DOM guide: Resolving URIs: Difference between revisions
No edit summary |
No edit summary |
||
Line 35: | Line 35: | ||
file:///c:/A/B/C/D/doc.dae | file:///c:/A/B/C/D/doc.dae | ||
here are some examples of how the following URIs will be normalized: | here are some examples of how the following URIs will be normalized: | ||
./path/document.dae | ./path/document.dae => file:///c:/A/B/C/D/path/document.dae | ||
../../../path/document.dae#elementID => file:///A/path/document.dae#elementID | ../../../path/document.dae#elementID => file:///A/path/document.dae#elementID | ||
/c:/path/document.dae#Light01 | /c:/path/document.dae#Light01 => file:///c:/path/document.dae#Light01 | ||
c:/path/document.dae | c:/path/document.dae => file:///c:/A/B/C/D/c:/path/document.dae | ||
Note that '''Windows file paths ARE NOT proper absolute path relative URIs!''' | Note that '''Windows file paths ARE NOT proper absolute path relative URIs!''' | ||
Note also that only the '/' character is used as a path delimeter. Windows uses the '\' character to delimit path segments. Using '\' character can result in incorrect URI processing! "A\B\C" is considered one path segment. If using "file:///A\B\C" as a base URI and trying to resolve "../doc.dae" the result will be "file:///doc.dae" and NOT "file:///A\B\doc.dae" as one might have expected. | Note also that only the '/' character is used as a path delimeter. Windows uses the '\' character to delimit path segments. Using '\' character can result in incorrect URI processing! "A\B\C" is considered one path segment. If using "file:///A\B\C" as a base URI and trying to resolve "../doc.dae" the result will be "file:///doc.dae" and NOT "file:///A\B\doc.dae" as one might have expected. | ||
===Accessing URI Data=== | ===Accessing URI Data=== | ||
The daeURI class provides accessors to all of the parts of a parsed URI. Please note that the COLLADA DOM refers to the "scheme" as "protocol" and the fragment as "id". | |||
Calling daeURI::getURI will return the full URI string. This may be the non-normalized URI string if getURI is called before the URI is normalized. URIs are normalized when they are needed for an operation, i.e. resolveElement or loading a document. | |||
Calling daeURI::getOriginalURI will return the original URI string either read from the document or supplied by the user before the URI was normalized. | |||
daeURI::getContainer will return a reference to the element that contains this URI data. This is helpful if you are given a daeURI object and need to find it's context or check which document it belongs to. | |||
===Resolving URIs=== | ===Resolving URIs=== | ||
As part of the document loading process, the COLLADA DOM will attempt to resolve all of the URIs that are loaded. | |||
The element that a URI resolves to, if successful, can be retrieved by calling daeURI::getElement. | |||
If daeURI::getElement returns NULL then there could have been an error attempting to resolve the URI or there hasn't been an attempt at all. Calling daeURI::getState will tell you the status of the URI. If getState returns a value of uri_failed_* then there was an error resolving the URI. If getState returns uri_pending then there hasn't been an attempt to resolve the URI. | |||
To resolve a URI call daeURI::resolveElement. | |||
All URIs that are found in a document upon load will attempt to resolve, but if an application creates elements programattically you may encounter URIs that are in the uri_pending unresolved state. To be safe, you should always check daeURI::getElement for NULL and if NULL, check the state and if needed attempt to resolve the URI manually by calling daeURI::resolveElement. | |||
If there have been many elements that have been created calling daeElement::resolveAll will resolve all URIs that are present in those newly created elements. An application that creates new elements and intends to pass COLLADA DOM data between different systems, like a conditioner in the [[COLLADA Refinery]], would benefit from calling daeElement::resolveAll before finishing. This could avoid common errors in which applications assume all URIs have attempted to be resolved. | |||
===Creating URI Strings=== | ===Creating URI Strings=== |
Revision as of 22:11, 26 March 2007
The daeURI
class represents URI data types in the COLLADA DOM. The daeURI
class provides an interface for accessing/manipulating the URI string and for resolving the URI reference.
Working with daeURI
The COLLADA DOM uses normalized absolute URIs only. The daeURI class contains code that converts relative URIs into absolute URIs.
The base URI syntax is
scheme://authority/filepath?query#fragment
An absolute URI contains a scheme and an authority. If the authority is left empty than localhost
is used, i.e.
file:///c:/path/document.dae#Geo_01
is a URI that references a file named document.dae on the localhost found in c:/path.
The fragment portion is used to identify elements by using their "id" attribute. The previous example references an element with an id "Geo_01".
A relative URI is any URI that does not contain a scheme and authority.
A relative URI can be a relative path, an absolute path, or just a fragment.
Examples of relative path URIs:
./path/document.dae ../../../path/document.dae#elementID document.dae#Geo_01
An absolute path must be preceded by a forward slash character '/'. An example:
/c:/path/document.dae#Light01
A fragment identifier references an element that can be found within the same document as the URI. An example of this:
#redMaterial
The COLLADA DOM parses URI strings and normalizes them into absolute URIs before it can use the URI for any operation, i.e. load or resolveElement.
