The London Perl and Raku Workshop takes place on 26th Oct 2024. If your company depends on Perl, please consider sponsoring and/or attending.

NAME

HTTP::DAV - A WebDAV client library for Perl5

SYNOPSIS

   # DAV script that connects to a webserver and safely
   # uploads the homepage.
   use HTTP::DAV;
  
   $d = new HTTP::DAV;
   $url = "http://host.org:8080/dav/";
  
   $d->credentials( -user=>"pcollins",-pass =>"mypass", 
                    -url =>$url,      -realm=>"DAV Realm" );
  
   $d->open( -url=>"$url )
      or die("Couldn't open $url: " .$d->message . "\n");
  
   $d->lock( -url=>"$url/index.html", -timeout=>"10m" ) 
      or die "Won't put unless I can lock\n";
  
   if ( $d->put( -local=>"/tmp/index.html", -url=>$url ) ) {
      print "/tmp/index.html successfully uploaded to $url\n";
   } else {
      print "put failed: " . $d->message . "\n";
   }
  
   $d->unlock( -url=>"$url/index.html");

DESCRIPTION

HTTP::DAV is a Perl API for interacting with and modifying content on webservers using the WebDAV protocol. Now you can LOCK, DELETE and PUT files and much more on a DAV-enabled webserver.

HTTP::DAV is part of the PerlDAV project hosted at http://www.webdav.org/perldav/ and has the following features:

  • Full RFC2518 method support. OPTIONS, TRACE, GET, HEAD, DELETE, PUT, COPY, MOVE, PROPFIND, PROPPATCH, LOCK, UNLOCK.

  • A fully object-oriented API.

  • Recursive GET and PUT for site backups and other scripted transfers.

  • Transparent lock handling when performing LOCK/COPY/UNLOCK sequences.

  • dave, a fully-functional ftp-style interface written on top of the HTTP::DAV API and bundled by default with the HTTP::DAV library. (If you've already installed HTTP::DAV, then dave will also have been installed (probably into /usr/local/bin). You can see it's man page by typing "perldoc dave" or going to http://www.webdav.org/perldav/dave/.

  • It is built on top of the popular LWP (Library for WWW access in Perl). This means that HTTP::DAV inherits proxy support, redirect handling, basic (and digest) authorization and many other HTTP operations. https is yet to be tested but is believed to be available with LWP. See LWP for more information.

  • Popular server support. HTTP::DAV has been tested against the following servers: mod_dav, IIS5, Xythos webfile server and mydocsonline. The library is growing an impressive interoperability suite which also serves as useful "sample scripts". See "make test" and t/*.

HTTP::DAV essentially has two API's, one which is accessed through this module directly (HTTP::DAV) and is a simple abstraction to the rest of the HTTP::DAV::* Classes. The other interface consists of the HTTP::DAV::* classes which if required allow you to get "down and dirty" with your DAV and HTTP interactions.

The methods provided in HTTP::DAV should do most of what you want. If, however, you need more control over the client's operations or need more info about the server's responses then you will need to understand the rest of the HTTP::DAV::* interfaces. A good place to start is with the HTTP::DAV::Resource and HTTP::DAV::Response documentation.

METHODS

METHOD CALLING: Named vs Unnamed parameters

You can pass parameters to HTTP::DAV methods in one of two ways: named or unnamed.

Named parameters provides for a simpler/easier to use interface. A named interface affords more readability and allows the developer to ignore a specific order on the parameters. (named parameters are also case insensitive)

Each argument name is preceded by a dash. Neither case nor order matters in the argument list. -url, -Url, and -URL are all acceptable. In fact, only the first argument needs to begin with a dash. If a dash is present in the first argument, HTTP::DAV assumes dashes for the subsequent ones.

Each method can also be called with unnamed parameters which often makes sense for methods with only one parameter. But the developer will need to ensure that the parameters are passed in the correct order (as listed in the docs).

 Doc:     method( -url=>$url, [-depth=>$depth] )
 Named:   $d->method( -url=>$url, -depth=>$d ); # VALID
 Named:   $d->method( -Depth=>$d, -Url=>$url ); # VALID
 Named:   $d->method( Depth=>$d,  Url=>$url );  # INVALID (needs -)
 Named:   $d->method( -Arg2=>$val2 ); # INVALID, ARG1 is not optional
 Unnamed: $d->method( $val1 );        # VALID
 Unnamed: $d->method( $val2,$val1 );  # INVALID, ARG1 must come first.

