Embperl::Syntax - base class for defining custom syntaxes
Embperl::Syntax provides a base class from which all custom
syntaxes should be derived. Currently Embperl comes with the following
derived syntaxes:
- EmbperlHTML
- all the HTML tag that Embperl recognizes by default
- EmbperlBlocks
- all the [ ] blocks that Embperl supports
- Embperl
- The default syntax; is derived from
"EmbperlHtml" and
"EmbperlBlocks"
- ASP
- <% %> and <%= %>, see perldoc Embperl::Syntax::ASP
- SSI
- Server Side Includes, see perldoc Embperl::Syntax::SSI
- Perl
- File contains pure Perl (similar to Apache::Registry), but can be used
inside EmbperlObject
- Text
- File contains only Text, no actions is taken on the Text
- Mail
- Defines the <mail:send> tag, for sending mail. This is an example
for a taglib, which could be a base for writing your own taglib to extent
the number of available tags
- POD
- Parses POD out of any file and creates a XML tree similar to pod2xml,
which can be formatted by XSLT afterwards.
You can choose which syntax is used inside your page, either by
the "EMBPERL_SYNTAX" configuration
directive, the "syntax", parameter to
"Execute" or the "[$
syntax $]" metacommand.
You can also specify multiple syntaxes e.g.
PerlSetEnv EMBPERL_SYNTAX "Embperl SSI"
Execute ({inputfile => '*', syntax => 'Embperl ASP'}) ;
The syntax metacommand allows you to switch the syntax or to add
or subtract syntaxes e.g.
[$ syntax + Mail $]
will add the Mail taglib so the <mail:send> tag is available
after this line.
[$ syntax - Mail $]
now the <mail:send> tag is unknown again
[$ syntax SSI $]
now you can only use SSI commands inside your page.
If you want to define your own syntax, you have to derive a new
class from one of the existing ones and extent it with new
tags/functionality. The best thing is to take a look at the syntax classes
that comes with Embperl. (inside the directory Embperl/Syntax/).
For example if you want to add new html tags, derive from
Embperl::Syntax::HTML, if you want to add new metacommands derive
from Embperl::Syntax::EmbperlBlocks.
Some of the classes define additional methods to easily add new
tags. See the respective pod file, which methods are available for a certain
class.
Embperl::Syntax defines the basic methods to create a
syntax:
Create a new syntax class. This method should only be called
inside a constructor of a derived class.
$self -> AddToRoot ($elements)
This adds a new element to the root of the parser tree.
$elements must be a hashref. See
Embperl::Syntax::ASP for an example.
$self -> AddInitCode ($compiletimecode, $initcode, $termcode,
$procinfo)
This gives you the possibility to add some Perl code, that is
always executed at the beginning of a document
($initcode), at the end of the document
($termcode) or at compile time
($compiletimecode). The three strings must be valid
Perl code. See Embperl::Syntax::SSI for an example.
$procinfo is a hashref that can consits of
additional processor infos (see below) for the document.
$self -> GetRoot
Returns the root of the parser tree.
Returns a syntax object which is build form the syntaxes named in
$name. If $oldname is given,
$name can start with a
"+" or
"-" to add or subtract a syntax. This is
normally only needed by Embperl itself or to implement a syntax switch
statement (see Embperl::Syntax::SSI for an example.)
$self -> CloneHash ($old, $replace)
Clones a hash which is given as hashref in
$old, optional replace the tags given in the hashref
$replace and return a hashref to the new hash.
Internaly the syntax object builds a data structure which serve as
base for the parser. This structure consists of a list of tokens and
options, which starts with a dash:
- '-lsearch' => 1
- Do an linear search instead of a binary search. This is necessary if the
tokens can't clearly separated.
- '-defnodetype' => ntypText,
- Defines the default type for text nodes. Without any specification the
type is CDATA, which mean no escaping takes places. With
"ntypText" all special characters are
escaped.
- '-rootnode'
- Name for a root node to insert always.
- <name> => \%tokendescription
- All items which does not start with a slash are treated as names. The name
of a token is only descriptive and is used in error messages. The item
must contain a hashref which describes the token.
Each token can have the following members:
- 'text' => '<'
- Start text
- 'end' => '>'
- End text
- 'matchall'
- when set to 1 new token starts at next character, when set to -1 new token
starts at next character, but only if it is the first token inside another
one.
- 'nodename'
- Text that should be outputed when node is stringifyed. Defaults to text.
If the first character is a ':' you can specify the sourounding delimiters
for this tag with :<start>:<end>:<text>:<endtag>.
Example: ':{:}:NAME' . If the nodename starts with a '!' a unique internal
id is generated, so two or more nodename of the same text, can have
different meaning in different contexts.
- 'contains' =>
'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_0123456789'
- Token consists of the following characters. Either
"start" and
"end" or
"contains" could be specified.
NOTE: If a item that only specfifies contains but no text
should be compiled, you must specfify a nodname.
- 'unescape' => 1
- If "optRawInput" isn't set unescape the
data of the inside the node
- 'nodetype' => ntypEndTag
- Type of the node
- 'cdatatype' => ntypAttrValue
- Type of nodes for data (which is not matched by 'inside' definitions)
inside this node. Set to zero to not generate any nodes for text inside of
this node, other then these that are matched by a 'inside'
definition.
- 'endtag'
- Name of the tag that marks the end of a block. This is used by the parser
to track correct nesting.
- 'follow' => \%tokenlist
- Hashref that specifices one or more tokens that must follow this
token.
