XSH acts as a command interpreter. Individual commands must be separated with a semicolon. Each command may be followed by a pipeline redirection to capture the command's output. In the interactive shell, backslash may be used at the end of line to indicate that the command follows on the next line. A pipeline redirections may be used either to feed the command's output to a unix command or to store it in a XSH string variable. In the first case, the syntax is xsh-command | shell-command ; where xsh-command is any XSH command and shell-command is any command (or code) recognized by the default shell interpreter of the operating system (i.e. on UNIX systems by sh or csh, on Windows systems by cmd). Brackets may be used to join more shell commands (may depend on which shell is used). xsh> ls //words/@* | grep foo | wc In order to store a command's output in a string variable, the pipeline redirection must take the form xsh-command |> $variable where xsh-command is any XSH command and $variable is any valid name for a string variable. xsh> count //words |> $count help command gives a list of all XSH commands. help type gives a list of all argument types. help followed by a command or type name gives more information on the particular command or argument type. XSH is intended to query and manipulate XML and HTML documents. Use one of the open/open-*/create commands to load an XML or HTML document from a local file, external URL (such as http:// or ftp://), string or pipe. While loading, XSH parses and optionally validates (see validation and load-ext-dtd) the document. Parsed documents are stored in memory as DOM trees, that can be navigated and manipulated quite similarly to a local filesystem. Every opened document is associated with an identifier (id), that is a symbolic name for the document in XSH and can be used for example as a prefix of xpath. In the current version, XSH is only able to save documents locally. To store a document on any other location, use ls command and pipe redirection to feed the XML representation of the document to any external program that is able to store it on a remote location. xsh> ls DOC:/ | ssh my.remote.org 'cat > test.xml' backups turn on backup file creation catalog use a catalog file during all parsing processes clone clone a given document close close document (do not save it, though) create make a new document from a given XML fragment documents display a list of open files filename file name id identifier nobackups turn off backup file creation open load an XML, HTML, or Docbook SGML document from a file, pipe or URI process-xinclude load and insert XInclude sections save save a document as XML or HTML select make a given document the current one stream process selected elements from an XML stream (EXPERIMENTAL) switch-to-new-documents set on/off changing current document to newly open/created files With XSH, it is possible to browse document trees as if they were a local filesystem, except that XPath expressions are used instead of ordinary UNIX paths. Current position in the document tree is called the current node. Current node's XPath may be queried with pwd command. In the interactive shell, current node is also displayed in the command line prompt. Remember, that beside cd command, current node (and document) is silently changed by all variant of open command, create command and temporarily also by the node-list variant of the foreach statement. Documents are specified in a similar way as harddrives on DOS/Windows(TM) systems (except that their names are not limitted to one letter in XSH), i.e. by a prefix of the form doc: where doc is the id associated with the document. To mimic the filesystem navigation as closely as possible, XSH contains several commands named by analogy of UNIX filesystem commands, such as cd, ls and pwd. xsh scratch:/> open docA="testA.xml" xsh docB:/> open docB="testB.xml" xsh> pwd docB:/ xsh docB:/> cd docA:/article/chapter[title='Conclusion'] xsh docA:/article/chapter[5]> pwd docA:/article/chapter[5] xsh docA:/article/chapter[5]> cd previous-sibling::chapter xsh docA:/article/chapter[4]> cd .. xsh docA:/article> select docB xsh docB:/> cd change current context node fold mark elements to be folded by list command locate show a given node location (as a cannonical XPath) ls list a given part of a document as XML pwd show current context node location (as a cannonical XPath) register-function define XPath extension function (EXPERIMENTAL) register-namespace register namespace prefix to use XPath expressions register-xhtml-namespace register XHTML namespace prefix to use XPath expressions register-xsh-namespace register XSH namespace prefix to use XPath expressions select make a given document the current one unfold unfold elements folded with fold command unregister-function undefine extension function (EXPERIMENTAL) unregister-namespace unregister namespace prefix xpath XPath expression XSH provides mechanisms not only to browse and inspect the DOM tree but also to modify its content by providing commands for copying, moving, and deleting its nodes as well as adding completely new nodes or XML fragments to it. It is quite easy to learn these commands since their names or aliases mimic their well-known filesystem analogies. On the other hand, many of these commands have two versions one of which is prefixed with a letter "x". This "x" stands for "cross", thus e.g. xcopy should be read as "cross copy". Let's explain the difference on the example of xcopy. When you copy, you have to specify what are you copying and where are you copying to, so you have to specify the source and the target. XSH is very much XPath-based so, XPath is used here to specify both of them. However, there might be more than one node that satisfies an XPath expression. So, the rule of thumb is that the "cross" variant of a command places one and every of the source nodes to the location of one and every destination node, while the plain variant works one-by-one, placing the first source node to the first destination, the second source node to the second destination, and so on (as long as there are both source nodes and destinations left). xsh> create a "<X><A/><Y/><A/></X>"; xsh> create b "<X><B/><C/><B/><C/><B/></X>"; xsh> xcopy a://A replace b://B; xsh> copy b://C before a://A; xsh> ls a:/; <?xml version="1.0" encoding="utf-8"?> <X><C/><A/><Y/><C/><A/></X> xsh> ls b:/; <?xml version="1.0" encoding="utf-8"?> <X><A/><A/><C/><A/><A/><C/><A/><A/></X> As already indicated by the example, another issue of tree modification is the way in which the destination node determines the target location. Should the source node be placed before, after, or into the resulting node? Should it replace it completely? This information has to be given in the location argument that usually precedes the destination XPath. Now, what happens if source and destination nodes are of incompatible types? XSH tries to avoid this by implicitly converting between node types when necessary. For example, if a text, comment, and attribute node is copied into, before or after an attribute node, the original value of the attribute is replaced, prepended or appended respectively with the textual content of the source node. Note however, that element nodes are never converted into text, attribute or any other textual node. There are many combinations here, so try yourself and see the results. You may even use some more sofisticated way to convert between node types, as shown in the following example, where an element is first commented out and than again uncommented. Note, that the particular approach used for resurrecting the commented XML material works only for well-balanced chunks of XML. create doc <<EOF; <?xml version='1.0'?