asin:4774133310
WEB+DB PRESS vol. 42
404 Blog Not Found:perl - Text::Darts 0.02 Released!
でも余裕があればText::Txとかも作ってみたいところ。

眠れないので勢いにまかせて作っちゃいましたよ岡野原さん。

svn co http://svn.coderepos.org/share/lang/perl/Text-Tx

で取って下さい。

CPANに上げるのはまだ。理由はこれから書きます。

CPANにまだ上げない理由その一。txはlibraryとして素直に使うにはちょっと問題があるのです。

tx.cpp 
140:    fprintf(stderr,"elemNum:%d nodeNum:%d\n",elemNum,nodeNum);

これ、tx::read()の一部ですが、成功した場合もこれがstderrに出力されます。デバッグモードとかならとにかく、libraryでこれはいかがなものか。現時点ではText::Txはこれを回避するために

lib/Text/Tx.pm 
sub open{
    my $pkg = shift;
    my $filename = shift;
    my $errfh;
    if (!$DEBUG){
        # temporarily close STDERR to suppress messages from tx_tool::tx::read;
        open $errfh, '>&', \*STDERR or die "Can't dup STDERR: $!";
        close STDERR;
    }
    my $dpi = xs_open($filename);
    if (!$DEBUG){
        # and restore STDERR
        open STDERR, '>&', $errfh or die "Can't dup OLDERR: $!";
    }
    carp __PACKAGE__, " cannot open $filename" unless $dpi;
    bless \$dpi, $pkg;
}

という具合に、一時的にstderrを対比させるというアクロバットをしこんでいます。これはきもい。

もう一つは、なぜか Mac OS X v10.4.11 の gcc 4.0できちんとcompileしないこと。署名の後に様子を張っておきますが、なんだかむずがってます。

とはいうものの、Txは元のWord ListよりもTrieが小さくなるというのが涙が出るほど嬉しい。

#!/usr/local/bin/perl
use strict;
use warnings;
use blib './Text-Tx/trunk';
use Text::Tx;

my $tx = Text::Tx->open("words.tx");
sub say { print @_, "\n" };
say $tx->gsub("The quick brown fox jumps over the black lazy dog",
                     sub{"<$_[0]>"});
print STDERR "wow\n"; # make sure STDERR is restored
CORE::dump if @ARGV and shift eq 'dump';

ここでwords.txは/usr/share/dict/wordsをtxbuildしたものですが、これを動かしてわざとcoreを吐かせたところ、FreeBSD 6-Stableで2512kB。

%wc /usr/share/dict/words 
  235882  235882 2493082 /usr/share/dict/words

ですから、これならmod_perlとかに組み込みで使うにも安心感があります。

Enjoy!

Dan the Perl Monger

OS Xでのmakeの結果 
cp lib/Text/Tx.pm blib/lib/Text/Tx.pm
/usr/local/bin/perl /usr/local/lib/perl5/5.10.0/ExtUtils/xsubpp  -C++ -typemap /usr/local/lib/perl5/5.10.0/ExtUtils/typemap  Tx.xs > Tx.xsc && mv Tx.xsc Tx.c
Please specify prototyping behavior for Tx.xs (see perlxs manual)
g++ -c  -I/usr/local/include/tx -fno-common -DPERL_DARWIN -no-cpp-precomp -DDEBUGGING -fno-strict-aliasing -pipe -I/usr/local/include -g -pipe -Os   -DVERSION=\"0.01\" -DXS_VERSION=\"0.01\"  "-I/usr/local/lib/perl5/5.10.0/darwin-thread-multi-2level/CORE"   Tx.c
In file included from /usr/include/c++/4.0.0/bits/basic_ios.h:44,
                 from /usr/include/c++/4.0.0/ios:50,
                 from /usr/include/c++/4.0.0/ostream:44,
                 from /usr/include/c++/4.0.0/iostream:44,
                 from /usr/local/include/tx/tx.hpp:5,
                 from Tx.xs:11:
/usr/include/c++/4.0.0/bits/locale_facets.h:4314:40: error: macro "do_open" requires 7 arguments, but only 2 given
/usr/include/c++/4.0.0/bits/locale_facets.h:4361:34: error: macro "do_close" requires 2 arguments, but only 1 given
/usr/include/c++/4.0.0/bits/locale_facets.h:4380:55: error: macro "do_open" requires 7 arguments, but only 2 given
/usr/include/c++/4.0.0/bits/locale_facets.h:4407:23: error: macro "do_close" requires 2 arguments, but only 1 given
In file included from /usr/include/c++/4.0.0/bits/locale_facets.h:4491,
                 from /usr/include/c++/4.0.0/bits/basic_ios.h:44,
                 from /usr/include/c++/4.0.0/ios:50,
                 from /usr/include/c++/4.0.0/ostream:44,
                 from /usr/include/c++/4.0.0/iostream:44,
                 from /usr/local/include/tx/tx.hpp:5,
                 from Tx.xs:11:
/usr/include/c++/4.0.0/i686-apple-darwin8/bits/messages_members.h:51:38: error: macro "do_open" requires 7 arguments, but only 2 given
/usr/include/c++/4.0.0/i686-apple-darwin8/bits/messages_members.h:60:71: error: macro "do_open" requires 7 arguments, but only 2 given
/usr/include/c++/4.0.0/i686-apple-darwin8/bits/messages_members.h:71:39: error: macro "do_close" requires 2 arguments, but only 1 given
/usr/include/c++/4.0.0/bits/locale_facets.h:4380: error: 'do_open' declared as a 'virtual' field
/usr/include/c++/4.0.0/bits/locale_facets.h:4380: error: expected ';' before 'const'
/usr/include/c++/4.0.0/bits/locale_facets.h:4407: error: variable or field 'do_close' declared void
/usr/include/c++/4.0.0/bits/locale_facets.h:4407: error: 'do_close' declared as a 'virtual' field
/usr/include/c++/4.0.0/bits/locale_facets.h:4407: error: expected ';' before 'const'
/usr/include/c++/4.0.0/i686-apple-darwin8/bits/messages_members.h:60: error: expected initializer before 'const'
/usr/include/c++/4.0.0/i686-apple-darwin8/bits/messages_members.h:71: error: expected initializer before 'const'
make: *** [Tx.o] Error 1