v6.18.1 Converted PDF (from epub): text is "invisible"

modified · 05-27-2023, 12:01 AM

I'm using `ebook-convert` to convert an epub (generated from a basic custom calibre recipe) into a pdf. The command looks basically like:

Code:

ebook-convert example-calibre6.18.1.epub example6.18.1.pdf --pdf-page-numbers --series=Test

With version 6.17.0 and earlier this was working perfectly, but with the latest version 6.18.1, the pdf generated contained text that is "white" or without colour. You cannot see the text in the PDF viewer (in my case, in macOS Preview) but if you do a "select-all", you can see the "invisible" text being selected.

I've attached the pdf output from 6.18.1 and 6.17.0 as well as the epub.

calibre version: 6.18.1
OS: Ubuntu 22.04.2 LTS, macOS Big Sur 11.7.5

Code:

DeDRM v10.0.3: Trying to decrypt example-calibre6.18.1.epub
DeDRM v10.0.3: Verifying zip archive integrity
DeDRM v10.0.3: Post-processing took 0.0 seconds
DeDRM v10.0.3: Finished after 0.1 seconds
Conversion options changed from defaults:
  series: 'Test'
  pdf_page_numbers: True
DeDRM v10.0.3: Trying to decrypt 1_bkwyyf.epub
DeDRM v10.0.3: Verifying zip archive integrity
DeDRM v10.0.3: Post-processing took 0.0 seconds
DeDRM v10.0.3: Finished after 0.0 seconds
1% Converting input to HTML...
qt.webenginecontext: 

GL Type: disabled
Surface Type: DefaultRenderableType
Surface Profile: CoreProfile
Surface Version: 4.1
QSG RHI Backend: OpenGL
Using Supported QSG Backend: yes
Using Software Dynamic GL: no
Using Multithreaded OpenGL: yes

Init Parameters:
  *  application-name calibre 
  *  browser-subprocess-path /Applications/calibre.app/Contents/Frameworks/QtWebEngineCore.framework/Helpers/QtWebEngineProcess.app/Contents/MacOS/QtWebEngineProcess 
  *  disable-features ConsolidatedMovementXY,InstalledApp,BackgroundFetch,WebOTP,WebPayments,WebUSB,PictureInPicture 
  *  disable-gpu  
  *  disable-speech-api  
  *  enable-features NetworkServiceInProcess,TracingServiceInProcess 
  *  enable-threaded-compositing  
  *  in-process-gpu  
  *  use-gl disabled 

InputFormatPlugin: EPUB Input running
on /var/folders/kr/bxxq54x51l714b1mrzyj0t1h0000gn/C/calibre_6.18.1_tmp_61l4ym7_/hw66hrhh.epub
Found HTML cover titlepage.xhtml
Parsing all content...
Referenced file 'feed_1/index.html' not found
34% Running transforms on e-book...
Merging user specified metadata...
Detecting structure...
Flattening CSS and remapping font sizes...
Source base font size is 12.00000pt
Removing fake margins...
Cleaning up manifest...
Trimming unused files from manifest...
Creating PDF Output...
67% Running PDF Output plugin
Failed to parse as XML, parsing as tag soup
68% Parsed all content for markup transformation
Link anchor: feed_0/index.html#article_3 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
Link anchor: feed_0/index.html#article_3 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
Link anchor: feed_0/index.html#article_4 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
Link anchor: feed_0/index.html#article_4 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
Link anchor: feed_0/index.html#article_2 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
Link anchor: feed_0/index.html#article_2 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
Link anchor: feed_0/index.html#article_0 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
Link anchor: feed_0/index.html#article_0 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
Link anchor: feed_0/index.html#article_1 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
Link anchor: feed_0/index.html#article_1 not found, linking to top of file instead
Link anchor: index_u3.html#feed_0 not found, linking to top of file instead
70% Completed markup transformation
90% Rendered all HTML as PDF
91% Added links to PDF content
93% Adding headers and footers
96% Headers and footers added
Merged 4 instances of AAAAAA+Times-Bold reducing size from 27.6 KB to 15 KB
Merged 2 instances of BAAAAA+Times-Roman reducing size from 9.4 KB to 7.8 KB
Merged 4 instances of CAAAAA+Times-Roman reducing size from 19.4 KB to 12.5 KB
Merged 2 instances of AAAAAA+Helvetica-Bold reducing size from 9.4 KB to 7.8 KB
Merged 2 instances of BAAAAA+Times-Bold reducing size from 10.7 KB to 9.1 KB
Merged 2 instances of DAAAAA+Times-Bold reducing size from 22.3 KB to 19.1 KB
Merged 4 instances of EAAAAA+Courier reducing size from 42.8 KB to 29.7 KB
Merged 4 instances of FAAAAA+Times-Roman reducing size from 50.1 KB to 25.1 KB
Merged 2 instances of BAAAAA+Courier reducing size from 10.8 KB to 10.7 KB
Merged 2 instances of CAAAAA+Helvetica-Bold reducing size from 9.9 KB to 8.3 KB
Merged 2 instances of DAAAAA+Helvetica-Oblique reducing size from 7.9 KB to 6.3 KB
Merged 6 instances of EAAAAA+Times-Bold reducing size from 48.2 KB to 18.5 KB
Merged 2 instances of GAAAAA+Times-Bold reducing size from 9.6 KB to 8.1 KB
Merged 8 instances of AAAAAA+Helvetica-Oblique reducing size from 37.6 KB to 7.7 KB
Merged 8 instances of BAAAAA+Helvetica-Bold reducing size from 39.7 KB to 8.3 KB
Merged 8 instances of CAAAAA+Times-Bold reducing size from 73 KB to 21.9 KB
Merged 6 instances of DAAAAA+Times-Roman reducing size from 82.2 KB to 27.9 KB
Merged 6 instances of FAAAAA+Times-Bold reducing size from 38.3 KB to 15.8 KB
Merged 2 instances of DAAAAA+Times-Italic reducing size from 16.7 KB to 15 KB
Merged 2 instances of EAAAAA+Times-Roman reducing size from 25.4 KB to 22.2 KB
Merged 2 instances of AAAAAA+Times-Roman reducing size from 8 KB to 7.4 KB
100% Updated metadata in PDF
PDF output written to /Users/me/Downloads/calibre_debugging/example6.18.1.pdf
Output saved to   /Users/me/Downloads/calibre_debugging/example6.18.1.pdf