> <book> <chapter> <title>Intro</title> </chapter> <chapter> <title>Rest</title> </chapter> </book> EOF # comment out the first chapter ls //chapter[1] |> $chapter_xml; add comment $chapter_xml replace //chapter[1]; ls / 0; # OUTPUT: <?xml version="1.0"?> <book> <!-- <chapter> <title>Intro</title> </chapter> --> <chapter> <title>Rest</title> </chapter> </book> # un-comment the chapter $comment = string(//comment()[1]); add chunk $comment replace //comment()[1]; ls / 0; # OUTPUT: <?xml version="1.0"?> <book> <chapter> <title>Intro</title> </chapter> <chapter> <title>Rest</title> </chapter> </book> clone clone a given document copy copy nodes (in the one-to-one mode) expression string-like expression insert create a node in on a given target location location relative destination specification (such as after, before, etc.) map quickly modify node value/data using Perl code move move nodes (in the one-to-one mode) node-type node type specification (such as element, attribute, etc.) normalize normalizes adjacent textnodes process-xinclude load and insert XInclude sections remove remove given nodes rename quickly rename nodes with in-line Perl code set-enc set document's charset (encoding) set-standalone set document's standalone flag strip-whitespace strip leading and trailing whitespace xcopy copy nodes (in the all-to-every mode) xinsert create nodes on all target locations xmove move nodes (in the all-to-every mode) xpath XPath expression xslt transform document with XSLT xupdate apply XUpdate commands on a document What a scripting language XSH would be had it not some kind of conditional statements, loops and other stuff that influences the way in which XSH commands are processed. Most notable XSH's feature in this area is that some of the basic flow control statements, namely if, unless, while and foreach have two variants, an XPath-based one and a Perl-based one. The XPath-based variant uses xpath expressions to specify the condition or node-lists to iterate, while the other one utilizes perl-code for this purpose. See descriptions of the individual statements for more detail. call call user-defined routine (macro) command-block single XSH command or a block of XSH commands def sub-routine (macro) declaration exit exit XSH shell foreach loop iterating over a node-list or perl array if if statement ifinclude conditionally include another XSH source in current position include include another XSH source in current position iterate iterate a block over current subtree last immediately exit an enclosing loop next start the next iteration of an enclosing loop prev restart an iteration on a previous node redo restart the innermost enclosing loop block return return from a subroutine run-mode switch into normal execution mode (quit test-mode) stream process selected elements from an XML stream (EXPERIMENTAL) test-mode do not execute any command, only check the syntax throw throw an exception try try/catch statement undef undefine sub-routine (macro) unless negated if statement while simple while loop Beside the possibility to browse the DOM tree and list some parts of it (as described in Navigation), XSH provides commands to obtain other information related to open documents as well as the XSH interpreter itself. These commands are listed bellow. count calculate a given XPath expression and enumerate node-lists defs list all user-defined routines (macros) doc-info displays various information about a document documents display a list of open files dtd show document's DTD enc show document's original character encoding help on-line documentation locate show a given node location (as a cannonical XPath) ls list a given part of a document as XML namespaces List namespaces in current scope (or in scope of given nodes) options list current settings using XSH syntax print print given stuff on standard console output pwd show current context node location (as a cannonical XPath) valid check if the document is valid (according to a DTD, RelaxNG, or XSD schemas) validate validate a document against a DTD, RelaxNG, or XSD schemas variables display a list of defined variables version show version information XSH commands accept different types of arguments, such as usual strings (expression) or XPath expressions. Notably, these two types and types based on them support string variable interpolation. See documentation of the individual types for more information. command-block single XSH command or a block of XSH commands enc-string character encoding (codepage) identifier expression string-like expression filename file name id identifier location relative destination specification (such as after, before, etc.) node-type node type specification (such as element, attribute, etc.) perl-code in-line code in Perl programming language xpath XPath expression In the current version, XSH supports two types of variables: string (scalar) variables and node-list variables. Perl programmers that might miss some other kinds of variables (arrays or hashes) may use the support for interacting with Perl to access these types (see some examples below). These two kinds of variables differ syntactically in the prefix: string variables are prefixed with a dollar sign ($) while node-list variables are prefixed with a percent sign (%). Every string variable name consists of a dollar sign ($) prefix and an id, that has to be unique among other scalar variables, e.g. $variable. Values are assigned to variables either by simple assignments of the form $variable = xpath or by capturing the output of some command with a variable redirection of the form command |> $variable. String variables may be used in string expressions, XPath expressions, or even in perl-code as $id or ${id}. In the first two cases, variables act as macros in the sense that all variables occurences are replaced by the corresponding values before the expression itself is evaluated. To display current value of a variable, use the print command, variables command or simply the variable name: xsh> $b="chapter"; xsh> $file="${b}s.xml"; xsh> open f=$file; xsh> ls //$b[count(descendant::para)>10] xsh> print $b chapter xsh> $b $b='chapter'; xsh> variables $a='chapters.xml'; $b='chapter'; Every string variable name consists of a percent sign (%) prefix and an id, that has to be unique among other node-list variables, e.g. %variable. Node-list variables can be used to store lists of nodes that result from evaluating an XPath. This is especially useful when several changes are performed on some set of nodes and evaluating the XPath expression repeatedly would take too long. Other important use is to remember a node that would otherwise be extremely hard or even impossible to locate by XPath expressions after some changes to