Created attachment 23471 [details] The ppt that contains a shape (throws NullPointerException) This problem happens when having extracting text from a .ppt file. Appears that the problems resides when extracting text from shapes, since I've made a simple test with a .ppt with a simple shape with text (test1) and another test with .ppt without the shape but with the same text (test2). On test1 we get a NullPointerException (org.apache.poi.hslf.model.SimpleShape.getClientRecords(SimpleShape.java:322)) and on test2 we get the desired text. Another thing is that when saving the test1 presentation has a .pptx file and then running a third test using that file, we get the desired result.
Created attachment 23472 [details] The ppt that doesn't contains a shape (all OK)
Created attachment 23473 [details] The pptx that contains a shape and runs OK
Created attachment 23474 [details] Test case Test case: pointing out the .ppt that throws NullPointerException Referenced libraries are: poi-3.5-beta5-20090219.jar poi-contrib-3.5-beta5-20090219.jar poi-ooxml-3.5-beta5-20090219.jar poi-scratchpad-3.5-beta5-20090219.jar log4j.jar dom4j.jar ooxml-schemas-1.0.jar
The problem is not reproducible with current trunk. The test case successfully runs against all the attached files. Yegor