Why merge pdf is flipped with example.. about pdfrw HOT 6 CLOSED

@sunqianxiang i wrote some code based on the fancy_watermark.py, no issue with flipping.

from pdfrw.

Either your watermark or your PDF pages are probably rotated. (Probably the PDF page, I think rotation for the watermark is handled.) If you have examples (especially if you have examples I can add to the test suite) please email them to me and I will fix.

Thanks,
Pat

from pdfrw.

but I maybe need normal merger..Such as for example with PyPDF2..

from PyPDF2 import PdfFileWriter, PdfFileReader
output = PdfFileWriter()
ipdf = PdfFileReader(open('output.pdf', 'rb'))
wpdf = PdfFileReader(open('transafe.pdf', 'rb'))
watermark = wpdf.getPage(0)
for i in xrange(ipdf.getNumPages()):
   try:
      page = ipdf.getPage(i)
      page.mergePage(watermark)
      output.addPage(page)
   except:
      pass
with open('newfile.pdf', 'wb') as f:
   output.write(f)

here generate not flipping pdf from code..but it have problem in pdf 1.4.
i think watemark need custom options or providing method to adjustment.
in pdf1.4 maybe have the problam.i can send mail to you for my pdf file..
Thanks in advance

from pdfrw.

Thank you for this bug report. I have checked in some code on a new branch that fixes this problem.

When I did a rewrite of my merge code, I forgot to save graphics state if the object was a regular page and not a form XObject. Most PDFs work fine this way, but your example PDF alters the graphics state and leaves it altered (which is perfectly legal, but which I did not have a failing PDF for).

I have not yet checked the fix into the master because I will first need to update expected checksums for the output of the watermark test. There is one changed file, here:

https://github.com/pmaupin/pdfrw/blob/pagemerge_state_save_restore/pdfrw/pagemerge.py

Please verify it works on all your files.

Also, if you can create a non-sensitive file that the old one fails on and the new one works on I will add it to the test suite.

@ThomasWaldmann -- you may also want this fix, depending on how much control you have over the PDFs you are watermarking -- it is a definite bug in pdfrw, but the PDFs you provided so far did not provoke it.

Thanks,
Pat

from pdfrw.

yup, works fine, had the same issue

from pdfrw.

The branch to fix this is now merged, so I'm closing this issue.

from pdfrw.