the tree structure are made, since such an XPath cannot be predicted in advance. Although node-list variables act just like XPath expressions that would result in the same node-list, for implementation reasons it is not possible to use node-list variables as parts of complex XPath expressions except for one case. They may be only used at the very beginning of an XPath expression. So while constructions such as %creatures[4], %creatures[@race='elf'], or %creatures/parents/father do work as expected, string(%creatures[2]/@name) //creature[%creatures[2]/@name=@name], or %creatures[@race='elf'][2] do not. In the first two cases it is because node-list variables cannot be evaluated in the middle of an XPath expression. The third case fails because this construction actually translates into a sequence of evaluations of self::*[@race='elf'][2] for each node in the %creatures node-list, which is not equivallent to the intended expression as the [2] filter does not apply to the whole result of %creatures[@race='elf'] at once but rather to the partial results. Fortunatelly, it is usually possible to work around these unsupported constructions quite easily. This is typically done by introducing some more variables as well as using the foreach statement. The following example should provide some idea on how to do this: # work around for $name=string(%creatures[2]/@name) xsh> foreach %creatures[2] $name=string(@name) # work around for ls //creature[%creatures[2]/@name=@name] xsh> ls //creature[$name=@name] # work around for ls %creatures[@race='elf'][2] xsh> %elves = %creatures[@race='elf'] xsh> ls %elves[2] Remember, that when a node is deleted from a tree it is at the same time removed from all node-lists it occurs in. Note also, that unlike string variables, node-list variables can not be (and are not intended to be) directly accessed from Perl code. All XSH string variables are usual Perl scalar variables from the XML::XSH::Map namespace, which is the default namespace for any Perl code evaluated from XSH. Thus it is possible to arbitrarily intermix XSH and Perl assignments: xsh> ls //chapter[1]/title <title>Introduction</title> xsh> $a=string(//chapter[1]/title) xsh> eval { $b="CHAPTER 1: ".uc($a); } xsh> print $b CHAPTER 1: INTRODUCTION If needed, it is, however, possible to use any other type of Perl variables by means of evaluating a corresponding perl code. The following example demonstrates using Perl hashes to collect and print some simple racial statistics about the population of Middle-Earth: foreach a:/middle-earth/creature { $race=string(@race); eval { $races{$race}++ }; } print "Middle-Earth Population (race/number of creatures)" eval { echo map "$_/$races{$_}\n", sort ($a cmp $b), keys(%races); }; assign variable assignment expression string-like expression id identifier local temporarily assign new value to a variable xpath XPath expression The following commands are used to modify the default behaviour of the XML parser or XSH itself. Some of the commands are switch between two different modes according to a given expression (which is expected to result either in zero or non-zero value). Other commands also working as a flip-flop have their own explicit counterpart (e.g. verbose and quiet or debug and nodebug). This misconsistency is due to historical reasons. The encoding and query-encoding options allow to specify character encoding that should be expected from user as well as the encoding to be used by XSH on output. This is particularly useful when you work with UTF-8 encoded documents on a console which supports only 8-bit characters. The options command displays current settings by means of XSH commands. Thus it can not only be used to review current values, but also to store them future use, e.g. in ~/.xshrc file. xsh> options | cat > ~/.xshrc backups turn on backup file creation debug display many annoying debugging messages empty-tags turn on/off serialization of empty tags enc-string character encoding (codepage) identifier encoding choose output charset indent turn on/off pretty-printing keep-blanks turn on/off ignorable whitespace preservation load-ext-dtd turn on/off external DTD fetching nobackups turn off backup file creation nodebug turn off debugging messages options list current settings using XSH syntax parser-completes-attributes turn on/off parser's ability to fill default attribute values parser-expands-entities turn on/off parser's tendency to expand entities parser-expands-xinclude turn on/off transparent XInclude insertion by parser pedantic-parser make the parser more pedantic query-encoding declare the charset of XSH source files and terminal input quiet turn off many XSH messages recovering turn on/off parser's ability to fix broken XML register-function define XPath extension function (EXPERIMENTAL) register-namespace register namespace prefix to use XPath expressions register-xhtml-namespace register XHTML namespace prefix to use XPath expressions register-xsh-namespace register XSH namespace prefix to use XPath expressions run-mode switch into normal execution mode (quit test-mode) skip-dtd turn on/off serialization of DTD DOCTYPE declaration switch-to-new-documents set on/off changing current document to newly open/created files test-mode do not execute any command, only check the syntax unregister-function undefine extension function (EXPERIMENTAL) unregister-namespace unregister namespace prefix validation turn on/off validation in XML parser verbose make XSH print many messages xpath-axis-completion sets TAB completion for axes in xpath expressions in the interactive mode xpath-completion turn on/off TAB completion for xpath expressions in the interactive mode To allow more complex tasks to be achieved, XSH provides ways for interaction with the Perl programming language and the system shell. Perl is a language optimized for scanning arbitrary text files, extracting information from those text files, and printing reports based on that information. It's also a good language for many system management tasks. The language is intended to be practical (easy to use, efficient, and complete). XSH itself is written in Perl, so it is extremely easy to support this language as an extension to XSH. Perl expressions or blocks of code can either be simply evaluated with the perl command, used to do quick changes to nodes of the DOM tree (see map command), used to provide list of strings to iterate over in a foreach loop, or to specify more complex conditions for if, unless, and while statements. To prevent conflict between XSH internals and the evaluated perl code, XSH runs such code in the context of a special namespace XML::XSH::Map. As described in the section Variables, XSH string variables may be accessed and possibly assigned from Perl code in the most obvious way, since they actually are Perl variables defined in the XML::XSH::Map namespace. The interaction between XSH and Perl actually works also the other way round, so that you may call back XSH from the evaluated Perl code. For this, Perl function xsh is defined in the XML::XSH::Map namespace. All parameters passed to this function are interpreted as XSH commands. To simplify evaluation of XPath expressions, another three functions: The first one, named