package SQL::Translator::Parser::XML::SQLFairy;
# -------------------------------------------------------------------
-# $Id: SQLFairy.pm,v 1.5 2003-11-19 17:04:18 grommit Exp $
+# $Id: SQLFairy.pm,v 1.6 2004-07-08 19:06:24 grommit Exp $
# -------------------------------------------------------------------
# Copyright (C) 2003 Mark Addison <mark.addison@itn.co.uk>,
#
=head1 NAME
-SQL::Translator::Parser::XML::SQLFairy - parser for SQL::Translator's XML
+SQL::Translator::Parser::XML::SQLFairy - parser for SQL::Translator's XML.
=head1 SYNOPSIS
This parser handles the flavor of XML used natively by the SQLFairy
project (SQL::Translator). The XML must be in the namespace
"http://sqlfairy.sourceforge.net/sqlfairy.xml."
+See L<SQL::Translator::Producer::XML::SQLFairy> for details of this format.
-To see an example of the XML translate one of your schema :) e.g.
+You do not need to specify every attribute of the Schema objects as any missing
+from the XML will be set to their default values. e.g. A field could be written
+using only;
- $ sqlt -f MySQL -t XML-SQLFairy schema.sql
+ <sqlf:field name="email" data_type="varchar" size="255" />
-=head2 attrib_values
+Instead of the full;
-The parser will happily parse XML produced with the attrib_values arg
-set. If it sees a value set as an attribute and a tag, the tag value
-will override that of the attribute.
+ <sqlf:field name="email" data_type="varchar" size="255" is_nullable="1"
+ is_auto_increment="0" is_primary_key="0" is_foreign_key="0" order="4">
+ <sqlf:comments></sqlf:comments>
+ </sqlf:field>
-e.g. For the xml below the table would get the name 'bar'.
-
- <sqlf:table name="foo">
- <sqlf:name>foo</name>
- </sqlf:table>
+If you do not explicitly set the order of items using order attributes on the
+tags then the order the tags appear in the XML will be used.
=head2 default_value
Leave the tag out all together to use the default in Schema::Field.
-Use empty tags or EMPTY_STRING for a zero lenth string. NULL for an
+Use empty tags or 'EMPTY_STRING' for a zero lenth string. 'NULL' for an
explicit null (currently sets default_value to undef in the
Schema::Field obj).
<sqlf:default_value></sqlf:default_value> <!-- Empty string -->
<sqlf:default_value>EMPTY_STRING</sqlf:default_value> <!-- Empty string -->
+ <sqlf:default_value/> <!-- Empty string -->
<sqlf:default_value>NULL</sqlf:default_value> <!-- NULL -->
- <sqlf:default_value/> <!-- Empty string BUT DON'T USE! See BUGS -->
-
=head2 ARGS
Doesn't take any extra parser args at the moment.
use strict;
use vars qw[ $DEBUG $VERSION @EXPORT_OK ];
-$VERSION = sprintf "%d.%02d", q$Revision: 1.5 $ =~ /(\d+)\.(\d+)/;
+$VERSION = sprintf "%d.%02d", q$Revision: 1.6 $ =~ /(\d+)\.(\d+)/;
$DEBUG = 0 unless defined $DEBUG;
use Data::Dumper;
#
my @nodes = $xp->findnodes('/sqlf:schema/sqlf:table');
for my $tblnode (
- sort {
+ sort {
"".$xp->findvalue('sqlf:order|@order',$a)
- <=>
- "".$xp->findvalue('sqlf:order|@order',$b)
+ <=>
+ "".$xp->findvalue('sqlf:order|@order',$b)
} @nodes
) {
debug "Adding table:".$xp->findvalue('sqlf:name',$tblnode);
#
my @nodes = $xp->findnodes('sqlf:fields/sqlf:field',$tblnode);
foreach (
- sort {
+ sort {
("".$xp->findvalue('sqlf:order',$a) || 0)
- <=>
- ("".$xp->findvalue('sqlf:order',$b) || 0)
+ <=>
+ ("".$xp->findvalue('sqlf:order',$b) || 0)
} @nodes
) {
my %fdata = get_tagfields($xp, $_, "sqlf:",
);