kovidgoyal · 05-27-2023, 12:48 AM

this will be a bug in the updated PoDoFo shipped with 6.18 you can probably work around it by not generating page numbers. I will look for a better workaround later.

modified · 05-27-2023, 02:42 AM

Quote:

Originally Posted by kovidgoyal

this will be a bug in the updated PoDoFo shipped with 6.18 you can probably work around it by not generating page numbers. I will look for a better workaround later.

Can confirm removing the page numbers generation resolves the text issue.

Thanks Kovid, I will look out for the fix/workaround.

Comfy.n · 05-27-2023, 04:55 PM

This might be somewhat related. I see jumbled diacritics in PDF bookmark headings after conversions.

Uninstalled 6.18.1, reinstalled 6.17, then diacritics in PDF bookmarks display correctly after conversions.

I usually prefer to do PDF conversions with 3.48 Portable, because of this issue, so I wonder if it would be possible to bundle Calibre with both PDF engines (Podofo and the original one from versions < 3.48)

kovidgoyal · 05-27-2023, 10:43 PM

@Comfy.n: Please attach a sample epub file reproducing the issue or it wont get fixed. As for using the old engine, no its not possible as the old engine is basically an entire web browser that is now unmaintained.

EDIT: Nevermind this should fix it: https://github.com/kovidgoyal/calibr...e5a2118022587e

kovidgoyal · 05-27-2023, 11:12 PM

Hmm the blanking is happening only with binary builds, when running from source it doesn't happen, even when using the podofo library from the binary. Weird. This is going to be a pain to debug.

kovidgoyal · 05-28-2023, 12:41 AM

https://github.com/kovidgoyal/calibr...a57643b04dfdab

Comfy.n · 06-06-2023, 08:42 PM

Quote:

Originally Posted by kovidgoyal

@Comfy.n: Please attach a sample epub file reproducing the issue or it wont get fixed. As for using the old engine, no its not possible as the old engine is basically an entire web browser that is now unmaintained.

EDIT: Nevermind this should fix it: https://github.com/kovidgoyal/calibr...e5a2118022587e

jecilop · 06-20-2023, 06:10 PM

Quote:

Originally Posted by kovidgoyal

this will be a bug in the updated PoDoFo shipped with 6.18 you can probably work around it by not generating page numbers. I will look for a better workaround later.

@Kovid,

I can confirm this helped me also.
If it helps to know (although it seems you may have a track on the problem already from later replies....):
I'm using epubs made from "scratch"....Text opened in html or created in html program...converted in Calibre to epub. Added images to epubs.

I wanted to be able to print as a letter size if ever desired. I tested and found only the cover and page numbers would show.
NOTE: I tried another method..opened epub to read in Calibre. I tried printing to PDF from there. The results were the same.

I removed the "generate page number" as you said to try and it worked. I did the same via Print to PDF, and it also worked.

Just FYI..hope this info is helpful.

Using Calibre 6.21 and the available book conversion and basic settings.

05-27-2023, 10:43 PM	#5
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	@Comfy.n: Please attach a sample epub file reproducing the issue or it wont get fixed. As for using the old engine, no its not possible as the old engine is basically an entire web browser that is now unmaintained. EDIT: Nevermind this should fix it: https://github.com/kovidgoyal/calibr...e5a2118022587e Last edited by kovidgoyal; 05-28-2023 at 12:59 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Converted EPUB from PDF has no text	Ansem_93	Conversion	2	03-23-2022 07:25 AM
Tool to OCR an "image" PDF → add text as extra layer?	Shohreh	PDF	5	12-19-2020 01:47 PM
I imported a "kepub.ebub" book into Calibre, but Calibre converted to "epub". Why?	droopy	Calibre	5	12-23-2019 08:33 AM
what's the difference between "compiled" and "converted" folders?	Gregg Bell	Kindle Formats	8	04-01-2014 01:45 AM
Problem with Epub to Pdf Conversion:text invisible on right hand side due to clipping	Feher	Conversion	2	06-08-2013 04:58 AM

05-27-2023, 12:48 AM	#2
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	this will be a bug in the updated PoDoFo shipped with 6.18 you can probably work around it by not generating page numbers. I will look for a better workaround later.

05-27-2023, 04:55 PM	#4
Comfy.n want to learn what I want Posts: 1,682 Karma: 7908443 Join Date: Sep 2020 Device: none	This might be somewhat related. I see jumbled diacritics in PDF bookmark headings after conversions. Uninstalled 6.18.1, reinstalled 6.17, then diacritics in PDF bookmarks display correctly after conversions. I usually prefer to do PDF conversions with 3.48 Portable, because of this issue, so I wonder if it would be possible to bundle Calibre with both PDF engines (Podofo and the original one from versions < 3.48)

05-27-2023, 11:12 PM	#6
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Hmm the blanking is happening only with binary builds, when running from source it doesn't happen, even when using the podofo library from the binary. Weird. This is going to be a pain to debug.

05-28-2023, 12:41 AM	#7
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	https://github.com/kovidgoyal/calibr...a57643b04dfdab