count, returns the same value as would be printed by count command in XSH on the same XPath expression. The second function, named literal, returns the result of XPath evaluation as if the whole expression was wrapped with the XPath string() function. In other words, literal('doc:expression') returns the same value as count('doc:string(expression)'). The third function, named xml_list, returns the result of the XPath search as a XML string which is equivallent to the output of a ls on the same XPath expression (without indentation and without folding or any other limitation on the depth of the listing). In the following examples we use Perl to populate the Middle-Earth with Hobbits whose names are read from a text file called hobbits.txt, unless there are some Hobbits in Middle-Earth already. unless (//creature[@race='hobbit']) { perl 'open $file, "hobbits.txt"'; perl '@hobbits=<$file>'; perl 'close $file'; foreach { @hobbits } { insert element "<creature name='$__' race='hobbit'>" into m:/middle-earth/creatures; } } perl { unless (count(//creature[@race='hobbit'])) { open my $file, "hobbits.txt"; foreach (<$file>) { xsh(qq{insert element "<creature name='$_' race='hobbit'>" into m:/middle-earth/creatures}); } close $file; } }; XSH allows the user to extend the set of XPath functions by providing an extension function written in Perl. This can be achieved using the register-function command. The perl code implementing an extension function works as a usual perl routine accepting its arguments in @_ and returning the result. The following conventions are used: The arguments passed to the perl implementation by the XPath engine are either simple scalars or XML::LibXML::NodeList objects, depending on the types of the XPath arguments. The implementation is responsible for checking the argument number and types. The implementation may use arbitrary XML::LibXML methods to process the arguments and return the result. (XML::LibXML perl module documentation can be found for example at http://search.cpan.org/author/PHISH/XML-LibXML-1.54/LibXML.pm). The implementation SHOULD NOT, however, MODIFY the document. Doing so could not only confuse the XPath engine but result in an critical error (such as segmentation fault). Calling XSH commands from extension function implementations is not currently allowed. The perl code must return a single value, which can be of one of the following types: a simple scalar (a number or string), XML::LibXML::Boolean object reference (result is a boolean value), XML::LibXML::Literal object reference (result is a string), XML::LibXML::Number object reference (resulat is a float), XML::LibXML::Node (or derived) object reference (result is a nodeset consisting of a single node), or XML::LibXML::NodeList (result is a nodeset). For convenience, simple (non-blessed) array references consisting of XML::LibXML::Node objects can also be used for a nodeset result instead of a XML::LibXML::NodeList. In the interactive mode, XSH interprets all lines starting with a exclamation mark (!) as shell commands and invokes the system shell to interpret them (this is to mimic FTP command-line interpreters). xsh> !ls -l -rw-rw-r-- 1 pajas pajas 6355 Mar 14 17:08 Artistic drwxrwxr-x 2 pajas users 128 Sep 1 10:09 CVS -rw-r--r-- 1 pajas pajas 14859 Aug 26 15:19 ChangeLog -rw-r--r-- 1 pajas pajas 2220 Mar 14 17:03 INSTALL -rw-r--r-- 1 pajas pajas 18009 Jul 15 17:35 LICENSE -rw-rw-r-- 1 pajas pajas 417 May 9 15:16 MANIFEST -rw-rw-r-- 1 pajas pajas 126 May 9 15:16 MANIFEST.SKIP -rw-r--r-- 1 pajas pajas 20424 Sep 1 11:04 Makefile -rw-r--r-- 1 pajas pajas 914 Aug 26 14:32 Makefile.PL -rw-r--r-- 1 pajas pajas 1910 Mar 14 17:17 README -rw-r--r-- 1 pajas pajas 438 Aug 27 13:51 TODO drwxrwxr-x 5 pajas users 120 Jun 15 10:35 blib drwxrwxr-x 3 pajas users 1160 Sep 1 10:09 examples drwxrwxr-x 4 pajas users 96 Jun 15 10:35 lib -rw-rw-r-- 1 pajas pajas 0 Sep 1 16:23 pm_to_blib drwxrwxr-x 4 pajas users 584 Sep 1 21:18 src drwxrwxr-x 3 pajas users 136 Sep 1 10:09 t -rw-rw-r-- 1 pajas pajas 50 Jun 16 00:06 test drwxrwxr-x 3 pajas users 496 Sep 1 20:18 tools -rwxr-xr-x 1 pajas pajas 5104 Aug 30 17:08 xsh To invoke a system shell command or program from the non-interactive mode or from a complex XSH construction, use the exec command. Since UNIX shell commands are very powerful tool for processing textual data, XSH supports direct redirection of XSH commands output to system shell command. This is very similarly to the redirection known from UNIX shells, except that here, of course, the first command in the pipe-line colone is an XSH command. Since semicolon (;) is used in XSH to separate commands, it has to be prefixed with a backslash if it should be used for other purposes. xsh> ls //chapter[5]/para | grep funny | less xsh> ls //chapter[5]/para | find "funny" | more exec execute a shell command expression string-like expression lcd change system working directory map quickly modify node value/data using Perl code perl evaluate in-line Perl code perl-code in-line code in Perl programming language rename quickly rename nodes with in-line Perl code assign $id=xpath $id=xpath assign %id=xpath %id=xpath In the first two cases (where dollar sign appears) store the result of evaluation of the xpath in a variable named $id. In this case, xpath is evaluated in a simmilar way as in the case of the count: if it results in a literal value this value is used. If it results in a node-list, number of nodes occuring in that node-list is used. Use the string() XPath function to obtain a literal values in these cases. xsh> $a=string(chapter/title) xsh> $b="hallo world" xsh> $a=5*100 xsh> $a $a=500 xsh> $a=(($a+5) div 10) xsh> $a $a=50.5 xsh> $a=//chapter xsh> $a $a=10 xsh> %chapters=//chapter xsh> $a=%chapters xsh> $a $a=10 xsh> ls ./creature <creature race='hobbit' name="Bilbo"/> ## WRONG (@name results in a singleton node-list) !!! xsh> $name=@name xsh> $name $name=1 ## CORRECT (use string() function) xsh> $name=string(@name) xsh> $name $name=Bilbo In the other two cases (where percent sign appears) find all nodes matching a given xpath and store the resulting node-list in the variable named %id. The variable may be later used instead of an XPath expression. variablesbackups Enable creating backup files on save (default). This command is equivalent to setting the $BACKUPS variable to 1. nobackupscall id [xpath | expression]* Call an XSH subroutine named id previously created using def. If the subroutine requires some paramters, these have to be specified after the id. Node-list parameters are given by means of xpath expressions. String parameters have to be string expressions. def, returncatalog expression Will use a given catalog file as a catalog during all parsing processes. Using a catalog will significantly speed up parsing processes if many external ressources are loaded into the parsed documents (such as DTDs or XIncludes) cd [xpath] chxpath Change current context node (and current document) to the first node matching a given xpath argument. clone id=id dup Make a copy of the document identified by the id following the equal sign assigning it the identifier of the first id. open, close, enc, documentsclose [id] Close the document identified by id, removing its