/* ** (c) COPYRIGHT MIT 1995. ** Please first read the full copyright statement in the file COPYRIGH. */
The cache contains details of persistent files which contain the contents of remote documents. The existing cache manager is somewhat naive - especially in its garbage collection but it is just an example of how it can be done.However, it is a fully HTTP/1.1 compliant cache manager. More advanced implementations are welcome!
This module is implemented by HTCache.c, and it is a part of the W3C Sample Code Library.
#ifndef HTCACHE_H
#define HTCACHE_H
#include "WWWLib.h"
#ifdef __cplusplus
extern "C" {
#endif
The cache_root is the URI of the location of the persistent
cache. An example is "file:/tmp/w3c-lib". If
cache_root is NULL then determine a cache root
using the following algorithm:
WWW_CACHE, TMP, and TEMP. If none
are set then then fall back on "/tmp".
w3c-cache" to the root identified above
The cache_root location does not have to exist, it will be created
automatically if not. An empty string will make '/' the cache root.
The size is the total size in MBytes - the default size is 20M. The cache can not be less than 5M.
We can only enable the cache if we are in secure mode where we can not access the local file system. This is for example the case if using an application as a telnet shell.
extern BOOL HTCacheInit (const char * cache_root, int size);
After the cache has been terminated it can not be used anymore unless you
do another HTCacheInit() call.
extern BOOL HTCacheTerminate (void);
The persistent cache has a set of overall parameters that you can adjust
The cache can be temporarily suspended by using the enable/disable flag. This does not prevent the cache from being enabled/disable at a later point in time.
extern void HTCacheMode_setEnabled (BOOL mode); extern BOOL HTCacheMode_enabled (void);
The cache can be setup to whether cache password protected documents thru the protected flag. By default this flag is turned off.
extern void HTCacheMode_setProtected (BOOL mode); extern BOOL HTCacheMode_protected (void);
Return the value of the cache root. The cache root can only be set through
the HTCacheInit() function. The string returned MUST be freed
by the caller
extern char * HTCacheMode_getRoot (void);
We set the default cache size to 20M. We set the minimum size to 5M in order not to get into weird problems while writing the cache. The size is indicated in Mega bytes. The size is given in MBytes and is also returned in MBytes. We don't consider the metainformation as part of the total cache size which is the the reason for why the min cache size should not be less than 5M.
extern BOOL HTCacheMode_setMaxSize (int size); extern int HTCacheMode_maxSize (void);
It is also possible to control the max size of a single cache entry so that the cache doesn't get filled with a very few, very large cached entries. The default max size for a single cached entry is 3M. The value indicated must be in Mbytes, for example, a vaue of 3 would mean 3 MBytes.
extern BOOL HTCacheMode_setMaxCacheEntrySize (int size); extern int HTCacheMode_maxCacheEntrySize (void);
If a response does not arrive with an expiration time and does not explicitly forbid its being cached, use the default expiration time. The time is given in seconds (e.g., 3,600 is one hour).
extern void HTCacheMode_setDefaultExpiration (const int exp_time); extern int HTCacheMode_DefaultExpiration (void);
There are various ways of handling Expires header when met in
a history list. Either it can be ignored all together, the user can
be notified with a warning, or the document can be reloaded automatically.
This flag decides what action to be taken. The default action is
HT_EXPIRES_IGNORE. In HT_EXPIRES_NOTIFY mode ,
we push a message on to the Error stack which is presented to the user.
typedef enum _HTExpiresMode {
HT_EXPIRES_IGNORE = 0,
HT_EXPIRES_NOTIFY,
HT_EXPIRES_AUTO
} HTExpiresMode;
extern void HTCacheMode_setExpires (HTExpiresMode mode);
extern HTExpiresMode HTCacheMode_expires (void);
The cache can be set to handle disconnected operation where it does not use
the network to validate entries and do not attempt to load new documents.
All requests that can not be fulfilled by the cache will be returned with
a "504 Gateway Timeout" response. There are two modes of how
the cache can operate in disconnected mode:
typedef enum _HTDisconnectedMode {
HT_DISCONNECT_NONE = 0,
HT_DISCONNECT_NORMAL = 1,
HT_DISCONNECT_EXTERNAL = 2
} HTDisconnectedMode;
extern void HTCacheMode_setDisconnected (HTDisconnectedMode mode);
extern HTDisconnectedMode HTCacheMode_disconnected (void);
extern BOOL HTCacheMode_isDisconnected (HTReload mode);
The persistent cache keeps an index of its current entries so that garbage collection and lookup becomes more efficient. This index is stored automatically at regular intervals so that we don't get out of sync. Also, it is automatically loaded at startup and saved at closedown of the cache.
