MEPBM: e-mail turns From: dave@casbah.acns.nwu.edu (Dave Wenger) Date: Mon, 05 Jun 1995 00:00:00 +0000 GSI sent me my turn for MPEBM via e-mail. The turn was corrupt, and I could not view the first page or print any of it. I found a workaround, and sent GSI a note which documented my problems. I am posting the workaround: if you have this corruption problem, read on. ----- The 4th line of actual uuencoded text was 2 characters shorter than the other lines, which our uudecoders (mine on the PC and someone else's on the Mac) registered as corrupt. They decoded anyway, and Acrobat Reader was then able to open the file, but for me on a PC in Windows, it registered the error: There was an error processing a page. The document may be damaged [109]. The turn would then come up, except that Page 1 displayed as blank. The rest of the turn displayed successfully. Attempts to print ANY page of the document would cause it to start counting to that page, and it would attempt to count the first page, and display the error: The document could not be printed. The document may be damaged [110]. This is under Windows. Someone else's turn behaved exactly the same on Windows. On the Macintosh, the other file would simply not display at all. We didn't try mine on the Mac. Although trying to decipher the missing uuencoded characters was not successful, I was able to uudecode the document on a unix box, and using a text editor (emacs) which displays all non-text characters as escape codes, were able to compare the actual pdf file to a copy of the test pdf that you sent us previously, that we had imported into Microsoft Word. The corruption was near the top, in something like a header, so I was able to type over the corrupted part in emacs, and then re-uuencode the file, download it to a PC, and open it up cleanly and print it. This should also be possible on the end PC or Mac, by uudecoding and then opening the file as straight text, retyping the appropriate part, and then saving it AS TEXT. On my PC, I could not get Microsoft Word to save it as straight text - even choosing the 'text only' save method put extra carriage returns in there for me that screwed up Acrobat, so I recommend something more basic, but which shows actual character codes as well, because there must be Cntl-M's in between the lines, which is not the default for PC's. Microsoft Word will understand the ^M's to be a newline, but will, when saving as text, put in a ^J (I think) as well, so that won't work. In terms of the actual pdf code, the corruption started after the .92 in the final parameter of the Matrix statement, and went until just before the << that comes before /Length. Here is what you should see. /Matrix [0.4497 0.2446 0.02518 0.3163 0.672 0.1412 0.1845 0.08334 0.9227] >> ] endobj 3 0 obj << /Length .....etc There should be no unrecognizable characters between /Matrix and /Length. Note that there are supposed to be unrecognizable characters in other places both before and after this section. I am therefore getting my turns, and other people on my team have succesfully gotten their turns by following these instructions. This is a lot of hoops to jump through, and probably not everyone will be so successful, but send GSI a note, and hopefully they will fix this. I don't know what has changed since the test message that GSI sent out, but I had no problems like this with that test message. -- David Wenger | "Showerists believe in the healing, restorative dave@casbah.acns.nwu.edu | and contemplative properties of the Shower _The_Way_of_the_Shower_ | and the rituals devoted to the Shower." "In the name of the Shower, and the Water, and the Holy Steam, Amen." Up