parse-tree from memory (note also that all nodes belonging to the document are removed from all nodelists they appear in). If id is omitted, the command closes the current document. copy xpath location xpath cp Copies nodes matching the first xpath to the destinations determined by the location directive relative to the second xpath. If more than one node matches the first xpath than it is copied to the position relative to the corresponding node matched by the second xpath according to the order in which are nodes matched. Thus, the n'th node matching the first xpath is copied to the location relative to the n'th node matching the second xpath. The possible values for location are: after, before, into, replace and cause copying the source nodes after, before, into (as the last child-node). the destination nodes. If replace location is used, the source node is copied before the destination node and the destination node is removed. Some kind of type conversion is used when the types of the source and destination nodes are not equal. Thus, text, cdata, comment or processing instruction node data prepend, append or replace value of a destination attribute when copied before,after/into or instead (replace) an attribute, and vice versa. Attributes may be copied after, before or into some other attribute to append, prepend or replace the destination attribute value. They may also replace the destination attribute completely (both its name and value). To copy an attribute from one element to another, simply copy the attribute node into the destination element. Elements may be copied into other elements (which results in appending the child-list of the destination element), or before, after or instead (replace) other nodes of any type except attributes. xsh> copy a://creature replace b://living-thing count xpath print_value, get Calculate a given xpath expression. If the result is a node-list, return number of nodes in the node-list. If the xpath results in a boolean, numeric or literal value, return the value. create id expression new Create a new document using expression to form the root element and associate it with a given identifier. xsh> create t1 root xsh> ls / <?xml version="1.0" encoding="utf-8"?> <root/> xsh> create t2 "<root id='r0'>Just a <b>test</b></root>" xsh> ls / <?xml version="1.0" encoding="utf-8"?> <root id='r0'>Just a <b>test</b></root> xsh> files scratch = new_document.xml t1 = new_document1.xml t2 = new_document2.xml open, clonedebug Turn on debugging messages. This is equivalent to setting $DEBUG variable to 1. nodebugdef id [$id | %id]* command-block or def id [$id | %id]*; define Define a new XSH subroutine named id. The subroutine may require zero or more parameters of nodelist or string type. These are declared as a whitespace-separated list of (so called) parametric variables (of nodelist or string type). The body of the subroutine is specified as a command-block. Note, that all subroutine declarations are processed during the parsing and not at run-time, so it does not matter where the subroutine is defined. The routine can be later invoked using the call command followed by the routine name and parameters. Nodelist parameters must be given as an XPath expressions, and are evaluated just before the subroutine's body is executed. String parameters must be given as (string) expressions. Resulting node-lists/strings are stored into the parametric variables before the body is executed. These variables are local to the subroutine's call tree (see also the local command). If there is a global variable using the same name as some parametric variable, the original value of the global variable is replaced with the value of the parametric variable for the time of the subroutine's run-time. Note that subroutine has to be declared before it is called with call. If you cannot do so, e.g. if you want to call a subroutine recursively, you have to pre-declare the subroutine using a def with no command-block. There may be only one full declaration (and possibly one pre-declaration) of a subroutine for one id and the declaration and pre-declaration has to define the same number of arguments and their types must match. def l3 %v { ls %v 3; # list given nodes upto depth 3 } call l3 //chapter; def comment %n # nodes to move to comments $mark # maybe some handy mark to recognize such comments { foreach %n { if ( . = ../@* ) { echo "Warning: attribute nodes are not supported!"; } else { echo "Commenting out:"; ls .; local $node = ""; ls . |> $node; add comment "$mark$node" replace .; } } } def uncomment %n $mark { foreach %n { if (. = ../comment()) { # is this node a comment node local $string = substring-after(.,"$mark"); add chunk $string replace .; } else { echo "Warning: Ignoring non-comment node:"; ls . 0; } } } # comment out all chapters with no paragraphs call comment //chapter[not(para)] "COMMENT-NOPARA"; # uncomment all comments (may not always be valid!) $mark="COMMENT-NOPARA"; call uncomment //comment()[starts-with(.,"$mark")] $mark; call, return, localdefs List names and parametric variables for all defined XSH routines. def, variablesdoc-info [expression] doc_info In the present implementation, this command displays information provided in the <?xml ...?> declaration of a document: version, encoding, standalone, plus information about level of gzip compression of the original XML file. set-enc, set-standalonefiles files, docs List open files and their identifiers. open, closedtd [id] Print external or internal DTD for a given document. If no document identifier is given, the current document is used. valid, validateempty-tags expression empty_tags If the value of expression is 1 (non-zero), empty tags are serialized as a start-tag/end-tag pair (<foo></foo>). This option affects both ls and save and possibly other commands. Otherwise, they are compacted into a short-tag form (<foo/>). Default value is 0. This command is equivalent to setting the $EMPTY_TAGS variable. enc [id] Print the original document encoding string. If no document identifier is given, the current document is used. set-encencoding enc-string Set the default output character encoding. This command is equivalent to setting the $ENCODING variable. exec expression [expression ...] system execute the system command(s) in expressions. exec echo hallo world; # prints hallo world exec "echo hallo word" | wc; # counts words in hallo world exec uname; # prints operating system name exit [expression] quit Exit xsh immediately, optionally with the exit-code resulting from a given expression. WARNING: No files are saved on exit. fold xpath [expression] This feature is still EXPERIMENTAL! Fold command may be used to mark elements matching the xpath with a xsh:fold attribute from the http://xsh.sourceforge.net/xsh/ namespace. When listing the DOM tree using ls xpath fold, elements marked in this way are folded to the depth given by the expression (default depth is 0 = fold immediately). xsh> fold //chapter 1 xsh> ls //chapter[1] fold <chapter id="intro" xsh:fold="1"> <title>...</title> <para>...</para> <para>...