PUBLIC METHODS

new(USERAGENT)

Creates a new HTTP::DAV client

 $d = HTTP::DAV->new()

The -useragent parameter expects an HTTP::DAV::UserAgent object. See the dave program for an advanced example of a custom UserAgent that interactively prompts the user for their username and password.

credentials(USER,PASS,[URL],[REALM])

sets authorization credentials for a URL and/or REALM.

When the client hits a protected resource it will check these credentials to see if either the URL or REALM match the authorization response.

Either URL or REALM must be provided.

returns no value

Example:

 $d->credentials( -url=>'myhost.org:8080/test/',
                  -user=>'pcollins',
                  -pass=>'mypass');
DebugLevel($val)

sets the debug level to $val. 0=off 3=noisy.

$val default is 0.

returns no value.

When the value is greater than 1, the HTTP::DAV::Comms module will log all of the client<=>server interactions into /tmp/perldav_debug.txt.

DAV OPERATIONS

For all of the following operations, URL can be absolute (http://host.org/dav/) or relative (../dir2/). The only operation that requires an absolute URL is open.

copy(URL,DEST,[OVERWRITE],[DEPTH])

copies one remote resource to another

-url

is the remote resource you'd like to copy. Mandatory

-dest

is the remote target for the copy command. Mandatory

-overwrite

optionally indicates whether the server should fail if the target exists. Valid values are "T" and "F" (1 and 0 are synonymous). Default is T.

-depth

optionally indicates whether the server should do a recursive copy or not. Valid values are 0 and (1 or "infinity"). Default is "infinity" (1).

The return value is always 1 or 0 indicating success or failure.

Requires a working resource to be set before being called. See open.

Note: if either 'URL' or 'DEST' are locked by this dav client, then the lock headers will be taken care of automatically. If the either of the two URL's are locked by someone else, the server should reject the request.

copy examples:

  $d->open(-url=>"host.org/dav_dir/");

Recursively copy dir1/ to dir2/

  $d->copy(-url=>"dir1/", -dest=>"dir2/");

Non-recursively and non-forcefully copy dir1/ to dir2/

  $d->copy(-url=>"dir1/", -dest=>"dir2/",-overwrite=>0,-depth=>0);

Create a copy of dir1/file.txt as dir2/file.txt

  $d->cwd(-url=>"dir1/");
  $d->copy("file.txt","../dir2");

Create a copy of file.txt as dir2/new_file.txt

  $d->copy("file.txt","/dav_dir/dir2/new_file.txt")
cwd(URL)

changes the remote working directory.

This is synonymous to open except that the URL can be relative.

  $d->open("host.org/dav_dir/dir1/");
  $d->cwd("../dir2");
  $d->cwd(-url=>"../dir1");

The return value is always 1 or 0 indicating success or failure.

Requires a working resource to be set before being called. See open.

You can not cwd to files, only collections (directories).

delete(URL)

deletes a remote resource

  $d->open("host.org/dav_dir/");
  $d->delete("index.html");
  $d->delete("./dir1");
  $d->delete(-url=>"/dav_dir/dir2/");

This command will recursively delete directories. BE CAREFUL of uninitialised file variables in situation like this: $d->delete("$dir/$file"). This will trash your $dir if $file is not set.

The return value is always 1 or 0 indicating success or failure.

Requires a working resource to be set before being called. See open.

get(URL,[TO],[CALLBACK])

downloads the file or directory at URL to the local location indicated by TO.

-url

is the remote resource you'd like to get. It can be a file or directory or a "glob".

-dest

is where you'd like to put the remote resource. If -dest is not specified then the get() function will routine the file contents as a scalar.

-callback

is a reference to a callback function which will be called everytime a file is completed downloading. The idea of the callback function is that some recursive get's can take a very long time and the user may require some visual feedback. Your callback function must accept 2 arguments as follows:

   # Print a message everytime a file get completes
   sub mycallback {
      my($success,$mesg) = @_;
      if ($success) {
         print "$mesg\n" 
      } else {
         print "Failed: $mesg\n" 
      }
   }

   $d->get( -url=>$url, -to=>$to, -callback=>\&mycallback );

The success argument specifies whether the get operation succeeded or not.