A base URI is needed to normalize relative URIs. Absolute URIs do not require a separate base URI. The COLLADA DOM will provide a base URI. If the URI is found within a COLLADA document then the base URI will be the document URI. If the URI is not part of a document then the current working directory is used as the base URI.
Assuming a base URI of
file:///c:/A/B/C/D/doc.dae
here are some examples of how the following URIs will be normalized:
./path/document.dae => file:///c:/A/B/C/D/path/document.dae ../../../path/document.dae#elementID => file:///A/path/document.dae#elementID /c:/path/document.dae#Light01 => file:///c:/path/document.dae#Light01 c:/path/document.dae => file:///c:/A/B/C/D/c:/path/document.dae
Note that Windows file paths ARE NOT proper absolute path relative URIs! Note also that only the '/' character is used as a path delimeter. Windows uses the '\' character to delimit path segments. Using '\' character can result in incorrect URI processing! "A\B\C" is considered one path segment. If using "file:///A\B\C" as a base URI and trying to resolve "../doc.dae" the result will be "file:///doc.dae" and NOT "file:///A\B\doc.dae" as one might have expected.
Accessing URI Data
The daeURI class provides accessors to all of the parts of a parsed URI. Please note that the COLLADA DOM refers to the "scheme" as "protocol" and the fragment as "id".
Calling daeURI::getURI will return the full URI string. This may be the non-normalized URI string if getURI is called before the URI is normalized. URIs are normalized when they are needed for an operation, i.e. resolveElement or loading a document.
Calling daeURI::getOriginalURI will return the original URI string either read from the document or supplied by the user before the URI was normalized.
daeURI::getContainer will return a reference to the element that contains this URI data. This is helpful if you are given a daeURI object and need to find it's context or check which document it belongs to.
Resolving URIs
As part of the document loading process, the COLLADA DOM will attempt to resolve all of the URIs that are loaded.
The element that a URI resolves to, if successful, can be retrieved by calling daeURI::getElement.
If daeURI::getElement returns NULL then there could have been an error attempting to resolve the URI or there hasn't been an attempt at all. Calling daeURI::getState will tell you the status of the URI. If getState returns a value of uri_failed_* then there was an error resolving the URI. If getState returns uri_pending then there hasn't been an attempt to resolve the URI.
To resolve a URI call daeURI::resolveElement.
All URIs that are found in a document upon load will attempt to resolve, but if an application creates elements programattically you may encounter URIs that are in the uri_pending unresolved state. To be safe, you should always check daeURI::getElement for NULL and if NULL, check the state and if needed attempt to resolve the URI manually by calling daeURI::resolveElement.
If there have been many elements that have been created calling daeElement::resolveAll will resolve all URIs that are present in those newly created elements. An application that creates new elements and intends to pass COLLADA DOM data between different systems, like a conditioner in the COLLADA Refinery, would benefit from calling daeElement::resolveAll before finishing. This could avoid common errors in which applications assume all URIs have attempted to be resolved.
Creating URI Strings
The COLLADA DOM provides two ways to create URI strings; by setting the string directly with daeURI::setURI or by using the daeURI::resolveURI method.
Calling daeURI::setURI allows you to pass in a new URI string to be used for the daeURI. The daeURI class will parse this string right away but will not normalize the new URI until it is needed. You may set the URI to any valid URI string, absolute or relative. Most commonly URIs are set to reference other COLLADA elements.
- Passing in a string "#elementID" will create a URI that references the elementID element from within the same document.
- Passing in a string "rel-path/document#elementID" will create a URI that references an element in an external document.
Note. It is the clients responsibility to ensure that strings passed into daeURI::setURI are valid URI strings.
Calling daeURI::resolveURI will create a URI string based on the element set with daeURI::setElement. The COLLADA DOM will create the URI based on the element's documentURI and the element's ID. This method of setting URI strings is very helpful if you have a reference to the element you wish to make the URI reference. Doing so will prevent the need of custom URI string generation. Here is an example of how this works:
//inst_geom is a pointer to an existing domInstance_geometry element //geom is a pointer to an existing domGeometry element. inst_geom->getUrl().setElement( geom ); inst_geom->getUrl().resolveURI();
External URI References
By default, the COLLADA DOM will load additional COLLADA documents to resolve external URI references. I will be refering to these COLLADA documents as "external documents". Client applications may want or need finer control over which documents are loaded and when.
The COLLADA DOM provides this functionality with daeURIResolver::setAutoLoadExternalDocuments and various methods of daeDocument class. To disable the DOM from implicitely loading documents when resolving URIs you must call
daeURIResolver::setAutoLoadExternalDocuments( false );
sometime before you attempt to load a document.
All URIs that reference elements in external documents will not be resolved. You will be returned the value uri_failed_external_document when querying the status of the URI by calling daeURI::getState.
Use daeDocument::getReferencedDocuments to retrieve a list of all documents that the current document has external URI references to.
Use daeDocument::getExternalURIs to retrieve a list of pointers to the daeURI objects that reference a specific document.
When a new document is loaded into the COLLADA DOM, via DAE::load, the COLLADA DOM checks if any currently loaded documents contain any external references to the newly loaded document. If a document does contain an external reference to the newly loaded document the reference is resolved.