyond the (str_m-1) -th character

+ * are discarded. If str_m is greater than zero it is guaranteed

+ * the resulting string will be null-terminated.

+ *

+ * NOTE that this matches the ISO C99, OpenBSD, and GNU C library 2.1,

+ * but is different from some older and vendor implementations,

+ * and is also different from XPG, XSH5, SUSv2 specifications.

+ * For historical discussion on changes in the semantics and standards

+ * of snprintf see printf(3) man page in the Linux programmers manual.

+ *

+ * Routines asprintf and vasprintf return a pointer (in the ptr argument)

+ * to a buffer sufficiently large to hold the resulting string. This pointer

+ * should be passed to free(3) to release the allocated storage when it is

+ * no longer needed. If sufficient space cannot be allocated, these functions

+ * will return -1 and set ptr to be a NULL pointer. These two routines are a

+ * GNU C library extensions (glibc).

+ *

+ * Routines asnprintf and vasnprintf are similar to asprintf and vasprintf,

+ * yet, like snprintf and vsnprintf counterparts, will write at most str_m-1

+ * characters into the allocated output string, the last character in the

+ * allocated buffer then gets the terminating null. If the formatted string

+ * length (the return value) is greater than or equal to the str_m argument,

+ * the resulting string was truncated and some of the formatted characters

+ * were discarded. These routines present a handy way to limit the amount

+ * of allocated memory to some sane value.

+ *

+ * AVAILABILITY

+ *   http://www.ijs.si/software/snprintf/

+ *

+ * REVISION HISTORY

+ * 1999-04	V0.9  Mark Martinec

+ *		- initial version, some modifications after comparing printf

+ *		  man pages for Digital Unix 4.0, Solaris 2.6 and HPUX 10,

+ *		  and checking how Perl handles sprintf (differently!);

+ * 1999-04-09	V1.0  Mark Martinec <mark.martinec@ijs.si>

+ *		- added main test program, fixed remaining inconsistencies,

+ *		  added optional (long long int) support;

+ * 1999-04-12	V1.1  Mark Martinec <mark.martinec@ijs.si>

+ *		- support the 'p' conversion (pointer to void);

+ *		- if a string precision is specified

+ *		  make sure the string beyond the specified precision

+ *		  will not be referenced (e.g. by strlen);

+ * 1999-04-13	V1.2  Mark Martinec <mark.martinec@ijs.si>

+ *		- support synonyms %D=%ld, %U=%lu, %O=%lo;

+ *		- speed up the case of long format string with few conversions;

+ * 1999-06-30	V1.3  Mark Martinec <mark.martinec@ijs.si>

+ *		- fixed runaway loop (eventually crashing when str_l wraps

+ *		  beyond 2^31) while copying format string without

+ *		  conversion specifiers to a buffer that is too short

+ *		  (thanks to Edwin Young <edwiny@autonomy.com> for

+ *		  spotting the problem);

+ *		- added macros PORTABLE_SNPRINTF_VERSION_(MAJOR|MINOR)

+ *		  to snprintf.h

+ * 2000-02-14	V2.0 (never released) Mark Martinec <mark.martinec@ijs.si>

+ *		- relaxed license terms: The Artistic License now applies.

+ *		  You may still apply the GNU GENERAL PUBLIC LICENSE

+ *		  as was distributed with previous versions, if you prefer;

+ *		- changed REVISION HISTORY dates to use ISO 8601 date format;

+ *		- added vsnprintf (patch also independently proposed by

+ *		  Caolan McNamara 2000-05-04, and Keith M Willenson 2000-06-01)

+ * 2000-06-27	V2.1  Mark Martinec <mark.martinec@ijs.si>

+ *		- removed POSIX check for str_m<1; value 0 for str_m is

+ *		  allowed by ISO C99 (and GNU C library 2.1) - (pointed out

+ *		  on 2000-05-04 by Caolan McNamara, caolan@ csn dot ul dot ie).