</para> </chapter> unfold, lsforeach xpath|perl-code command|command-block for If the first argument is an xpath expression, execute the command-block for each node matching the expression making it temporarily the current node, so that all relative XPath expressions are evaluated in its context. If the first argument is a perl-code, it is evaluated and the resulting perl-list is iterated setting the variable $__ (note that there are two underscores!) to be each element of the list in turn. It works much like perl's foreach, except that the variable used consists of two underscores. xsh> foreach //company xmove ./employee into ./staff; xsh> foreach { glob('*.xml') } { open f=$__; list f:/; } help command|argument-type ? Print help on a given command or argument type. if xpath|perl-code command if xpath|perl-code command-block [ elsif command-block ]* [ else command-block ] Execute command-block if a given xpath or perl-code expression evaluates to a non-emtpty node-list, true boolean-value, non-zero number or non-empty literal. If the first test fails, check all possibly following elsif conditions and execute the corresponding command-block for the first one of them which is true. If none of them succeeds, execute the else command-block (if any). def node_type %n { foreach (%n) { if ( . = self::* ) { # XPath trick to check if . is an element echo 'element'; } elsif ( . = ../@* ) { # XPath trick to check if . is an attribute echo 'attribute'; } elsif ( . = ../processing-instruction() ) { echo 'pi'; } elsif ( . = ../text() ) { echo 'text'; } elsif ( . = ../comment() ) { echo 'comment' } else { # well, this should not happen, but anyway, ... echo 'unknown-type'; } } } ifinclude filename Include a file named filename and execute all XSH commands therein unless the file was already included using either include of ifinclude. includeinclude filename . Include a file named filename and execute all XSH commands therein. ifincludeindent expression If the value of expression is 1, format the XML output while saving a document by adding some nice ignorable whitespace. If the value is 2 (or higher), XSH will act as in case of 1, plus it will add a leading and a trailing linebreak to each text node. Note, that since the underlying C library (libxml2) uses a hardcoded indentation of 2 space characters per indentation level, the amount of whitespace used for indentation can not be altered on runtime. This command is equivalent to setting the $INDENT variable. insert node-type expression [namespace expression] location xpath add Works just like xadd, except that the new node is attached only the first node matched. xinsert, move, xmoveiterate xpath command-block Iterate works very much like the XPath variant of foreach, except that iterate evaluates the command-block as soon as a new node matching a given xpath is found. As a limitation, the xpath expresion used with iterate may only consist of one XPath step, i.e. it cannot contain an XPath step separator /. What are the benefits of iterate over a foreach loop, then? Well, under some circumstances it is efficiency, under other there are none. To clarify this, we have to dive a bit deeper into the details of XPath implementation. By definition, the node-list resulting from evaluation of an XPath has to be ordered in the canonical document order. That means that an XPath implementation must contain some kind of a sorting algorithm. This would not itself be much trouble if a relative document order of two nodes of a DOM tree could be determined in a constant time. Unfortunately, the libxml2 library, used behind XSH, does not implement mechanisms that would allow this complexity restriction (which is, however, quite natural and reasonable approach if all the consequences are considered). Thus, when comparing two nodes, libxml2 traverses the tree to find their nearest common ancestor and at that point determines the relative order of the two subtrees by trying to seek one of them in a list of right siblings of the other. This of course cannot be handled in a constant time. As a result, the sorting algorithm, reasonably efficient for a constant time comparison (polynomial of a degree < 1.5) or small node-lists, becomes rather unusable for huge node-lists with linear time comparison (still polynomial but of a degree > 2). The iterate command provides a way to avoid sorting the resulting nodelist by limiting allowed XPath expression to one step (and thus one axis) at a time. On the other hand, since iterate is implemented in Perl, a proxy object glueing the C and Perl layers has to be created for every node the iterator passes by. This (plus some extra subroutine calls) makes it about two to three times slower compared to a similar tree-traversing algorithm used by libxml2 itself during XPath evaluation. Our experience shows that iterate beats foreach in performance on large node-lists (>=1500 nodes, but your milage may vary) while foreach wins on smaller node-lists. The following two examples give equivallent results. However, the one using iterate may be faster esp. if the number of nodes being counted is very large. cd rohan/inhabitants; iterate child::*[@age>=18 and @age<60] { perl $productive++ }; echo "$productive inhabitants in productive age"; $productive=count(rohan/inhabitants/*[@age>=18 and @age<60]); echo "$productive inhabitants in productive age"; Use e.g. | time cut pipe-line redirection to benchmark a XSH command on a UNIX system. foreach, next, prev, lastkeep_blanks expression keep_blanks Allows you to turn off XML::LibXML's default behaviour of maintaining whitespace in the document. Non-zero expression forces the XML parser to preserve all whitespace. This command is equivalent to setting the $KEEP_BLANKS variable. last [expression] The last command is like the break statement in C (as used in loops); it immediately exits an enclosing loop. The optional expression argument may evaluate to a positive integer number that indicates which level of the nested loops to quit. If this argument is omitted, it defaults to 1, i.e. the innermost loop. Using this command outside a subroutine causes an immediate run-time error. foreach, while, iterate, next, lastlcd expression chdir Changes the filesystem working directory to expression, if possible. If expression is omitted, changes to the directory specified in HOME environment variable, if set; if not, changes to the directory specified by LOGDIR environment variable. load_ext_dtd expression load_ext_dtd If the expression is non-zero, XML parser loads external DTD subsets while parsing. By default, this option is enabled. This command is equivalent to setting the $LOAD_EXT_DTD variable. local $id = xpathlocal %id = xpathlocal $id|%id [ $id|%id ... ] This command acts in a very similar way as assign does, except that the variable assignment is done temporarily and lasts only for the rest of the nearest enclosing command-block. At the end of the enclosing block or subroutine the original value is restored. This command may also be used without the assignment part and assignments may be done later using the usual assign command. Note, that the variable itself is not lexically is still global in the sense that it is still visible to any subroutine called subsequently from within the same block. A local just gives temporary values to global (meaning package) variables. Unlike Perl's my declarations it does not create