- 'inside' => \%tokenlist
- Hashref that specifices one or more tokens that could occur inside a node
that is started with this token.
- exitinside
- when the token found, the parser stop searching in the current level and
continues with the tokens that are defined in the hash from there the
current one was "called" via inside
- donteat
- set to 1 to don't eat the start text, so it will be matched again, by any
tokens set under "inside". Set 2 to
don't the end text. Set to 3 for both.
- 'procinfo' =>
- Processor info. Hashref with information how to process this token.
The processor info gives information how to compile this token to
valid code that can be executed later on by the processor. There could be
information for multiple processors. At the moment only the embperl
processor is defined. Normally you must not worry about different processor,
because the syntax object knows inside that all procinfo is for the
embperl processor. procinfo is a parameter to many methods, it
is a hashref and can take the following items:
- perlcode =>
<string> or <arrayref>
- Code to generate. You can also specify a arrayref of strings. The first
string which contains matching attributes are used. The following special
strings are replaced:
- %#<N>%
- Text of childnode number <N> (starting with zero)
- %><N>%
- Text of sibling node number <N> . 0 gives the current node, > 0
gives the Nth next node, < 0 gives the Nth previous node.
- %&<attr>%
- Value of attribute <attr>.
- %^<stackname>%
- Stringvalue of given stack
- %?<stackname>%
- Set if stackvalue was used
- %$n%
- Source Dom Tree, Index of current node.
- %$t%
- Source Dom Tree
- %$x%
- Index of current node
- %$l%
- Index of last node
- %$c%
- Sets the current node Index, if not already done
- %$q%
- Index of source Dom Tree
- %$p%
- Number of current checkpoint
- %%
- Gives a single %
All of the above special values (expect those start with $) allows
the following modifiers:
- %<X>*<N>%
- Attribute/Child etc. must exist.
- %<X>!<N>%
- Attribute/Child etc. must not exist.
- %<X>=<N>:<value1>|<value2>|<value3>%
- Attribute/Child etc. must have the value = <value1> or
<value2> etc.
- %<X>~<N>:<value1>|<value2>|<value3>%
- Attribute/Child etc. must contain the substring <value1> or
<value2> etc. and a non alphanum character must follow the
substring.
writing a minus sign (-) after * ! = or ~ will cause the
child/attribute not to be included, but the condition is evaluated. Writing
an ' will cause the value to be quoted.
- perlcodeend
=> <string>
- Code to generate at the end of the block.
- compiletimeperlcode
=> <string> or <arrayref>
- Code that is executed at compile time. You can also specify a arrayref of
string. The first string which contains matching attributes are used. The
same special strings are replaced as in
"perlcode".
$_[0] contains the Embperl request
object. The method "Code" can be used
to get or set the perl code that should be generated by this node.
If the code begins with #!- all newlines are removed in the
code. This is basically useful to keep all code on the same line, so the
line number in error reporting matches the line in the source.
- compiletimeperlcodeend
=> <string>
- Code that is executed at compile time, but at the end of the tag. The same
special strings are replaced as in
"perlcode".
$_[0] contains the Embperl request
object. The method "Code" can be used
to get or set the perl code that should be generated by this node.
If the code begins with #!- all newlines are removed in the
code. This is basically useful to keep all code on the same line, so the
linenumber in error reporting matches the line in the source.
- perlcoderemove
=> 0/1
- Remove perlcode if perlcodeend condition is not met.
- removenode =>
<removelevel>
- Remove node after compiling. <removelevel> could be one of the
following, values could be added:
- 1.
- Remove this node only
- 2.
- Remove next node if it consists of only white spaces and optKeepSpaces
isn't set.
- 3.
- Replace next node with one space if next node consists only of white
spaces and optKeepSpaces isn't set.
- 4.
- Set this node to ignore for output.
- 5.
- Remove all child nodes
- 6.
- Set all child nodes to ignore for output.
- 7.
- Calculate Attributes values of this node also for nodes that are set to
ignore for output (makes only sense if 8 is also set).
- removespaces
=> <removeflags>
- Remove spaces before or after tag.
- 1.
- Remove all white spaces before tag
- 2.
- Remove all white spaces after tag
- 3.
- Remove spaces and tabs before tag
- 4.
- Remove spaces and tabs after tag
- 5.
- Remove all spaces and tabs but one before tag
- 6.
- Remove all whihe space after text inside of tag
- 7.
- Remove spaces and tabs after text inside of tag
- mayjump =>
0/1
- If set, tells the compiler that this code may jump to another program
location. (e.g. if, while, goto etc.). Could also be a condition as
described under perlcode.
- compilechilds
=> 0/1
- Compile child nodes. Default: 1
- stackname =>
<name>
- Name of stack for "push",
"stackmatch"
- stackname2 =>
<name>
- Name of stack for "push2"
- push =>
<value>
- Push value on stack which name is given with
"stackname". Value could include the
same specical values as "perlcode"
- push2 =>
<value>
- Push value on stack which name is given with
"stackname2". Value could include the
same specical values as "perlcode"
- stackmatch =>
<value>
- Check if value on stack which name is given with
"stackname" is the same as the given
value. If not give a error message about tag mismatch. Value could include
the same specical values as
"perlcode"
- switchcodetype
=> <1/2>
- 1 means put the following code into normal code which is executed every
time the page is requested
2 means put the following code put into code which is executed
direct after compilation. This is mainly for defining subs, or using
modules etc.
- addflags
- cdatatype
- forcetype
- insidemustexist
- matchall
- exitinside
- addfirstchild
- starttag
- endtag
- parsetimeperlcode
- contains