+ *		  Besides relaxed license this change in standards adherence

+ *		  is the main reason to bump up the major version number;

+ *		- added nonstandard routines asnprintf, vasnprintf, asprintf,

+ *		  vasprintf that dynamically allocate storage for the

+ *		  resulting string; these routines are not compiled by default,

+ *		  see comments where NEED_V?ASN?PRINTF macros are defined;

+ *		- autoconf contributed by Caolan McNamara

+ * 2000-10-06	V2.2  Mark Martinec <mark.martinec@ijs.si>

+ *		- BUG FIX: the %c conversion used a temporary variable

+ *		  that was no longer in scope when referenced,

+ *		  possibly causing incorrect resulting character;

+ *		- BUG FIX: make precision and minimal field width unsigned

+ *		  to handle huge values (2^31 <= n < 2^32) correctly;

+ *		  also be more careful in the use of signed/unsigned/size_t

+ *		  internal variables - probably more careful than many

+ *		  vendor implementations, but there may still be a case

+ *		  where huge values of str_m, precision or minimal field

+ *		  could cause incorrect behaviour;

+ *		- use separate variables for signed/unsigned arguments,

+ *		  and for short/int, long, and long long argument lengths

+ *		  to avoid possible incompatibilities on certain

+ *		  computer architectures. Also use separate variable

+ *		  arg_sign to hold sign of a numeric argument,

+ *		  to make code more transparent;

+ *		- some fiddling with zero padding and "0x" to make it

+ *		  Linux compatible;

+ *		- systematically use macros fast_memcpy and fast_memset

+ *		  instead of case-by-case hand optimization; determine some

+ *		  breakeven string lengths for different architectures;

+ *		- terminology change: 'format' -> 'conversion specifier',

+ *		  'C9x' -> 'ISO/IEC 9899:1999 ("ISO C99")',

+ *		  'alternative form' -> 'alternate form',

+ *		  'data type modifier' -> 'length modifier';

+ *		- several comments rephrased and new ones added;

+ *		- make compiler not complain about 'credits' defined but

+ *		  not used;

+ */

+

+

+/* Define HAVE_SNPRINTF if your system already has snprintf and vsnprintf.

+ *

+ * If HAVE_SNPRINTF is defined this module will not produce code for

+ * snprintf and vsnprintf, unless PREFER_PORTABLE_SNPRINTF is defined as well,

+ * causing this portable version of snprintf to be called portable_snprintf

+ * (and portable_vsnprintf).

+ */

+/* #define HAVE_SNPRINTF */

+

+/* Define PREFER_PORTABLE_SNPRINTF if your system does have snprintf and

+ * vsnprintf but you would prefer to use the portable routine(s) instead.

+ * In this case the portable routine is declared as portable_snprintf

+ * (and portable_vsnprintf) and a macro 'snprintf' (and 'vsnprintf')

+ * is defined to expand to 'portable_v?snprintf' - see file snprintf.h .

+ * Defining this macro is only useful if HAVE_SNPRINTF is also defined,

+ * but does does no harm if defined nevertheless.

+ */

+/* #define PREFER_PORTABLE_SNPRINTF */

+

+/* Define SNPRINTF_LONGLONG_SUPPORT if you want to support

+ * data type (long long int) and length modifier 'll' (e.g. %lld).

+ * If undefined, 'll' is recognized but treated as a single 'l'.

+ *

+ * If the system's sprintf does not handle 'll'

+ * the SNPRINTF_LONGLONG_SUPPORT must not be defined!

+ *

+ * This is off by default as (long long int) is a language extension.

+ */

+/* #define SNPRINTF_LONGLONG_SUPPORT */

+

+/* Define NEED_SNPRINTF_ONLY if you only need snprintf, and not vsnprintf.

+ * If NEED_SNPRINTF_ONLY is defined, the snprintf will be defined directly,

+ * otherwise both snprintf and vsnprintf routines will be defined

+ * and snprintf will be a simple wrapper around vsnprintf, at the expense

+ * of an extra procedure call.

+ */

+/* #define NEED_SNPRINTF_ONLY */

+

+/* Define NEED_V?ASN?PRINTF macros if you need library extension

+ * routines asprintf, vasprintf, asnprintf, vasnprintf respectively,

+ * and your system library does not provide them. They are all small

+ * wrapper routines around portable_vsnprintf. Defining any of the four

+ * NEED_V?ASN?PRINTF macros automatically turns off NEED_SNPRINTF_ONLY

+ * and turns on PREFER_PORTABLE_SNPRINTF.

+ *

+ * Watch for name conflicts with the system library if these routines

+ * are already present there.

