| 1 | package File::Kvpar; | ||
| 2 | |||
| 3 | use strict; | ||
| 4 | use warnings; | ||
| 5 | |||
| 6 | use constant HEAD => 1; | ||
| 7 | use constant TAIL => 2; | ||
| 8 | |||
| 9 | sub new { | ||
| 10 | my $cls = shift; | ||
| 11 | my ($mode, $file, $fh); | ||
| 12 | my %self = ( | ||
| 13 | 'elements' => [], | ||
| 14 | 'have_read' => 0, | ||
| 15 | 'pos' => 0, | ||
| 16 | ); | ||
| 17 | if (@_ == 1) { | ||
| 18 | $file = shift; | ||
| 19 | if (ref $file) { | ||
| 20 | ($fh, $file) = ($file, "<$file>"); | ||
| 21 | $mode = '<'; | ||
| 22 | } | ||
| 23 | elsif ($file =~ s/^([<>+]+)//) { | ||
| 24 | $mode = $1; | ||
| 25 | } | ||
| 26 | else { | ||
| 27 | $mode = '<'; | ||
| 28 | } | ||
| 29 | } | ||
| 30 | elsif (@_ == 2) { | ||
| 31 | ($mode, $file) = @_; | ||
| 32 | } | ||
| 33 | if (!defined $fh) { | ||
| 34 | open $fh, $mode, $file or die "Can't open $file: $!"; | ||
| 35 | } | ||
| 36 | @self{qw(file fh)} = ($file, $fh); | ||
| 37 | bless \%self, $cls; | ||
| 38 | } | ||
| 39 | |||
| 40 | sub iter { | ||
| 41 | my ($self) = @_; | ||
| 42 | my @elems = @{ $self->{'elements'} }; | ||
| 43 | return sub { | ||
| 44 | return shift @elems if @elems; | ||
| 45 | return if $self->{'have_read'} & TAIL; | ||
| 46 | return $self->_read_one; | ||
| 47 | } | ||
| 48 | } | ||
| 49 | |||
| 50 | sub write { | ||
| 51 | my $self = shift; | ||
| 52 | return if !@_; | ||
| 53 | my ($fh, $elems, $pos) = @$self{qw(fh elements pos)}; | ||
| 54 | my @elems = @$elems; | ||
| 55 | splice @elems, $pos, @elems - $pos, @_ if $pos > 0; | ||
| 56 | my @pars = map { _hash2par($_) } @_; | ||
| 57 | print $fh @pars; | ||
| 58 | my $ofs = tell $fh; | ||
| 59 | truncate $fh, $ofs; | ||
| 60 | push @elems, @_; | ||
| 61 | $self->{'pos'} = @elems; | ||
| 62 | @$elems = @elems; | ||
| 63 | return $self; | ||
| 64 | } | ||
| 65 | |||
| 66 | sub append { | ||
| 67 | my $self = shift; | ||
| 68 | return if !@_; | ||
| 69 | my ($fh, $elems, $pos) = @$self{qw(fh elements pos)}; | ||
| 70 | seek $fh, 0, 2 or die "Can't seek to end: $!"; | ||
| 71 | my @elems = @$elems; | ||
| 72 | push @elems, @_; | ||
| 73 | my @pars = map { _hash2par($_) } @_; | ||
| 74 | print $fh @pars; | ||
| 75 | $self->{'pos'} = @$elems; | ||
| 76 | @$elems = @elems; | ||
| 77 | return $self; | ||
| 78 | } | ||
| 79 | |||
| 80 | sub truncate { | ||
| 81 | my ($self) = @_; | ||
| 82 | my ($fh, $elems, $pos) = @$self{qw(fh elements pos)}; | ||
| 83 | my $ofs = tell $fh; | ||
| 84 | truncate $fh, $ofs; | ||
| 85 | splice @$elems, $pos; | ||
| 86 | return $self; | ||
| 87 | } | ||
| 88 | |||
| 89 | sub _hash2par { | ||
| 90 | my ($hash) = @_; | ||
| 91 | my $par = ''; | ||
| 92 | my %hash = %$hash; | ||
| 93 | my ($at, $lb) = delete @hash{'@','#'}; | ||
| 94 | if (defined $at) { | ||
| 95 | $par .= '@' . $at; | ||
| 96 | $par .= ' ' . $lb if defined $lb; | ||
| 97 | $par .= "\n"; | ||
| 98 | } | ||
| 99 | foreach my $k (sort keys %hash) { | ||
| 100 | my $v = $hash{$k}; | ||
| 101 | $par .= "$k $v\n" if defined $v; | ||
| 102 | } | ||
| 103 | $par = "#empty\n" if !length $par; | ||
| 104 | return $par . "\n"; | ||
| 105 | } | ||
| 106 | |||
| 107 | sub _par2hash { | ||
| 108 | my ($par) = @_; | ||
| 109 | chomp $par; | ||
| 110 | my %hash; | ||
| 111 | foreach (split /\n/, $par) { | ||
| 112 | if (/^[@](\S*)(?: (.+))?$/) { | ||