a local variable. This is known as dynamic scoping. Lexical scoping is not implemented in XSH. To sum up for Perl programmers: local in XSH works exactly the same as local in Perl. assign, deflocate xpath Print canonical XPaths leading to nodes matched by a given xpath. pwdls xpath [expression] list List the XML representation of all nodes matching xpath. The optional expression argument may be provided to specify the depth of XML tree listing. If negative, the tree will be listed to unlimited depth. If the expression results in the word fold, elements marked with the fold command are folded, i.e. listed only to a certain depth (this feature is still EXPERIMENTAL!). Unless in quiet mode, this command prints also number of nodes matched on stderr. If the xpath parameter is omitted, current context node is listed to the depth of 1. count, fold, unfoldmap perl-code xpath sed This command provides an easy way to modify node's data (content) using arbitrary Perl code. Each of the nodes matching xpath is passes its data to the perl-code via the $_ variable and receives the (possibly) modified data using the same variable. Since element nodes do not really have any proper content (they are only a storage for other nodes), node's name (tag) is used in case of elements. Note, however, that recent versions of XSH provide a special command rename with a very similar syntax to map, that should be used for renaming element, attribute, and processing instruction nodes. xsh> map { $_=ucfirst($_) } //hobbit/@name xsh> map { s/goblin/orc/gi } //hobbit/tale/text() move xpath location xpath mv move command acts exactly like copy, except that it removes the source nodes after a succesfull copy. Remember that the moved nodes are actually different nodes from the original ones (which may not be obvious when moving nodes within a single document into locations that do not require type conversion). So, after the move, the original nodes do not exist neither in the document itself nor any nodelist variable. See copy for more details on how the copies of the moved nodes are created. copy, xmove, insert, xinsertnamespaces [xpath] For each node matching given xpath lists all namespaces that are valid in its scope in the form of xmlns:prefix="uri" declarations. If no xpath is given, lists namespaces in the scope of the current node. next [expression] The next command is like the continue statement in C; it starts the next iteration of an enclosing loop. The optional expression argument may evaluate to a positive integer number that indicates which level of the nested loops should be restarted. If omitted, it defaults to 1, i.e. the innermost loop. Using this command outside a loop causes an immediate run-time error. foreach, while, iterate, redo, last, prevnobackups Disable creating backup files on save. This command is equivalent to setting the $BACKUPS variable to 0. nobackupsnodebug Turn off debugging messages. This is equivalent to setting $DEBUG variable to 0. debugnormalize xpath normalize puts all text nodes in the full depth of the sub-tree underneath each node selected by a given xpath, into a "normal" form where only structure (e.g., elements, comments, processing instructions, CDATA sections, and entity references) separates text nodes, i.e., there are neither adjacent Text nodes nor empty Text nodes. [open [HTML|XML|DOCBOOK] [FILE|PIPE|STRING]] id=expression Load a new XML, HTML or SGML DOCBOOK document from the file (actually arbitrary URL), command output or string provided by the expression. In XSH the document is given a symbolic name id. To identify the documentin commands like close, save, validate, dtd or enc simply use id. In commands which work on document nodes, give id: prefix to XPath expressions to point the XPath to the document. xsh> open x=mydoc.xml # open a document # open a HTML document from the Internet xsh> open HTML h="http://www.google.com/?q=xsh" # quote file name if it contains whitespace xsh> open y="document with a long name with spaces.xml" # you may omit the word open when loading an XML file/URI. xsh> z=mybook.xml # use HTML or DOCBOOK keywords to load these types xsh> open HTML z=index.htm # use PIPE keyword to read output of a command xsh> open HTML PIPE z='wget -O - xsh.sourceforge.net/index.html' # use z: prefix to identify the document opened with the # previous comand in an XPath expression. xsh> ls z://chapter/title options flags List current values of all XSH flags and options (such as validation flag or query-encoding). xsh> options | cat > ~/.xshrc parser-completes-attributes expression complete_attributes, complete-attributes, parser_completes_attributes If the expression is non-zero, this command allows XML parser to complete the elements attributes lists with the ones defaulted from the DTDs. By default, this option is enabled. This command is equivalent to setting the $PARSER_COMPLETES_ATTRIBUTES variable. parser_expands_entities expression parser_expands_entities Enable the entity expansion during the parse process if the expression is non-zero, disable it otherwise. If entity expansion is off, any external parsed entities in the document are left as entities. Defaults to on. This command is equivalent to setting the $PARSER_EXPANDS_ENTITIES variable. parser_expands_xinclude expression parser_expands_xinclude If the expression is non-zero, the parser is allowed to expand XIinclude tags imidiatly while parsing the document. This command is equivalent to setting the $PARSER_EXPANDS_XINCLUDE variable. process-xincludepedantic_parser expression pedantic_parser If you wish, you can make XML::LibXML more pedantic by passing a non-zero expression to this command. This command is equivalent to setting the $PEDANTIC_PARSER variable. eval perl-code eval Evaluate a given perl expression. countprev [expression] This command is only allowed inside an iterate loop. It returns the iteration one step back, to the previous node on the iterated axis. The optional expression argument may be used to indicate to which level of nested loops the command applies to. iterate, redo, last, nextprint expression [expression ...] echo Interpolate and print a given expression(s). process_xinclude [id] process_xinclude, process-xincludes, process_xincludes, xinclude, xincludes, load_xincludes, load-xincludes, load_xinclude, load-xinclude Process any xinclude tags in the document id. parser-expands-xincludepwd Print XPath leading to the current context node. This is equivalent to locate .. locatequery-encoding enc-string query_encoding Set the default query character encoding. This command is equivalent to setting the $QUERY_ENCODING variable. quiet Turn off verbose messages. This command is equivalent to setting the $QUIET variable. verboserecovering expression Turn on recovering parser mode if the expression is non-zero or off otherwise. Defaults to off. Note, that the in the recovering mode, validation is not performed by the parser even if the validation flag is on and that recovering mode flag only influences parsing of XML documents (not HTML). The recover mode helps to efficiently recover documents that are almost well-formed. This for example includes documents without a close tag for the document element (or any other element inside