+ *

+ * NOTE: vasprintf and vasnprintf routines need va_copy() from stdarg.h, as

+ * specified by C99, to be able to traverse the same list of arguments twice.

+ * I don't know of any other standard and portable way of achieving the same.

+ * With some versions of gcc you may use __va_copy(). You might even get away

+ * with "ap2 = ap", in this case you must not call va_end(ap2) !

+ *   #define va_copy(ap2,ap) ap2 = ap

+ */

+/* #define NEED_ASPRINTF   */

+/* #define NEED_ASNPRINTF  */

+/* #define NEED_VASPRINTF  */

+/* #define NEED_VASNPRINTF */

+

+

+/* Define the following macros if desired:

+ *   SOLARIS_COMPATIBLE, SOLARIS_BUG_COMPATIBLE,

+ *   HPUX_COMPATIBLE, HPUX_BUG_COMPATIBLE, LINUX_COMPATIBLE,

+ *   DIGITAL_UNIX_COMPATIBLE, DIGITAL_UNIX_BUG_COMPATIBLE,

+ *   PERL_COMPATIBLE, PERL_BUG_COMPATIBLE,

+ *

+ * - For portable applications it is best not to rely on peculiarities

+ *   of a given implementation so it may be best not to define any

+ *   of the macros that select compatibility and to avoid features

+ *   that vary among the systems.

+ *

+ * - Selecting compatibility with more than one operating system

+ *   is not strictly forbidden but is not recommended.

+ *

+ * - 'x'_BUG_COMPATIBLE implies 'x'_COMPATIBLE .

+ *

+ * - 'x'_COMPATIBLE refers to (and enables) a behaviour that is

+ *   documented in a sprintf man page on a given operating system

+ *   and actually adhered to by the system's sprintf (but not on

+ *   most other operating systems). It may also refer to and enable

+ *   a behaviour that is declared 'undefined' or 'implementation specific'

+ *   in the man page but a given implementation behaves predictably

+ *   in a certain way.

+ *

+ * - 'x'_BUG_COMPATIBLE refers to (and enables) a behaviour of system's sprintf

+ *   that contradicts the sprintf man page on the same operating system.

+ *

+ * - I do not claim that the 'x'_COMPATIBLE and 'x'_BUG_COMPATIBLE

+ *   conditionals take into account all idiosyncrasies of a particular

+ *   implementation, there may be other incompatibilities.

+ */

+

+

+

+/* ============================================= */

+/* NO USER SERVICABLE PARTS FOLLOWING THIS POINT */

+/* ============================================= */

+

+#define PORTABLE_SNPRINTF_VERSION_MAJOR 2

+#define PORTABLE_SNPRINTF_VERSION_MINOR 2

+

+#if defined(NEED_ASPRINTF) || defined(NEED_ASNPRINTF) || defined(NEED_VASPRINTF) || defined(NEED_VASNPRINTF)

+# if defined(NEED_SNPRINTF_ONLY)

+# undef NEED_SNPRINTF_ONLY

+# endif

+# if !defined(PREFER_PORTABLE_SNPRINTF)

+# define PREFER_PORTABLE_SNPRINTF

+# endif

+#endif

+

+#if defined(SOLARIS_BUG_COMPATIBLE) && !defined(SOLARIS_COMPATIBLE)

+#define SOLARIS_COMPATIBLE

+#endif

+

+#if defined(HPUX_BUG_COMPATIBLE) && !defined(HPUX_COMPATIBLE)

+#define HPUX_COMPATIBLE

+#endif

+

+#if defined(DIGITAL_UNIX_BUG_COMPATIBLE) && !defined(DIGITAL_UNIX_COMPATIBLE)

+#define DIGITAL_UNIX_COMPATIBLE

+#endif

+

+#if defined(PERL_BUG_COMPATIBLE) && !defined(PERL_COMPATIBLE)

+#define PERL_COMPATIBLE

+#endif

+

+#if defined(LINUX_BUG_COMPATIBLE) && !defined(LINUX_COMPATIBLE)

+#define LINUX_COMPATIBLE

+#endif

+

+#include <sys/types.h>

+#include <string.h>

+#include <stdlib.h>

+#include <stdio.h>

+#include <stdarg.h>

+#include <assert.h>

+#include <errno.h>

+

+#ifdef isdigit

+#undef isdigit

+#endif

+#define isdigit(c) ((c) >= '0' && (c) <= '9')

+

+/* For copying strings longer or equal to 'breakeven_point'

+ * it is more efficient to call memcpy() than to do it inline.

