比萨生成的pdf在Chrome中可见，但在Adobe Reader中不可见

Question

I am trying to save a rendered html page in flask as pdf. I tried the following things:

pdf=render_template('exp1_post.html',some_data=some_data)
filename = "simplePrint.pdf"        
pisa.CreatePDF(pdf, file(filename, "w"))

and

pdf=render_template('exp1_post.html',some_data=some_data)
filename = "simplePrint.pdf"        
pisa.CreatePDF(pdf.encode("ISO-8859-1"), file(filename, "w"))

but generated pdf file is perfectly visible in Google Chrome (opened directly by going to the location on disk), but i get blank page in Adobe acrobat reader.

Similar problem that i can find: PDF text show in Google Chrome but not in Adobe Acrobat

but i am not sure how to implement above solution in python Chrome Version 41.0.2272.118 m Adobe Reader XI

Answer 1

I'm not a Python expert (I know the name, that's about it) so my solution may not be correct. But I do think I know what the problem is and that should help you find the solution you're looking for.

When you open your file with a binary editor (or a good text editor showing invisible characters) you'll see that each line ending consists of three line-ending characters:

x0D x0D x0A

Or otherwise said

Carriage Return, Carriage Return, Line feed

This is wrong. From the PDF specification:

"As a matter of convention, the tokens in a PDF file are arranged into lines; see 7.2, "Lexical Conventions."Each line shall be terminated by an end-of-line (EOL) marker, which may be a CARRIAGE RETURN (0Dh), a LINE FEED (0Ah), or both. PDF files with binary data may have arbitrarily long lines."

I think this is what breaks Adobe Reader. It's strange that Adobe Reader and Acrobat throw a fit on this file, while many other (worse) PDF readers (such as Mac OS X Preview) seem to show it without any problem.

All of this said, you seem to have a problem with line endings. Given my limited knowledge of Python all I might point you to is the line:

file(filename, "w")

I read in Python documentation that on some platforms this can treat files as ASCII files and ruin binary files. As a PDF is definitely a binary file, I would change that to:

file(filename, "wb")

and see what happens.

I can tell you that for a far as I could see the rest of the file structure seems correct. So I think you have all necessary objects etc to show the file correct (as proven by Chrome and Mac Preview), so I really think the line ending problem is the one you need to solve.