man SVN::Core () - Core module of the subversion perl bindings


SVN::Core - Core module of the subversion perl bindings


    use SVN::Core; # does apr_initialize and cleanup for you

    # create a root pool and set it as default pool for later use
    my $pool = SVN::Pool->new_default;

    sub something {
        # create a subpool of the current default pool
        my $pool = SVN::Pool->new_default_sub;
        # some svn operations...

        # $pool gets destroyed and the previous default pool
        # is restored when $pool's lexical scope ends

    # svn_stream_t as native perl io handle
    my $stream = $txn->root->apply_text('trunk/filea', undef);
    print $stream $text;
    close $stream;

    # native perl io handle as svn_stream_t
    SVN::Repos::dump_fs($repos, \*STDOUT, \*STDERR,
                        0, $repos->fs->youngest_rev, 0);


SVN::Core implements higher level functions of fundamental subversion functions.


SVN::Core::auth_open([auth provider array]);
Takes a reference to an array of authentication providers and returns an auth_baton. If you use prompt providers you can not use this function, but need to use the auth_open_helper.
SVN::Core::auth_open_helper([auth provider array);
Prompt providers return two values instead of one. The 2nd parameter is a reference to whatever was passed into them as the callback. auth_open_helper splits up these arguments, passing the provider objects into auth_open which gives it an auth_baton and putting the other ones in an array. The first return value of this function is the auth_baton, the second is a reference to an array containing the references to the callbacks. These callback arrays should be stored in the object the auth_baton is attached to.


svn_stream_t - SVN::Stream

You can use native perl io handles (including io globs) as svn_stream_t in subversion functions. Returned svn_stream_t are also translated into perl io handles, so you could access them with regular print, read, etc.

Note that some functions take a stream to read or write, while it does not close it but still hold the reference to the handle. In this case the handle won't be destroyed properly. You should always use correct default pool before calling such functions.

svn_pool_t - SVN::Pool

The perl bindings significantly simplify the usage of pools, while still being manually adjustable.

Functions requiring pool as the last argument (which are, almost all of the subversion functions), the pool is optionally. The default pool is used if it is omitted. If default pool is not set, a new root pool will be created and set as default automatically when the first function requiring a default pool is called.

For callback functions providing pool to your subroutine, you could also use CW$pool->default to make it the default pool in the scope.


new ([$parent])
Create a new pool. The pool is a root pool if CW$parent is not supplied.
new_default ([$parent])
Create a new pool. The pool is a root pool if CW$parent is not supplied. Set the new pool as default pool.
Create a new subpool of the current default pool, and set the resulting pool as new default pool.
Clear the pool.
Destroy the pool. If the pool is the default pool, restore the previous default pool as default. This is normally called automatically when the SVN::Pool object is no longer used and destroyed by the perl garbage collector.

svn_error_t - SVN::Error

By default the perl bindings handle exceptions for you. The default handler automatically croaks with an appropriate error message. This is likely sufficient for simple scripts, but more complex usage may demand handling of errors.

You can override the default exception handler by changing the CW$SVN::Error::handler variable. This variable holds a reference to a perl sub that should be called whenever an error is returned by a svn function. This sub will be passed a svn_error_t object. Its return value is ignored.

If you set the CW$SVN::Error::handler to undef then each call will return an svn_error_t object as its first return in the case of an error, followed by the normal return values. If there is no error then a svn_error_t will not be returned and only the normal return values will be returned. When using this mode you should be careful only to call functions in array context. For example: my ($ci) = CW$ctx->mkdir('http://svn/foo'); In this case CW$ci will be an svn_error_t object if an error occurs and a svn_client_commit_info object otherwise. If you leave the parenthesis off around CW$ci (scalar context) it will be the commit_info object, which in the case of an error will be undef.

If you plan on using this exception handling, understanding the exception handling system the C API uses is helpful. You can find information on it in the HACKING file and the API documentation. Looking at the implementation of SVN::Error::croak_on_error and SVN::Error::expanded_message may be helpful as well.