The mesg argument is a status message. The status message could contain any string and often contains useful error messages or success messages. The error messages set during a recursive get are also retrievable via the errors() function discussed further down under ERROR HANDLING

The return value of get is normally 1 or 0 indicating whether the entire get sequence was a success or if there was ANY failures. For instance, in a recursive get, if the server couldn't open 1 of the 10 remote files, for whatever reason, then the return value will be 0. This is so that you can have your script call the errors() routine to handle error conditions.

If you did not specify a -to location then get will return the file's contents as a scalar. In this case, you should expect "undef", or "file contents".

Requires a working resource to be set before being called. See open.

get examples:

  $d->open("host.org/dav_dir/");

Recursively get remote my_dir/ to .

  $d->get("my_dir/",".");

Recursively get remote my_dir/ to /tmp/my_dir/ calling &mycallback($success,$mesg) everytime a file operation is completed.

  $d->get("my_dir","/tmp",\&mycallback);

Get remote my_dir/index.html to /tmp/index.html

  $d->get(-url=>"/dav_dir/my_dir/index.html",-to=>"/tmp");

Get remote index.html to /tmp/index1.html

  $d->get("index.html","/tmp/index1.html");

Get remote index.html as a scalar (into the string $file_contents):

  $file_contents = $d->get("index.html");

Get all of the files matching the globs file1* and file2*:

  $d->get("file[12]*","/tmp");

Get all of the files matching the glob file?.html:

  $d->get("file?.html","/tmp"); # downloads file1.html and file2.html but not file3.html or file1.txt

Invalid glob:

  $d->get("/dav_dir/*/index.html","/tmp"); # Can not glob like this.
lock([URL],[OWNER],[DEPTH],[TIMEOUT],[SCOPE],[TYPE])

locks a resource. If URL is not specified, it will lock the current working resource (opened resource).

   $d->lock( -url     => "index.html",
             -owner   => "Patrick Collins",
             -depth   => "infinity",
             -scope   => "exclusive",
             -type    => "write",
             -timeout => "10h" )

See HTTP::DAV::Resource lock() for details of the above parameters.

The return value is always 1 or 0 indicating success or failure.

Requires a working resource to be set before being called. See open.

When you lock a resource, the lock is held against the current HTTP::DAV object. In fact, the locks are held in a HTTP::DAV::ResourceList object. You can operate against all of the locks that you have created as follows:

  ## Print and unlock all locks that we own.
  my $rl_obj = $d->get_lockedresourcelist();
  foreach $resource ( $rl_obj->get_resources() ) {
      @locks = $resource->get_locks(-owned=>1);
      foreach $lock ( @locks ) { 
        print $resource->get_uri . "\n";
        print $lock->as_string . "\n";
      }
      ## Unlock them?
      $resource->unlock;
  }

Typically, a simple $d->unlock($uri) will suffice.

lock example

  $d->lock($uri, -timeout=>"1d");
  ...
  $d->put("/tmp/index.html",$uri);
  $d->unlock($uri);
mkcol(URL)

make a remote collection (directory)

The return value is always 1 or 0 indicating success or failure.

Requires a working resource to be set before being called. See open.

  $d->open("host.org/dav_dir/");
  $d->mkcol("new_dir");                  # Should succeed
  $d->mkcol("/dav_dir/new_dir");         # Should succeed
  $d->mkcol("/dav_dir/new_dir/xxx/yyy"); # Should fail
move(URL,DEST,[OVERWRITE],[DEPTH])

moves one remote resource to another

-url

is the remote resource you'd like to move. Mandatory

-dest

is the remote target for the move command. Mandatory

-overwrite

optionally indicates whether the server should fail if the target exists. Valid values are "T" and "F" (1 and 0 are synonymous). Default is T.

Requires a working resource to be set before being called. See open.

The return value is always 1 or 0 indicating success or failure.