the document). This command is equivalent to setting the $RECOVERING variable. redo [expression] The redo command restarts a loop block without evaluating the conditional again. The optional expression argument may evaluate to a positive integer number that indicates which level of the nested loops should be restarted. If omitted, it defaults to 1, i.e. the innermost loop. Using this command outside a loop causes an immediate run-time error. while ($i<100) { # ... foreach //para { # some code if $param { redo; # redo foreach loop } else { redo 2; # redo while loop } } } foreach, while, iterate, next, lastregister-function expression perl-code regfunc EXPERIMENTAL! Register given perl code as a new XPath extension function under a name provided in the first argument (expression). XML::LibXML DOM API may be used in the perl code for object processing. If the name contains a colon, then the first part before the colon must be a registered namespace prefix (see register-namespace) and the function is registered within the corresponding namespace. register-namespace expression expression regns Registers the first argument as a prefix for the namespace given in the second argument. The prefix can later be used in XPath expressions. register-xhtml-namespace expression regns-xhtml Registers a prefix for the XHTML namespace. The prefix can later be used in XPath expressions. register-xsh-namespace expression regns-xsh Registers a new prefix for the XSH namespace. The prefix can later be used in XPath expressions. Note, that XSH namespace is by default registered with xsh prefix. This command is thus, in general, useful only when some document uses xsh prefix for a different namespace. remove xpath rm, prune, delete, del Remove all nodes matching xpath. xsh> del //creature[@manner='evil'] rename perl-code xpath This command is very similar to the map command, except that it operates on nodes' names rather than their data/values. For every element, attribute or processing-instruction matched by the xpath expression the following procedure is used: 1) the name of the node is stored into Perl's $_ variable, 2) the perl-code is evaluated, and 3) the (posibly changed) content of the $_ variable is used as a new name for the node. xsh> map $_='halfling' //hobbit xsh> map { $_=uc($_) } (//*|//@*) mapreturn This command immediatelly stops the execution of a procedure it occurs in and returns the execution to the place of the script from which the subroutine was called. Using this command outside a subroutine causes an immediate run-time error. def, callrun-mode run_mode Switch into normal XSH mode in which all commands are executed. This is equivalent to setting $TEST_MODE variable to 0. test-modesave [HTML|XML|XInclude] [FILE|PIPE|STRING] id expression [encoding enc-string] or save id or save Save the document identified by id. Using one of the FILE, PIPE, STRING keywords the user may choose to save the document to a file send it to a given command's input via a pipe or simply return its content as a string. If none of the keywords is used, it defaults to FILE. If saving to a PIPE, the expression argument must provide the coresponding command and all its parameters. If saving to a FILE, the expression argument may provide a filename; if omitted, it defaults to the original filename of the document. If saving to a STRING, the expression argument is ignored and may freely be omitted. The output format is controlled using one of the XML, HTML, XInclude keywords (see below). If the format keyword is ommited, save it defaults to XML. Note, that a document should be saved as HTML only if it actually is a HTML document. Note also, that the optional encoding parameter forces character conversion only; it is up to the user to declare the document encoding in the appropriate HTML <META> tag. The XInclude keyword automatically implies XML format and can be used to force XSH to save all already expanded XInclude sections back to their original files while replacing them with <xi:include> tags in the main XML file. Moreover, all material included within <include> elements from the http://www.w3.org/2001/XInclude namespace is saved to separate files too according to the href attribute, leaving only empty <include> element in the root file. This feature may be used to split the document to new XInclude fragments. The encoding keyword followed by a enc-string can be used to convert the document from its original encoding to a different encoding. In case of XML output, the <?xml?> declaration is changed accordingly. The new encoding is also set as the document encoding for the particular document. save HTML PIPE mydoc 'lynx -stdin' open, close, enc, documentsselect id Make id the document identifier to be used in the next xpath evaluation without identifier prefix. xsh> a=mydoc1.xml# opens and selects a xsh> ls /# lists a xsh> b=mydoc2.xml# opens and selects b xsh> ls /# lists b xsh> ls a:/# lists and selects a xsh> select b# does nothing except selecting b xsh> ls /# lists b set-enc enc-string [id] Changes character encoding of a given document. If no document id is given, the command applies to the current document. This has two effects: changing the XMLDecl encoding declaration in the document prolog to display the new encoding and making all future save operations on the document default to the given charset. xsh> ls <?xml version="1.0" encoding="iso-8859-1"?> <foo>...</foo> xsh> set-enc "utf-8" xsh> ls <?xml version="1.0" encoding="utf-8"?> <foo>...</foo> xsh> save# saves the file in UTF-8 encoding enc, doc-infoset-standalone expression [id] Changes the value of standalone declaration in the XMLDecl prolog of a document. The expression should evaluate to either 1 or 0 or 'yes' or 'no'. The result of applying the command on other values is not specified. If no document id is given, the command applies to the current document. doc-infoskip-dtd expression skip_dtd If the value of expression is 1 (non-zero), DTD DOCTYPE declaration is omitted from any serialization of XML documents (including ls and save). Default value is 0. This command is equivalent to setting the $SKIP_DTD variable. sort xpath|perl-code perl-code %id EXPERIMENTAL! This command is not yet guaranteed to remain in the future releases. DOCUMENTATION OBSOLETE! Syntax changed! This command may be used to sort the node-list stored in the node-list variable id. First, for each node in the node-list %id, the first argument (either a xpath or perl-code expression), which serves as a sorting criterion, is evaluated in the context of the node and the obtained value is stored together with the node. (In case of xpath the result of whatever type is cast to a string). Then perl's sorting algorithm is used to sort the nodelist, consulting the second, perl-code, argument to compare nodes. Before the perl-code is evaluated, the values obtained from the previous evaluation of the sorting crierion argument on the two nodes being compared are stored into $a and $b variables in the respective order. The perl-code being consulted is supposed to return either -1 (the first node should come first), 0 (no order precedence), or 1 (the second node should come first). Note that Perl provides very convenient operators cmp and <