在for循环中使用RapidImageLoader加载本地图片，出现内存溢出问题

[jaffwu]

01-12 17:45:17.383 29243-31124 E/dalvikvm-heap: Out of memory on a 3686416-byte allocation.
01-12 17:45:17.383 29243-31124 I/dalvikvm: "pool-1-rapidview_thread_pool-thread-17" prio=1 tid=20 RUNNABLE
01-12 17:45:17.383 29243-31124 I/dalvikvm: | group="main" sCount=0 dsCount=0 obj=0x42ca65b8 self=0x6b8e1fc8
01-12 17:45:17.383 29243-31124 I/dalvikvm: | sysTid=31124 nice=19 sched=0/0 cgrp=apps/bg_non_interactive handle=1792847824
01-12 17:45:17.383 29243-31124 I/dalvikvm: | state=R schedstat=( 0 0 0 ) utm=36 stm=2 core=0
01-12 17:45:17.383 29243-31124 I/dalvikvm: at android.graphics.BitmapFactory.nativeDecodeByteArray(Native Method)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at android.graphics.BitmapFactory.decodeByteArray(BitmapFactory.java:500)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at android.graphics.BitmapFactory.decodeByteArray(BitmapFactory.java:523)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at com.tencent.rapidview.utils.RapidImageLoader.getRapid(RapidImageLoader.java:281)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at com.tencent.rapidview.utils.RapidImageLoader.access$200(RapidImageLoader.java:35)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at com.tencent.rapidview.utils.RapidImageLoader$1.run(RapidImageLoader.java:162)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:422)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at java.util.concurrent.FutureTask.run(FutureTask.java:237)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1112)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:587)
01-12 17:45:17.383 29243-31124 I/dalvikvm: at java.lang.Thread.run(Thread.java:841)
01-12 17:45:17.383 29243-31124 D/skia: --- decoder->decode returned false

[Earlonus]

感谢反馈，本地已经修复，外网验证后再提交上来。

RapidPool的中对文件byte[]流缓存本意是对.lua或.xml等影响加载速度的极小文件进行缓存，提高加载速度而又对内存占用影响微乎其微。

而消耗大量内存资源的图片，建议采用glide等图片缓存库封装成控件展示。

本地修改尝试跳过了对图片的缓存，验证后提交上来。