Note: if either 'URL' or 'DEST' are locked by this dav client, then the lock headers will be taken care of automatically. If either of the two URL's are locked by someone else, the server should reject the request.

move examples:

  $d->open(-url=>"host.org/dav_dir/");

move dir1/ to dir2/

  $d->move(-url=>"dir1/", -dest=>"dir2/");

non-forcefully move dir1/ to dir2/

  $d->move(-url=>"dir1/", -dest=>"dir2/",-overwrite=>0);

Move dir1/file.txt to dir2/file.txt

  $d->cwd(-url=>"dir1/");
  $d->move("file.txt","../dir2");

move file.txt to dir2/new_file.txt

  $d->move("file.txt","/dav_dir/dir2/new_file.txt")
open(URL)

opens the directory (collection resource) at URL.

open will perform a propfind against URL. If the server does not understand the request then the open will fail.

Similarly, if the server indicates that the resource at URL is NOT a collection, the open command will fail.

options([URL])

Performs an OPTIONS request against the URL or the working resource if URL is not supplied.

Requires a working resource to be set before being called. See open.

The return value is a string of comma separated OPTIONS that the server states are legal for URL or undef otherwise.

A fully compliant DAV server may offer as many methods as: OPTIONS, TRACE, GET, HEAD, DELETE, PUT, COPY, MOVE, PROPFIND, PROPPATCH, LOCK, UNLOCK

Note: IIS5 does not support PROPPATCH or LOCK on collections.

Example:

 $options = $d->options($url);
 print $options . "\n";
 if ($options=~ /\bPROPPATCH\b/) {
    print "OK to proppatch\n";
 }

Or, put more simply:

 if ( $d->options($url) =~ /\bPROPPATCH\b/ ) {
    print "OK to proppatch\n";
 }
propfind([URL],[DEPTH])

Perform a propfind against URL at DEPTH depth.

-depth can be used to specify how deep the propfind goes. "0" is collection only. "1" is collection and it's immediate members (This is the default value). "infinity" is the entire directory tree. Note that most DAV compliant servers deny "infinity" depth propfinds for security reasons.

Requires a working resource to be set before being called. See open.

The return value is an HTTP::DAV::Resource object on success or 0 on failure.

The Resource object can be used for interrogating properties or performing other operations.

 ## Print collection or content length
 if ( $r=$d->propfind( -url=>"/my_dir", -depth=>1) ) {
    if ( $r->is_collection ) {
       print "Collection\n" 
       print $r->get_resourcelist->as_string . "\n"
    } else {
       print $r->get_property("getcontentlength") ."\n";
    }
 }

Please note that although you may set a different namespace for a property of a resource during a set_prop, HTTP::DAV currently ignores all XML namespaces so you will get clashes if two properties have the same name but in different namespaces. Currently this is unavoidable but I'm working on the solution.

proppatch([URL],[NAMESPACE],PROPNAME,PROPVALUE,ACTION)

If -action equals "set" then we set a property named -propname to -propvalue in the namespace -namespace for -url.

If -action equals "remove" then we unset a property named -propname in the namespace -namespace for -url.

If no action is supplied then the default action is "set".

The return value is an HTTP::DAV::Resource object on success or 0 on failure.

The Resource object can be used for interrogating properties or performing other operations.

Requires a working resource to be set before being called. See open.

It is recommended that you use set_prop and unset_prop instead of proppatch for readability.

set_prop simply calls proppatch(-action=set)> and unset_prop calls proppatch(-action="remove")>

See set_prop and unset_prop for examples.

put(LOCAL,[URL],[CALLBACK])

uploads the files or directories at -local to the remote destination at -url.

-local points to a file, directory or series of files or directories (indicated by a glob).

If the filename contains any of the characters `*', `?' or `[' it is a candidate for filename substitution, also known as ``globbing''. This word is then regarded as a pattern (``glob-pattern''), and replaced with an alphabetically sorted list of file names which match the pattern.

One can upload/put a string by passing a reference to a scalar in the -local parameter. See example below.

put requires a working resource to be set before being called. See open.

The return value is always 1 or 0 indicating success or failure.

