oofy - 18-5-2011 at 13:03
The PDF for Gergel--Isopropyl Bromide in the library is poisoned to the point where I have to reboot my OS X Mac or my iPad (1) if I try to open it.
This is a very bad crash. I have tried running it through pdftk with no success. It is the first PDF I have ever had this problem with. I see that it
was created with ABBYY Finereader. I wonder if someone could possibly re-encode this PDF with a later version or more compatibility options so that it
doesn't irretrievably crash on a Mac or iPad.
I'm looking forward to reading it. Thanks.
[Edited on 18-5-2011 by oofy]
Polverone - 18-5-2011 at 14:59
The MD5 checksum for the file is f019e2722bd64684e2093a0933e390f4. Ensure that it has downloaded cleanly.
If you have to reboot OS X to recover from a PDF problem I think that there is something wrong with your hardware. I don't have a Mac myself but I
have successfully viewed this PDF using Adobe Acrobat 9, evince, and xpdf under Windows and Linux.
oofy - 18-5-2011 at 16:39
I have redownloaded it several times. I have tried it on three different OS X machines (and my iPad, which runs IOS, not OS X).
In both cases (IOS and OS X) the machines lock up because of excessive resource usage and have to be rebooted. This is extremely rare for OS X and
unheard of on an iPad.
I have tried a variety of PDF software for both platforms with the same result.
It seems reasonable to assume that there is a Postscript bug in the PDF file that is creating a loop that uses up all of the available resources of
the various machines. Postscript from the PDF itself is the only executable code common to all these environments.
**EDIT** the MD5 matches. I just checked it. The file is downloaded correctly. I would simply install Acrobat Professional and re-encode the file
myself, except that I am on a boat at sea and it is therefore inconvenient at the moment. And that would not solve the problem for the other Mac and
iPad users out there.
[Edited on 19-5-2011 by oofy]
Polverone - 18-5-2011 at 17:50
If there were an infinite loop in the file's Postscript code it should lock up all readers, just as an infinite loop in a web page's Javascript should
cause problems on all web browsers. Yet I have viewed the book successfully with Adobe's Acrobat Reader 9 and two other less common PDF viewers. I
have only 1 GB of RAM on my test machine, so I don't think it needs a lot of resources. You have tried to open the file under OS X with a recent
version of Adobe Acrobat Reader, and it consumes so much RAM that the machine becomes unresponsive and needs rebooting?
I was successfully able to run operations on the file with pdftk 1.41. When you say you tried pdftk without success, do you mean that pdftk had
problems processing the file or simply that the processed file emitted by pdftk still cannot be read on your machines?
Sedit - 18-5-2011 at 18:25
Can we link to the file? My computers a piece of shit so if I manage to open it I would be pretty sure its on your end.
oofy - 19-5-2011 at 07:50
<a href=http://library.sciencemadness.org/library/books/gergel_isopropyl_bromide.pdf>Here is the file.</a>
Pdftk successfully reads the file, does whatever I tell it to do, and successfully writes the output file. No errors. But also no change in the
crashing behaviour.
Acrobat Reader v8 displays the pdf on OS X (I tried that after my last posting--I did not have Acrobat readily available, did not really have the
bandwidth to download it because it has become huge, and do not wish to use it). Nothing else can read the file. Preview chews on the file for a
while, doesn't display anything, then goes unresponsive and requires a force-kill and reboot, if enough time has elapsed. Acrobat Reader won't let me
makes changes or save or print the file to see if I can create something that doesn't crash all my other pdf readers on OS X or iPad.
I do not have the power available to read the entire book on my laptop in Acrobat (we're 100% solar around here). I can really only read it on my
iPad. But that is peripheral to the fact that PDF is explicitly a cross-platform document format and this file crashes Preview and all other
third-party Mac and iPad software. There is something wrong with it. Hopefully, pointing that out will help others who wish to read it with something
other than Windows/Adobe. The question is: what is wrong with this file that Acrobat can tolerate but nothing else on Mac can?
I read hundreds of PDFs on my iPad every month. Not one of them has ever crashed, even once, since April 2010. There is nothing wrong at my end.
Thanks for the suggestions.
Polverone - 19-5-2011 at 09:07
The file's page images are compressed with JBIG2, which is a little "exotic" by the standards of PDF. But it's been in fairly common use since 2003.
There are two possibilities: 1) Preview and whatever reads PDF on the iPad do not handle JBIG2. 2) There is something specifically problematic about
the Gergel file, not JBIG2 in general.
To cross-check, please try downloading/reading Ignition!, which was created by a similar process and is also encoded in JBIG2. If you have trouble with that too I guess the Apple-bundled
applications don't handle JBIG2.
Sedit - 19-5-2011 at 09:11
It worked on both my computers no problem.
not_important - 19-5-2011 at 10:55
No problems here on several machines, *NIX and Win and OS X and Vista over OS X. Recent versions of Acrobat and Foxit Reader.
oofy - 20-5-2011 at 00:34
I downloaded Ignition! and it crashes and burns just like the Gergel file. JBIG2 appears to be the problem. Now to find out what the heck JBIG2 is.
Thanks.
*EDIT* lots of reports out there of JBIG2 crashing Coregraphics with a heap overflow and many others report the same behaviour: excessive allocation
of memory with eventual slowdown, crash etc. Funny that it affects both OS X and IOS, but apparently that is so.
I had someone convert the file to HTML and back to PDF and it resolves the crash, but the images are so poor that there is garbage scattered
throughout the text, making it unreadable. It looks perfect until you try to read it, then you run across the garbage and get stuck. Must be OCR
errors.
[Edited on 20-5-2011 by oofy]
#maverick# - 20-5-2011 at 08:30
works on the shitty machines at school i believe its probably on your part
mr.crow - 20-5-2011 at 11:15
I have had multiple PDFs from the reference section crash my Mac pdf viewer. I only use it because its my computer at work. Macs aren't all they are
cracked up to be
Perhaps there are some alternative PDF viewers? I use one for windows. Adobe pdf is notorious
kmno4 - 21-5-2011 at 03:04
IrfanView and gsview32 do not read at all JBIG2 compressed files, also this "poisoned" one.
I removed compression, try it:
http://ifile.it/z2jx0oe
Both viewers can read it now.
[Edited on 21-5-2011 by kmno4]