[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
On Wed, 4 May 2011, Martijn Grooten wrote:
Bit off an odd thing I noticed today. I recently sent someone a PDF I had generated using LaTeX (via latex, dvipdf). He then wanted to have the LaTeX source and as I had several versions of the .tex file lying around, I took the one I thought was the latest version and regenerated a PDF. Then I took md5 checksums of the PDF files... and they turned out to be different. After I convinced myself that the .tex file was indeed the latest version by other means, I ran dvipdf again... and the checksum was different was again. I repeated this a few times and the checksum was different every time. Is there a reason for this apparent randomness? The files are binaries so diff doesn't tell me anything but the fact that they are different. I can view them with less, bar some non-printable characters, and they look the same.
Might they have a date/time generated encoded into them? You could always get the source for dvipdf :)Ah, I've just done a check myself - you can 'less' the .pdf file, but go to the end:
In one old file: /CreationDate(D:20100212090550Z00'00') /ModDate(D:20100212090550Z00'00')>>endobj When I've re-dvips'd it, I see: /CreationDate(D:20110504222730+01'00') /ModDate(D:20110504222730+01'00')>>endobj Gordon -- The Mailing List for the Devon & Cornwall LUG http://mailman.dclug.org.uk/listinfo/list FAQ: http://www.dcglug.org.uk/listfaq