+ * The value depends mostly on the processor architecture,

+ * but also on the compiler and its optimization capabilities.

+ * The value is not critical, some small value greater than zero

+ * will be just fine if you don't care to squeeze every drop

+ * of performance out of the code.

+ *

+ * Small values favor memcpy, large values favor inline code.

+ */

+#if defined(__alpha__) || defined(__alpha)

+#  define breakeven_point   2	/* AXP (DEC Alpha)     - gcc or cc or egcs */

+#endif

+#if defined(__i386__)  || defined(__i386)

+#  define breakeven_point  12	/* Intel Pentium/Linux - gcc 2.96 */

+#endif

+#if defined(__hppa)

+#  define breakeven_point  10	/* HP-PA               - gcc */

+#endif

+#if defined(__sparc__) || defined(__sparc)

+#  define breakeven_point  33	/* Sun Sparc 5         - gcc 2.8.1 */

+#endif

+

+/* some other values of possible interest: */

+/* #define breakeven_point  8 */  /* VAX 4000          - vaxc */

+/* #define breakeven_point 19 */  /* VAX 4000          - gcc 2.7.0 */

+

+#ifndef breakeven_point

+#  define breakeven_point   6	/* some reasonable one-size-fits-all value */

+#endif

+

+#define fast_memcpy(d,s,n) \

+  { register size_t nn = (size_t)(n); \

+    if (nn >= breakeven_point) memcpy((d), (s), nn); \

+    else if (nn > 0) { /* proc call overhead is worth only for large strings*/\

+      register char *dd; register const char *ss; \

+      for (ss=(s), dd=(d); nn>0; nn--) *dd++ = *ss++; } }

+

+#define fast_memset(d,c,n) \

+  { register size_t nn = (size_t)(n); \

+    if (nn >= breakeven_point) memset((d), (int)(c), nn); \

+    else if (nn > 0) { /* proc call overhead is worth only for large strings*/\

+      register char *dd; register const int cc=(int)(c); \

+      for (dd=(d); nn>0; nn--) *dd++ = cc; } }

+

+/* prototypes */

+

+#if defined(NEED_ASPRINTF)

+int asprintf   (char **ptr, const char *fmt, /*args*/ ...);

+#endif

+#if defined(NEED_VASPRINTF)

+int vasprintf  (char **ptr, const char *fmt, va_list ap);

+#endif

+#if defined(NEED_ASNPRINTF)

+int asnprintf  (char **ptr, size_t str_m, const char *fmt, /*args*/ ...);

+#endif

+#if defined(NEED_VASNPRINTF)

+int vasnprintf (char **ptr, size_t str_m, const char *fmt, va_list ap);

+#endif

+

+#if defined(HAVE_SNPRINTF)

+/* declare our portable snprintf  routine under name portable_snprintf  */

+/* declare our portable vsnprintf routine under name portable_vsnprintf */

+#else

+/* declare our portable routines under names snprintf and vsnprintf */

+#define portable_snprintf snprintf

+#if !defined(NEED_SNPRINTF_ONLY)

+#define portable_vsnprintf vsnprintf

+#endif

+#endif

+

+#if !defined(HAVE_SNPRINTF) || defined(PREFER_PORTABLE_SNPRINTF)

+int portable_snprintf(char *str, size_t str_m, const char *fmt, /*args*/ ...);

+#if !defined(NEED_SNPRINTF_ONLY)

+int portable_vsnprintf(char *str, size_t str_m, const char *fmt, va_list ap);

+#endif

+#endif

+

+/* declarations */

+

+static char credits[] = "\n\

+@(#)snprintf.c, v2.2: Mark Martinec, <mark.martinec@ijs.si>\n\

+@(#)snprintf.c, v2.2: Copyright 1999, Mark Martinec. Frontier Artistic License applies.\n\

+@(#)snprintf.c, v2.2: http://www.ijs.si/software/snprintf/\n";

+

+#if defined(NEED_ASPRINTF)

+int asprintf(char **ptr, const char *fmt, /*args*/ ...) {

+  va_list ap;