Next: , Up: Extracting meta data

4.1 Plugin management

Using GNU libextractor from a multi-threaded parent process requires some care. The problem is that on most platforms GNU libextractor starts sub-processes for the actual extraction work. This is useful to isolate the parent process from potential bugs; however, it can cause problems if the parent process is multi-threaded. The issue is that at the time of the fork, another thread of the application may hold a lock (i.e. in gettext or libc). That lock would then never be released in the child process (as the other thread is not present in the child process). As a result, the child process would then deadlock on trying to acquire the lock and never terminate. This has actually been observed with a lock in GNU gettext that is triggered by the plugin startup code when it interacts with libltdl.

The problem can be solved by loading the plugins using the EXTRACTOR_OPTION_IN_PROCESS option, which will run GNU libextractor in-process and thus avoid the locking issue. In this case, all of the functions for loading and unloading plugins, including EXTRACTOR_plugin_add_defaults and EXTRACTOR_plugin_remove_all, are thread-safe and reentrant. However, using the same plugin list from multiple threads at the same time is not safe.

All plugin code is expected required to be reentrant and state-less, but due to the extensive use of 3rd party libraries this cannot be guaranteed.

— C Struct: EXTRACTOR_PluginList

A plugin list represents a set of GNU libextractor plugins. Most of the GNU libextractor API is concerned with either constructing a plugin list or using it to extract meta data. The internal representation of the plugin list is of no concern to users or plugin developers.

— Function: void EXTRACTOR_plugin_remove_all (struct EXTRACTOR_PluginList *plugins)

Unload all of the plugins in the given list.

— Function: struct EXTRACTOR_PluginList * EXTRACTOR_plugin_remove (struct EXTRACTOR_PluginList *plugins, const char*name)

Unloads a particular plugin. The given name should be the short name of the plugin, for example “mime” for the mime-type extractor or “mpeg” for the MPEG extractor.

— Function: struct EXTRACTOR_PluginList * EXTRACTOR_plugin_add (struct EXTRACTOR_PluginList *plugins, const char* name,const char* options, enum EXTRACTOR_Options flags)

Loads a particular plugin. The plugin is added to the existing list, which can be NULL. The second argument specifies the name of the plugin (i.e. “ogg”). The third argument can be NULL and specifies plugin-specific options. Finally, the last argument specifies if the plugin should be executed out-of-process (EXTRACTOR_OPTION_DEFAULT_POLICY) or not.

— Function: struct EXTRACTOR_PluginList * EXTRACTOR_plugin_add_config (struct EXTRACTOR_PluginList *plugins, const char* config, enum EXTRACTOR_Options flags)

Loads and unloads plugins based on a configuration string, modifying the existing list, which can be NULL. The string has the format “[-]NAME(OPTIONS){:[-]NAME(OPTIONS)}*”. Prefixing the plugin name with a “-” means that the plugin should be unloaded.

— Function: struct EXTRACTOR_PluginList * EXTRACTOR_plugin_add_defaults (enum EXTRACTOR_Options flags)

Loads all of the plugins in the plugin directory. This function is what most GNU libextractor applications should use to setup the plugins.