APR error value, possibly SVN_ custom error.
Details from producer of error.
svn_error_t object of the error that's wrapped.
The pool holding this error and any child errors it wraps.
Source file where the error originated.
Source line where the error originated.
Returns the english description of the status code.
Returns the english description of the apr_err status code set on the CW$svn_error_t. This is short for: SVN::Error::strerror($svn_error_t->apr_err()); Returns a new svn_error_t object with the error status specified in CW$apr_err, the child as CW$child, and error message of CW$message. A quick n' easy way to create a wrappered exception with your own message before throwing it up the stack. $child is the svn_error_t object you want to wrap and CW$new_msg is the new error string you want to set. Add new_err to the end of CW$chain's chain of errors. The CW$new_err chain will be copied into CW$chain's pool and destroyed, so CW$new_err itself becomes invalid after this function. Free the memory used by CW$svn_error_t, as well as all ancestors and descendants of CW$svn_error_t. You must call this on every svn_error_t object you get or you will leak memory. Returns the error message by tracing through the svn_error_t object and its children and concatenating the error messages. This is how the internal exception handlers get their error messages.
Returns true if the value is an svn_error type return. Returns false if the value is anything else or undefined. This is useful for seeing if a call has returned an error.
Default error handler. It takes an svn_error_t and extracts the error messages from it and croaks with those messages. It can be used two ways. The first is detailed above as setting it as the automatic exception handler via setting CW$SVN::Error::handler. The 2nd is if you have CW$SVN::Error::handler set to undef as a wrapper for calls you want to croak on when there is an error but don't want to have to write an explicit error handler for example: my CW$result_rev=SVN::Error::croak_on_error($ctx->checkout($url,$path,'HEAD',1)); If there is no error then croak_on_error will return the arguments passed to it unchanged.
The same as croak_on_error except it will give a more detailed stack backtrace. Including showing internal calls within the implementations of the perl bindings. This is useful if you're working on developing the bindings.
This is useful for wrapping around calls which you wish to ignore any potential error. It checks to see if the first parameter is an error and if it is it clears it. It then returns all the other parameters.


'A'dd, 'D'elete, 'R'eplace, 'M'odify
Source path of copy (if any).
Source revision of copy (if any).

svn_node_kind_t - SVN::Node

An enum of the following constants:

$SVN::Node::none, CW$SVN::Node::file, CW$SVN::Node::dir, CW$SVN::Node::unknown.



Opaque object describing a set of configuration options.


Node kind. One of these constants: CW$SVN::Node::none, CW$SVN::Node::file, CW$SVN::Node::dir, CW$SVN::Node::unknown.
Length of file text, or 0 for directories.
Does the node have props?
Last rev in which this node changed.
Time of created_rev (mod-time).
Author of created rev.


Indicates if the credentials may be saved (to disk).


Indicates if the credentials may be saved (to disk).


Indicates if the credentials may be saved (to disk).
Bit mask of the accepted failures.


Primary CN.
ASCII fingerprint.
ASCII date from which the certificate is valid.
ASCII date until which the certificate is valid.
DN of the certificate issuer.
Base-64 encoded DER certificate representation.


Full paths to the certificate file.
Indicates if the credentials may be saved (to disk).


Certificate password.
Indicates if the credentials may be saved (to disk).



Certificate is not yet valid.
Certificate has expired.
Certificate's CN (hostname) does not match the remote hostname.
Certificate authority is unknown (i.e. not trusted).
Other failure. This can happen if neon has introduced a new failure bit that we do not handle yet.


Chia-liang Kao <>


Copyright (c) 2003 CollabNet. All rights reserved.

This software is licensed as described in the file COPYING, which you should have received as part of this distribution. The terms are also available at If newer versions of this license are posted there, you may use a newer version instead, at your option.

This software consists of voluntary contributions made by many individuals. For exact contribution history, see the revision history and logs, available at