scripts/library: add human_to_size

This is a bash wrapper around an awk function that parses human readable
sizes and returns their representative values in bytes, as a string. A
small test harness is added to validate the functionality.

Signed-off-by: Dave Reisner <dreisner@archlinux.org>

Makefile.am

@@ -41,6 +41,8 @@ test-vercmp: test/util src/util
 test-parseopts: test/scripts scripts
 	$(BASH_SHELL) $(top_srcdir)/test/scripts/parseopts_test.sh \
 		$(top_srcdir)/scripts/library/parseopts.sh
+	$(BASH_SHELL) $(top_srcdir)/test/scripts/human_to_size_test.sh \
+		$(top_srcdir)/scripts/library/human_to_size.sh
 
 # create the pacman DB and cache directories upon install
 install-data-local:

scripts/Makefile.am

@@ -27,7 +27,8 @@ EXTRA_DIST = \
 
 LIBRARY = \
 	library/output_format.sh \
-	library/parseopts.sh
+	library/parseopts.sh \
+	library/human_to_size.sh
 
 # Files that should be removed, but which Automake does not know.
 MOSTLYCLEANFILES = $(bin_SCRIPTS)

scripts/library/README

@@ -27,3 +27,11 @@ Reccommended Usage:
 Returns:
   0: parse success
   1: parse failure (error message supplied)
+
+human_to_size.sh:
+A function to convert human readable sizes (such as "5.3 GiB") to raw byte
+equivalents. base10 and base2 suffixes are supported, case sensitively. If
+successful, the converted byte value is written to stdout and the function
+returns 0. If an error occurs, nothing in written and the function returns 1.
+Results may be inaccurate when using a broken implementation of awk, such
+as mawk or busybox awk.

scripts/library/human_to_size.sh

@@ -0,0 +1,51 @@
+human_to_size() {
+  awk -v human="$1" '
+  function trim(s) {
+    gsub(/^[[:space:]]+|[[:space:]]+$/, "", s)
+    return s
+  }
+
+  function parse_units(units) {
+    if (!units || units == "B")
+      return 1
+    if (match(units, /^.iB$/))
+      return 1024
+    if (match(units, /^.B$/))
+      return 1000
+    if (length(units) == 1)
+      return 1024
+
+    # parse failure: invalid base
+    return -1
+  }
+
+  function parse_scale(s) {
+    return index("BKMGTPE", s) - 1
+  }
+
+  function isnumeric(string) {
+    return match(string, /^[-+]?[[:digit:]]*(\.[[:digit:]]*)?/)
+  }
+
+  BEGIN {
+    # peel off the leading number as the size, fail on invalid number
+    human = trim(human)
+    if (isnumeric(human))
+      size = substr(human, RSTART, RLENGTH)
+    else
+      exit 1
+
+    # the trimmed remainder is assumed to be the units
+    units = trim(substr(human, RLENGTH + 1))
+
+    base = parse_units(units)
+    if (base < 0)
+      exit 1
+
+    scale = parse_scale(substr(units, 1, 1))
+    if (scale < 0)
+      exit 1
+
+    printf "%d\n", size * base^scale + (size + 0 > 0 ? 0.5 : -0.5)
+  }'
+}

test/scripts/Makefile.am

@@ -1,5 +1,6 @@
 check_SCRIPTS = \
-	parseopts_test.sh
+	parseopts_test.sh \
+	human_to_size_test.sh
 
 noinst_SCRIPTS = $(check_SCRIPTS)
 

test/scripts/human_to_size_test.sh

@@ -0,0 +1,75 @@
+#!/bin/bash
+
+# source the library function
+if [[ -z $1 || ! -f $1 ]]; then
+  printf "error: path to human_to_size library not provided or does not exist\n"
+  exit 1
+fi
+. "$1"
+
+if ! type -t human_to_size >/dev/null; then
+  printf 'human_to_size function not found\n'
+  exit 1
+fi
+
+parse_hts() {
+  local input=$1 expected=$2 result
+
+  (( ++testcount ))
+
+  result=$(human_to_size "$1")
+  if [[ $result = "$expected" ]]; then
+    (( ++pass ))
+  else
+    (( ++fail ))
+    printf '[TEST %3s]: FAIL\n' "$testcount"
+    printf '      input: %s\n' "$input"
+    printf '     output: %s\n' "$result"
+    printf '   expected: %s\n' "$expected"
+  fi
+}
+
+summarize() {
+  if (( !fail )); then
+    printf 'All %s tests successful\n' "$testcount"
+    exit 0
+  else
+    printf '%s of %s tests failed\n' "$fail" "$testcount"
+    exit 1
+  fi
+}
+trap 'summarize' EXIT
+
+printf 'Beginning human_to_size tests\n'
+
+# parse_hts <input> <expected output>
+
+parse_hts '1MiB' 1048576
+
+parse_hts '10XiB' ''
+
+parse_hts '10 MiB' 10485760
+
+parse_hts '10 XiB' ''
+
+parse_hts '.1 TiB' 109951162778
+
+parse_hts '  -3    KiB   ' -3072
+
+parse_hts 'foo3KiB' ''
+
+parse_hts '3KiBfoo' ''
+
+parse_hts '3kib' ''
+
+parse_hts '+1KiB' 1024
+
+parse_hts '+1.0 KiB' 1024
+
+parse_hts '1MB' 1000000
+
+parse_hts '1M' 1048576
+
+parse_hts ' 1 G ' 1073741824
+
+parse_hts '1Q' ''