if (
- exists $fdata{'default_value'} and
+ exists $fdata{'default_value'} and
defined $fdata{'default_value'}
) {
if ( $fdata{'default_value'} =~ /^\s*NULL\s*$/ ) {
#
# TODO:
- # - We should be able to make the table obj spot this when
+ # - We should be able to make the table obj spot this when
# we use add_field.
# - Deal with $field->extra
#
);
$schema->add_view( %data ) or die $schema->error;
}
-
+
#
# Triggers
#
);
$schema->add_trigger( %data ) or die $schema->error;
}
-
+
#
# Procedures
#
);
$schema->add_procedure( %data ) or die $schema->error;
}
-
+
return 1;
}
# -------------------------------------------------------------------
sub get_tagfields {
#
-# get_tagfields XPNODE, NAMESPACE => qw/TAGNAMES/;
+# get_tagfields XP, NODE, NAMESPACE => qw/TAGNAMES/;
# get_tagfields $node, "sqlf:" => qw/name type fields reference/;
#
-# Returns hash of data. If a tag isn't in the file it is not in this
-# hash.
-# TODO Add handling of and explicit NULL value.
+# Returns hash of data.
+# TODO - Add handling of an explicit NULL value.
#
my ($xp, $node, @names) = @_;
if ( m/:$/ ) { $ns = $_; next; } # Set def namespace
my $thisns = (s/(^.*?:)// ? $1 : $ns);
- foreach my $path ( "\@$thisns$_", "$thisns$_" ) {
- $data{ $_ } = "".$xp->findvalue( $path, $node )
- if $xp->exists( $path, $node );
+ my $is_attrib = m/^sql|comments|action$/ ? 0 : 1;
+ my $attrib_path = "\@$thisns$_";
+ my $tag_path = "$thisns$_";
+ if ( $xp->exists($attrib_path,$node) ) {
+ $data{$_} = "".$xp->findvalue($attrib_path,$node);
+ warn "Use of '$_' as an attribute is depricated."
+ ." Use a child tag instead."
+ ." To convert your file to the new version see the Docs.\n"
+ unless $is_attrib;
+ debug "Got $_=".( defined $data{ $_ } ? $data{ $_ } : 'UNDEF' );
+ }
+ elsif ( $xp->exists($tag_path,$node) ) {
+ $data{$_} = "".$xp->findvalue($tag_path,$node);
+ warn "Use of '$_' as a child tag is depricated."
+ ." Use an attribute instead."
+ ." To convert your file to the new version see the Docs.\n"
+ if $is_attrib;
debug "Got $_=".( defined $data{ $_ } ? $data{ $_ } : 'UNDEF' );
}
}
=head1 BUGS
-B<Empty Tags> e.g. <sqlf:default_value/> Will be parsed as "" and
-hence also false. This is a bit counter intuative for some tags as
-seeing <sqlf:is_nullable /> you might think that it was set when it
-fact it wouldn't be. So for now it is safest not to use them until
-their handling by the parser is defined.
+Ignores the order attribute for Constraints, Views, Indices,
+Views, Triggers and Procedures, using the tag order instead. (This is the order
+output by the SQLFairy XML producer).
=head1 TODO
=over 4
-=item *
-
-Support sqf:options.
-
-=item *
+=item *
-Test forign keys are parsed ok.
+Support options and extra attributes.
-=item *
+=item *
-Sort out sane handling of empty tags <foo/> vs tags with no content
-<foo></foo> vs it no tag being there.
+Test foreign keys are parsed ok.
-=item *
+=item *
Control over defaulting of non-existant tags.