Read the saved set of cached entries from disk. we only allow the index ro be read when there is no entries in memory. That way we can ensure consistancy.
extern BOOL HTCacheIndex_read (const char * cache_root);
Walk through the list of cached objects and save them to disk. We override any existing version but that is normally OK as we have already read its contents.
extern BOOL HTCacheIndex_write (const char * cache_root);
The cache object is what we store about a cached objet in memory.
typedef struct _HTCache HTCache;
Filling the cache is done as all other transportation of bulk data in libwww using streams. The cache object creater is a stream which in many cases sits on a T stream so that we get the original feed and at the same time can parse the contents.
In some situations, we want to append data to an already exiting cache entry. This is the case when a use has interrupted a download and we are stuck with a subpart of the document. If the user later on whishes to download the object again we can issue a range request and continue from where we were. This will in many situations save a lot of bandwidth.
extern HTConverter HTCacheWriter, HTCacheAppend;
This function writes the metainformation along with the data object stored by the HTCacheWriter stream above. If no headers are available then the meta file is empty
extern BOOL HTCache_writeMeta (HTCache * cache, HTRequest * request,
HTResponse * response);
In case we received a "304 Not Modified" response then we do
not have to tough the body but must merge the metainformation with the previous
version. Therefore we need a special metainformation update function.
extern BOOL HTCache_updateMeta (HTCache * cache, HTRequest * request,
HTResponse * response);
Clear a cache entry
extern BOOL HTCache_resetMeta (HTCache * cache, HTRequest * request,
HTResponse * response);
After we get a response back, we should check whether we can still cache an entry and/or we should add an entry for a resource that has just been created so that we can remember the etag and other things. The latter allows us to guarantee that we don't loose data due to the lost update problem.
extern HTCache * HTCache_touch (HTRequest * request, HTResponse * response,
HTParentAnchor * anchor);
Loading a cached object is also done as all other loads in libwww by using a protocol load module. For the moment, this load function handles the persistent cache as if it was on local file but in fact it could be anywhere.
extern HTProtCallback HTLoadCache;
Remove a HTCache object from memory and from disk. You must explicitly remove a lock before this operation can succeed
extern BOOL HTCache_remove (HTCache * cache);
Destroys all cache entried in memory but does not write anything to disk. Use the index methods above for doing that. We do not delete the disk contents.
extern BOOL HTCache_deleteAll (void);
Destroys all cache entried in memory and on disk. This call basically resets the cache to the inital state but it does not terminate the cache. That is, you don't have to reinitialize the cache before you can use it again.
extern BOOL HTCache_flushAll (void);
Verifies if a cache object exists for this URL and if so returns a URL for the cached object. It does not verify whether the object is valid or not, for example it might have expired. Use the cache validation methods for checking this.
extern HTCache * HTCache_find (HTParentAnchor * anchor, char * default_name);
This function checks whether a document has expired or not. The check is based on the metainformation passed in the anchor object The function returns the level of validation needed for getting a fresh version. We also check the cache control directives in the request to see if they change the freshness discission.
extern HTReload HTCache_isFresh (HTCache * me, HTRequest * request);
As a cache hit may occur several places, we have a public function where we can declare a download to be a true cache hit. The number of hits a cache object has affects its status when we are doing garbage collection.
extern BOOL HTCache_addHit (HTCache * cache);
Is we have a valid entry in the cache then we also need a location where we can get it. Hopefully, we may be able to access it thourgh one of our protocol modules, for example the local file module. The name returned is in URL syntax and must be freed by the caller
extern char * HTCache_name (HTCache * cache);
While we are creating a new cache object or while we are validating an existing one, we must have a lock on the entry so that not other requests can get to it in the mean while. A lock can be broken if the same request tries to create the cache entry again. This means that we have tried to validate the cache entry but we got a new shipment of bytes back from the origin server or an intermediary proxy.
extern BOOL HTCache_getLock (HTCache * cache, HTRequest * request); extern BOOL HTCache_breakLock (HTCache * cache, HTRequest * request); extern BOOL HTCache_hasLock (HTCache * cache); extern BOOL HTCache_releaseLock (HTCache * cache);
#ifdef __cplusplus } #endif #endif /* HTCACHE_H */