| 113 | $hash{'@'} = $1 if length $1; | ||
| 114 | $hash{'#'} = $2 if defined $2; | ||
| 115 | } | ||
| 116 | elsif (/^[#]/) { | ||
| 117 | next; | ||
| 118 | } | ||
| 119 | elsif (/^(\S+) (.*)$/) { | ||
| 120 | $hash{$1} = $2; | ||
| 121 | } | ||
| 122 | } | ||
| 123 | return \%hash; | ||
| 124 | } | ||
| 125 | |||
| 126 | sub _read_one { | ||
| 127 | my ($self) = @_; | ||
| 128 | my $fh = $self->{'fh'}; | ||
| 129 | local $/ = ''; | ||
| 130 | my $elems = $self->{'elements'}; | ||
| 131 | if (my $par = <$fh>) { | ||
| 132 | $self->{'pos'}++; | ||
| 133 | my $hash = _par2hash($par); | ||
| 134 | push @{ $elems }, $hash; | ||
| 135 | $self->{'have_read'} |= HEAD; | ||
| 136 | return $hash; | ||
| 137 | } | ||
| 138 | else { | ||
| 139 | $self->{'have_read'} |= TAIL; | ||
| 140 | return; | ||
| 141 | } | ||
| 142 | } | ||
| 143 | |||
| 144 | sub _read_remainder { | ||
| 145 | my ($self) = @_; | ||
| 146 | my ($hash, @rem); | ||
| 147 | while ($hash = $self->_read_one) { | ||
| 148 | push @rem, $hash; | ||
| 149 | } | ||
| 150 | return @rem; | ||
| 151 | } | ||
| 152 | |||
| 153 | sub head { | ||
| 154 | my ($self) = @_; | ||
| 155 | my $elems = $self->{'elements'}; | ||
| 156 | return $self->_read_one if !@$elems; | ||
| 157 | return $elems->[0]; | ||
| 158 | } | ||
| 159 | |||
| 160 | sub tail { | ||
| 161 | my ($self) = @_; | ||
| 162 | my $elems = $self->{'elements'}; | ||
| 163 | $self->_read_remainder if !( $self->{'have_read'} & TAIL ); | ||
| 164 | return if @$elems < 2; | ||
| 165 | return @$elems[1..$#$elems]; | ||
| 166 | } | ||
| 167 | |||
| 168 | sub reset { | ||
| 169 | my ($self) = @_; | ||
| 170 | seek $self->{'fh'}, 0, 0 or die "Can't seek to beginning of file: $!"; | ||
| 171 | $self->{'pos'} = 0; | ||
| 172 | $self->{'elements'} = []; | ||
| 173 | $self->{'have_read'} = 0; | ||
| 174 | return $self; | ||
| 175 | } | ||
| 176 | |||
| 177 | sub elements { | ||
| 178 | my ($self) = @_; | ||
| 179 | my $elems = $self->{'elements'}; | ||
| 180 | $self->_read_remainder if ( $self->{'have_read'} & (HEAD|TAIL) ) != (HEAD|TAIL); | ||
| 181 | return @$elems; | ||
| 182 | } | ||
| 183 | |||
| 184 | 1; | ||
| 185 | |||
| 186 | =pod | ||
| 187 | |||
| 188 | =head1 NAME | ||
| 189 | |||
| 190 | File::Kvpar - read and write files containing key-value paragraphs | ||
| 191 | |||
| 192 | =head1 SYNOPSIS | ||
| 193 | |||
| 194 | $kv = File::Kvpar->new($file); | ||
| 195 | $kv = File::Kvpar->new($mode, $file); # $mode = '<' || '>' || ... | ||
| 196 | $kv = File::Kvpar->new($fh); | ||
| 197 | $iter = $kv->iter; | ||
| 198 | while ($elem = &$iter) { | ||
| 199 | ... | ||
| 200 | } | ||
| 201 | @a = $kv->elements; | ||
| 202 | $a = $kv->head; | ||
| 203 | @a = $kv->tail; | ||
| 204 | foreach ($kv->elements) { | ||
| 205 | ... | ||
| 206 | } | ||
| 207 | |||
| 208 | =head1 DESCRIPTION | ||
| 209 | |||
| 210 | A B<kvpar> file consists of zero or more paragraphs (delimited by blank lines). | ||
| 211 | Each line in a paragraph has the form C<< I<KEY> I<VAL> >>; the key and value | ||
| 212 | are separated by a single space (ASCII character 32). | ||
| 213 | |||
| 214 | A single element may have the form C<< @I<STR> I<VAL> >> -- if it does, the | ||
| 215 | element's hash will be augmented with the key/value pairs | ||
| 216 | ( | ||
| 217 | C<@> => I<STR>, | ||
| 218 | C<#> => I<VAL> | ||
