下面总结一下媒体扫描的工作流程，它并不复杂，就是有些绕，如图10-2所示： :-:  图10-2 MediaScanner扫描流程图 通过上图可以发现，MS扫描的流程还是比较清晰的，就是四渡赤水这一招，让很多初学者摸不着头脑。不过读者千万不要像我当初那样，觉得这是垃圾代码的代表。实际上这是码农有意而为之，在MediaScanner.java中通过一段比较详细的注释，对整个流程做了文字总结，这段总结非常简单，这里就不翻译了。 **MediaScanner.java** ~~~ //前面还有一段话，读者可自行阅读。下面是流程的文件总结。 * In summary: * JavaMediaScannerService calls * JavaMediaScanner scanDirectories, which calls * JavaMediaScanner processDirectory (native method), which calls * nativeMediaScanner processDirectory, which calls * nativeMyMediaScannerClient scanFile, which calls * JavaMyMediaScannerClient scanFile, which calls * JavaMediaScannerClient doScanFile, which calls * JavaMediaScanner processFile (native method), which calls * nativeMediaScanner processFile, which calls * nativeparseMP3, parseMP4, parseMidi, parseOgg or parseWMA, which calls * nativeMyMediaScanner handleStringTag, which calls * JavaMyMediaScanner handleStringTag. * OnceMediaScanner processFile returns, an entry is inserted in to the database. ~~~ 看完这么详细的注释，想必你也会认为，码农真是故意这么做的。但他们为什么要设计成这样呢？以后会不会改呢？注释中也说明了目前设计的流程是这样，估计以后有可能改。