Sorry about the slight delay in pickup. You can blame a few lame cheaters. This is to make their cheating a bit harder.
#!/usr/bin/env awk
function shift( array, \
junk, elm0, l )
{
elm0 = array[0]
for ( l = 0; l < asorti( array, junk ) - 1; l++ )
array[l] = array[l+1];
delete array[l]
return elm0
}
function init_cpp_src_line()
{
logical_line = ""
delete break_pos
}
function shift_valid_bp( array, \
junk, elm )
{
elm = -1
if ( 0 < asorti( array, junk ) )
do {
elm = shift( array )
} while ( 0 > elm );
return elm
}
function check_cpp_src_line_break_pos( \
i, junk )
{
printf( "break_pos:" )
for ( i = 0; i < asorti( break_pos, junk ); i++ )
printf( " %d", break_pos[i] );
printf( "\n" )
}
function check_cpp_src_line()
{
printf( "logical_line[%s]\n", logical_line )
check_cpp_src_line_break_pos()
}
function append_line( phys_line, \
filt_line, bp_len )
{
filt_line = phys_line
sub( /\\$/, " ", filt_line )
logical_line = logical_line filt_line
bp_len = asorti( break_pos, junk )
break_pos[bp_len] = length( logical_line ) - 1
}
function print_line( \
c0, c1, i, junk, part_str )
{
c0 = 0
while( asorti( break_pos, junk ) > 1 )
{
if ( ( c1 = shift_valid_bp( break_pos ) ) < 1 )
{
part_str = substr( logical_line, c0 + 1 )
printf( "%s\n", part_str )
return
}
part_str = substr( logical_line, c0 + 1, c1 - c0 + 1 )
gsub( / $/, "\\", part_str )
printf( "%s\n", part_str )
c0 = c1 + 1
}
part_str = substr( logical_line, c0 + 1 )
printf( "%s\n", part_str )
}
function shrink_spaces( pos, \
tail, removed_length, k )
{
tail = substr( logical_line, pos )
sub( /^[ \t]+/, " ", tail )
removed_length = length( logical_line ) - pos - length( tail ) + 1
logical_line = substr( logical_line, 0, pos - 1 ) tail
for ( k = 0; k < asorti( break_pos, junk ); k++ )
if ( ( pos + removed_length ) <= break_pos[k] )
break_pos[k] = break_pos[k] - removed_length;
else if ( pos <= break_pos[k] )
break_pos[k] = -1;
return removed_length
}
function shrink_spaces_to_linebreak( pos, \
junk, part_str, removed_length, i )
{
for ( i = 0; i < asorti( break_pos, junk ) && break_pos[i] < pos ; i++ )
;
if ( break_pos[i] < 1 )
return;
part_str = substr( logical_line, pos, break_pos[i] - pos + 1 )
sub( /^[ \t]+/, " ", part_str )
removed_length = ( break_pos[i] - pos + 1 ) - length( part_str )
tail = substr( logical_line, pos + removed_length )
logical_line = substr( logical_line, 0, pos - 1 ) tail
for ( ; i < asorti( break_pos, junk ); i++ )
break_pos[i] -= removed_length;
return removed_length
}
function delete_linebreaks_in_2nd_token( \
tail, paren_depth, junk, i, j, k, l )
{
if ( logical_line ~ /^[ \t]*#[ \t]*define[ \t]+[0-9A-Za-z_]+\(/ )
{
tail = logical_line
sub( /^[ \t]*#[ \t]*define[ \t]+[0-9A-Za-z_]+/, "", tail )
paren_depth = 0
l = 0
i = length( logical_line ) - length( tail ) + 1 # seek to the 1st op paren
j = i
do {
if ( substr( logical_line, j, 2 ) ~ /[ \t][ \t]/ )
l = shrink_spaces( j );
else if ( substr( logical_line, j, 1 ) == "(" )
paren_depth += 1;
else if ( substr( logical_line, j, 1 ) == ")" )
paren_depth -= 1;
j += 1
} while ( j < length( logical_line ) && paren_depth != 0 )
for ( k = 0; k < asorti( break_pos, junk ); k++ )
if ( i <= break_pos[k] && break_pos[k] < j )
break_pos[k] = -1;
if ( l > 0 )
shrink_spaces_to_linebreak( j );
}
}
BEGIN{
init_cpp_src_line()
}
{
append_line( $0 )
if ( $0 !~ /\\$/ )
{
delete_linebreaks_in_2nd_token()
print_line()
init_cpp_src_line()
}
}
END{
if ( 0 < length( logical_line ) )
{
delete_linebreaks_in_2nd_token()
print_line()
}
}