| 219 | ). No further interpretation is made, but one use of this might be to treat | ||
| 220 | I<STR> as the object's B<type> and I<VAL> as the object's B<primary key>. | ||
| 221 | |||
| 222 | =head1 METHODS | ||
| 223 | |||
| 224 | =over 4 | ||
| 225 | |||
| 226 | =item B<new> | ||
| 227 | |||
| 228 | $kv = File::Kvpar->new($file); | ||
| 229 | $kv = File::Kvpar->new($mode, $file); | ||
| 230 | $kv = File::Kvpar->new($fh); | ||
| 231 | |||
| 232 | Open or create a kvpar-formatted file. | ||
| 233 | |||
| 234 | I<$mode> defaults to C<< E<lt> >> if a file path is the only argument. | ||
| 235 | |||
| 236 | For greater control, pass a filehandle as the only argument. Passing both a | ||
| 237 | mode and filehandle will cause an exception to be thrown. | ||
| 238 | |||
| 239 | =item B<iter> | ||
| 240 | |||
| 241 | $iter = $kv->iter; | ||
| 242 | while ($hash = &$iter) { ... } | ||
| 243 | while ($hash = $iter->()) { ... } | ||
| 244 | |||
| 245 | Return a CODE reference that may be invoked to obtain the next element in the | ||
| 246 | file. It returns the undefined value if all elements have been read. | ||
| 247 | |||
| 248 | =item B<elements> | ||
| 249 | |||
| 250 | Return a list (not an array reference) of all elements in the file. Any unread | ||
| 251 | elements will be read. | ||
| 252 | |||
| 253 | =item B<head> | ||
| 254 | |||
| 255 | Return the first element in the file. It will be read if it hasn't already been. | ||
| 256 | |||
| 257 | =item B<tail> | ||
| 258 | |||
| 259 | @elems = $kv->tail; | ||
| 260 | |||
| 261 | Return a list of all elements after the first. Any unread elements will be read. | ||
| 262 | |||
| 263 | =item B<write> | ||
| 264 | |||
| 265 | $kv->write(@hashes); | ||
| 266 | |||
| 267 | Write I<@hashes> to the current position in the file. The file will be | ||
| 268 | truncated afterwards, since the alternative would totally mess up files in most | ||
| 269 | cases. The file must have been opened for write access, of course. | ||
| 270 | |||
| 271 | =item B<append> | ||
| 272 | |||
| 273 | $kv->append(@hashes); | ||
| 274 | |||
| 275 | Write I<@hashes> to the end of the file. The file must have been opened for | ||
| 276 | write access, of course. | ||
| 277 | |||
| 278 | =item B<truncate> | ||
| 279 | |||
| 280 | $kv = File::Kvpar->new('<+', $file); | ||
| 281 | $kv->head; # Read and discard the first element | ||
| 282 | $kv->truncate; | ||
| 283 | $kv->write({...}); | ||
| 284 | |||
| 285 | Truncate the file at the current position and discard any elements that may | ||
| 286 | have been read at or after the current position. If nothing has been read from | ||
| 287 | the file, the file will become empty. | ||
| 288 | |||
| 289 | =item B<reset> | ||
| 290 | |||
| 291 | $kv->reset; | ||
| 292 | |||
| 293 | Set the file position to 0 -- i.e., seek to the beginning of the file -- and | ||
| 294 | discard any elements that have already been read. If the file has been opened | ||
| 295 | for write access, the next B<write> will truncate it. If it has been opened | ||
| 296 | for append access, the next B<append> will wreak havoc unless you call | ||
| 297 | B<truncate> first. | ||
| 298 | |||
| 299 | This will throw an exception if the underlying filehandle is not seekable. | ||
| 300 | |||
| 301 | =back | ||
| 302 | |||
| 303 | =cut | ||
| 304 |