See get() for a description of what the optional callback parameter does.

put examples:

Put a string to the server:

  my $myfile = "This is the contents of a file to be uploaded\n";
  $d->put(-local=>\$myfile,-url=>"http://www.host.org/dav_dir/file.txt");

Put a local file to the server:

  $d->put(-local=>"/tmp/index.html,-url=>"http://www.host.org/dav_dir/");

Put a series of local files to the server:

  In these examples, /tmp contains file1.html, file1, file2.html, 
  file2.txt, file3.html, file2/

  $d->put(-local=>"/tmp/file[12]*",-url=>"http://www.host.org/dav_dir/");
  
  uploads file1.html, file1, file2.html, file2.txt and the directory file2/ to dav_dir/.
set_prop([URL],[NAMESPACE],PROPNAME,PROPVALUE)

Sets a property named -propname to -propvalue in the namespace -namespace for -url.

Requires a working resource to be set before being called. See open.

The return value is an HTTP::DAV::Resource object on success or 0 on failure.

The Resource object can be used for interrogating properties or performing other operations.

Example:

 if ( $r = $d->set_prop(-url=>$url,
              -namespace=>"dave",
              -propname=>"author",
              -propvalue=>"Patrick Collins"
             ) ) {
    print "Author property set\n";
 } else {
    print "set_prop failed:" . $d->message . "\n";
 }

See the note in propfind about namespace support in HTTP::DAV. They're settable, but not readable.

steal([URL])

forcefully steals any locks held against URL.

steal will perform a propfind against URL and then, any locks that are found will be unlocked one by one regardless of whether we own them or not.

Requires a working resource to be set before being called. See open.

The return value is always 1 or 0 indicating success or failure. If multiple locks are found and unlocking one of them fails then the operation will be aborted.

 if ($d->steal()) {
    print "Steal succeeded\n";
 } else {
    print "Steal failed: ". $d->message() . "\n";
 }
unlock([URL])

unlocks any of our locks on URL.

Requires a working resource to be set before being called. See open.

The return value is always 1 or 0 indicating success or failure.

 if ($d->unlock()) {
    print "Unlock succeeded\n";
 } else {
    print "Unlock failed: ". $d->message() . "\n";
 }
unset_prop([URL],[NAMESPACE],PROPNAME)

Unsets a property named -propname in the namespace -namespace for -url. Requires a working resource to be set before being called. See open.

The return value is an HTTP::DAV::Resource object on success or 0 on failure.

The Resource object can be used for interrogating properties or performing other operations.

Example:

 if ( $r = $d->unset_prop(-url=>$url,
              -namespace=>"dave",
              -propname=>"author",
             ) ) {
    print "Author property was unset\n";
 } else {
    print "set_prop failed:" . $d->message . "\n";
 }

See the note in propfind about namespace support in HTTP::DAV. They're settable, but not readable.

ACCESSOR METHODS

get_user_agent

Returns the clients' working HTTP::DAV::UserAgent object.

You may want to interact with the HTTP::DAV::UserAgent object to modify request headers or provide advanced authentication procedures. See dave for an advanced authentication procedure.

get_last_request

Takes no arguments and returns the clients' last outgoing HTTP::Request object.

You would only use this to inspect a request that has already occurred.

If you would like to modify the HTTP::Request BEFORE the HTTP request takes place (for instance to add another header), you will need to get the HTTP::DAV::UserAgent using get_user_agent and interact with that.

get_workingresource

Returns the currently "opened" or "working" resource (HTTP::DAV::Resource).

The working resource is changed whenever you open a url or use the cwd command.

e.g. $r = $d->get_workingresource print "pwd: " . $r->get_uri . "\n";

get_workingurl

Returns the currently "opened" or "working" URL.

The working resource is changed whenever you open a url or use the cwd command.

  print "pwd: " . $d->get_workingurl . "\n";
get_lockedresourcelist

Returns an HTTP::DAV::ResourceList object that represents all of the locks we've created using THIS dav client.

  print "pwd: " . $d->get_workingurl . "\n";
get_absolute_uri(REL_URI,[BASE_URI])

This is a useful utility function which joins BASE_URI and REL_URI and returns a new URI.

If BASE_URI is not supplied then the current working resource (as indicated by get_workingurl) is used. If BASE_URI is not set and there is no current working resource the REL_URI will be returned.

For instance: $d->open("http://host.org/webdav/dir1/");

 # Returns "http://host.org/webdav/dir2/"
 $d->get_absolute_uri(-rel_uri=>"../dir2");

 # Returns "http://x.org/dav/dir2/file.txt"
 $d->get_absolute_uri(-rel_uri  =>"dir2/file.txt",
                      ->base_uri=>"http://x.org/dav/");

Note that it subtly takes care of trailing slashes.

