#!/bin/sh
# grep 3.4 through 3.7 mishandled matching '.' against the valid UTF-8
# sequences (ED)(90-9F)(80-BF) corresponding to U+D400 through U+D7FF,
# which are some Hangul Syllables and Hangul Jamo Extended-B.  They
# also mishandled (F4)(88-8F)(80-BF)(80-BF) which correspond to
# U+108000 through U+10FFFF (Supplemental Private Use Area plane B).

. "${srcdir=.}/init.sh"; path_prepend_ ../src

require_en_utf8_locale_

LC_ALL=en_US.UTF-8
export LC_ALL

# Check that '.' completely matches $1, i.e., that $1 is a single UTF-8 char.
check_char ()
{
  printf "$1\\n" >in || framework_failure_

  grep $2 '^.$' in >out || fail=1
  cmp in out || fail=1
}

# Check that '.*' does not completely match $1, i.e., that
# $1 contains an encoding error.
check_nonchar ()
{
  printf "$1\\n" >in || framework_failure_

  grep -a -v '^.*$' in >out || fail=1
  cmp in out || fail=1
}

fail=0

# "." should match U+D45C HANGUL SYLLABLE PYO.
check_char '\355\221\234'

# Check boundary-condition characters, and non-characters,
# while we are at it.

check_char '\0' -a
check_char '\177'
check_nonchar '\200'
check_nonchar '\277'
check_nonchar '\300\200'
check_nonchar '\301\277'

for i in 302 337; do
  for j in 200 277; do
    check_char "\\$i\\$j"
  done
  for j in 177 300; do
    check_nonchar "\\$i\\$j"
  done
done
for i in 340; do
  for j in 240 277; do
    for k in 200 277; do
      check_char "\\$i\\$j\\$k"
    done
    for k in 177 300; do
      check_nonchar "\\$i\\$j\\$k"
    done
  done
  for j in 239 300; do
    for k in 177 200 277 300; do
      check_nonchar "\\$i\\$j\\$k"
    done
  done
done
for i in 341 354 356 357; do
  for j in 200 277; do
    for k in 200 277; do
      check_char "\\$i\\$j\\$k"
    done
    for k in 177 300; do
      check_nonchar "\\$i\\$j\\$k"
    done
  done
  for j in 177 300; do
    for k in 177 200 277 300; do
      check_nonchar "\\$i\\$j\\$k"
    done
  done
done
for i in 355; do
  for j in 200 237; do
    for k in 200 277; do
      check_char "\\$i\\$j\\$k"
    done
    for k in 177 300; do
      check_nonchar "\\$i\\$j\\$k"
    done
  done
  for j in 177 240; do
    for k in 177 200 277 300; do
      check_nonchar "\\$i\\$j\\$k"
    done
  done
done

# On platforms like 32-bit AIX where WCHAR_MAX == 0xFFFF, skip checks
# where the corresponding Unicode characters are not supported.
if test $fail -eq 0; then
  printf '\360\220\200\200\n' >in || framework_failure_
  grep '^.$' in >out 2>&1 || fail=1
  cmp in out || skip_ 'platform does not support U+10000'
fi

for i in 360; do
  for j in 220 277; do
    for k in 200 277; do
      for l in 200 277; do
        check_char "\\$i\\$j\\$k\\$l"
      done
      for l in 177 300; do
        check_nonchar "\\$i\\$j\\$k\\$l"
      done
    done
    for k in 177 300; do
      for l in 177 200 277 300; do
        check_nonchar "\\$i\\$j\\$k\\$l"
      done
    done
  done
  for j in 217 300; do
    for k in 177 200 277 300; do
      for l in 177 200 277 300; do
        check_nonchar "\\$i\\$j\\$k\\$l"
      done
    done
  done
done
for i in 361 363; do
  for j in 200 277; do
    for k in 200 277; do
      for l in 200 277; do
        check_char "\\$i\\$j\\$k\\$l"
      done
      for l in 177 300; do
        check_nonchar "\\$i\\$j\\$k\\$l"
      done
    done
    for k in 177 300; do
      for l in 177 200 277 300; do
        check_nonchar "\\$i\\$j\\$k\\$l"
      done
    done
  done
  for j in 177 300; do
    for k in 177 200 277 300; do
      for l in 177 200 277 300; do
        check_nonchar "\\$i\\$j\\$k\\$l"
      done
    done
  done
done
for i in 364; do
  for j in 200 217; do
    for k in 200 277; do
      for l in 200 277; do
        check_char "\\$i\\$j\\$k\\$l"
      done
      for l in 177 300; do
        check_nonchar "\\$i\\$j\\$k\\$l"
      done
    done
    for k in 177 300; do
      for l in 177 200 277 300; do
        check_nonchar "\\$i\\$j\\$k\\$l"
      done
    done
  done
  for j in 177 220; do
    for k in 177 200 277 300; do
      for l in 177 200 277 300; do
        check_nonchar "\\$i\\$j\\$k\\$l"
      done
    done
  done
done

Exit $fail