Created attachment 21923 [details] Word Escher image extraction This patch adds support for extracting Escher drawing information from the body of a word document. It also adds support for the FSPA plex to determine what shapes correspond to characters in the document, although no code has been made to do so. The patch also extends PicturesTable.getAllPictures() to extract images from Escher drawings. Note that only pictures from the main body of the document (not headers, etc.) are extracted from Escher drawings with this patch. There is also a minor fix to EscherClientAnchorRecord. The word format of this record is different from excel, etc., but I haven't found the exact format. This patch fixes the record from throwing an exception when read. The patch includes a small test case.
Could you please upload GaiaTest.doc and GaiaTestImg.png? (svn patches do not include binary files) The rest of the patch has been applied to svn trunk, thanks for that
There were some compiler errors in FSPATable. It looks like these were some tiny mistakes being hidden by (1.5) auto-boxing. Fixes applied in r658984. I also made the class a little better encapsulated and type safe. It doesn't seem to be exercised by any test cases yet.
Created attachment 22135 [details] Document for test case
Created attachment 22136 [details] Image for test case