ERROR HANDLING METHODS

message

message gets the last success or error message.

The return value is always a scalar (string) and will change everytime a dav operation is invoked (lock, cwd, put, etc).

See also errors for operations which contain multiple error messages.

errors

Returns an @array of error messages that had been set during a multi-request operation.

Some of HTTP::DAV's operations perform multiple request to the server. At the time of writing only put and get are considered multi-request since they can operate recursively requiring many HTTP requests.

In these situations you should check the errors array if to determine if any of the requests failed.

The errors function is used for multi-request operations and not to be confused with a multi-status server response. A multi-status server response is when the server responds with multiple error messages for a SINGLE request. To deal with multi-status responses, see HTTP::DAV::Response.

 # Recursive put
 if (!$d->put( "/tmp/my_dir", $url ) ) {
    # Get the overall message
    print $d->message;
    # Get the individual messages
    foreach $err ( $d->errors ) { print "  Error:$err\n" }
 }
is_success

Returns the status of the last DAV operation performed through the HTTP::DAV interface.

This value will always be the same as the value returned from an HTTP::DAV::method. For instance:

  # This will always evaluate to true
  ($d->lock($url) eq $d->is_success) ?

You may want to use the is_success method if you didn't capture the return value immediately. But in most circumstances you're better off just evaluating as follows: if($d->lock($url)) { ... }

get_lastresponse

Takes no arguments and returns the last seen HTTP::DAV::Response object.

You may want to use this if you have just called a propfind and need the individual error messages returned in a MultiStatus.

If you find that you're using get_last_response() method a lot, you may be better off using the more advanced HTTP::DAV interface and interacting with the HTTP::DAV::* interfaces directly as discussed in the intro. For instance, if you find that you're always wanting a detailed understanding of the server's response headers or messages, then you're probably better off using the HTTP::DAV::Resource methods and interpreting the HTTP::DAV::Response directly.

To perform detailed analysis of the server's response (if for instance you got back a multistatus response) you can call get_lastresponse which will return to you the most recent response object (always the result of the last operation, PUT, PROPFIND, etc). With the returned HTTP::DAV::Response object you can handle multi-status responses.

For example:

   # Print all of the messages in a multistatus response
   if (! $d->unlock($url) ) {
      $response = $d->get_lastresponse();
      if ($response->is_multistatus() ) {
        foreach $num ( 0 .. $response->response_count() ) {
           ($err_code,$mesg,$url,$desc) =
              $response->response_bynum($num);
           print "$mesg ($err_code) for $url\n";
        }
      }
   }

ADVANCED METHODS

new_resource

Creates a new resource object with which to play. This is the preferred way of creating an HTTP::DAV::Resource object if required. Why? Because each Resource object needs to sit within a global HTTP::DAV client. Also, because the new_resource routine checks the HTTP::DAV locked resource list before creating a new object.

 $dav->new_resource( -uri => "http://..." );
set_workingresource(URL)

Sets the current working resource to URL.

You shouldn't need this method. Call open or cwd to set the working resource.

You CAN call set_workingresource but you will need to perform a propfind immediately following it to ensure that the working resource is valid.

INSTALLATION, TODO, MAILING LISTS and REVISION HISTORY

Please see the primary HTTP::DAV webpage at (http://www.webdav.org/perldav/http-dav/) or the README file in this library.

SEE ALSO

You'll want to also read: HTTP::DAV::Response, HTTP::DAV::Resource, dave

and maybe if you're more inquisitive: LWP::UserAgent,HTTP::Request, HTTP::DAV::Comms,HTTP::DAV::Lock, HTTP::DAV::ResourceList, HTTP::DAV::Utils

AUTHOR AND COPYRIGHT

This module is Copyright (C) 2001 by

    Patrick Collins
    G03 Gloucester Place, Kensington
    Sydney, Australia

    Email: pcollins@cpan.org
    Phone: +61 2 9663 4916

All rights reserved.

You may distribute this module under the terms of either the GNU General Public License or the Artistic License, as specified in the Perl README file.