Closed Bug 845786 Opened 11 years ago Closed 11 years ago

PDF "Preview on Firefox" not render some letters

Categories

(Firefox :: PDF Viewer, defect, P3)

19 Branch
x86
Windows XP
defect

Tracking

()

RESOLVED FIXED
Firefox 26

People

(Reporter: sir_maximinus, Unassigned)

References

Details

(Whiteboard: [pdfjs-c-rendering][pdfjs-d-font-conversion][pdfjs-f-fixed-upstream]https://github.com/mozilla/pdf.js/pull/3433#issuecomment-21750439)

Attachments

(3 files)

User Agent: Mozilla/5.0 (Windows NT 5.1; rv:19.0) Gecko/20100101 Firefox/19.0
Build ID: 20130215130331

Steps to reproduce:

With FireFox 19, I go to a page with an iframe that shows a PDF Document.




Actual results:

With the default option that FireFox has in Options > Applications > PDF, that is Preview in Firefox:

The PDF document doesn't view correctly. Some letters are not shown: C and P, but not in all paragraphs.

Using the plugin Adobe Acrobat (Options > Applications > PDF: Use Adobe Acrobat (in Firefox)) all the letters of the document PDF are shown correctly, as in previous version of Firefox.


Expected results:

The PDF must render well all letters.
Component: General → PDF Viewer
Product: Core → Firefox
Attachment #718943 - Attachment mime type: application/octet-stream → application/pdf
Test case to reproduce de bug. Shows an iframe that points to the previous attachment PDF.
Attached file TestCase.html
Priority: -- → P3
Whiteboard: [pdfjs-c-rendering][pdfjs-d-font-conversion]
i also have this problem and i have more info about the problem in many of my PDFs

The problem have its origins in Oracle... all the PDFs with this "missing letters" problem are generated by Oracle reports (or Oracle PDF driver if you want). 

At least version 10.2 of oracle forms/reports have this problem and i had in the past the same problem with xpdf and poppler library.

At that time i contacted the xpdf author and he found that oracle was using the font index 0 (??) for storing the first font characters, yep the pdf standard seems to say that it should start at 1, not zero. (i'm not a programmer, nor a pdf expert, so i may be saying things the wrong way, but that is the main idea)

The result is clear, for those PDF readers that start to read from index 1, they will drop the first character used by each font. In this testcase is the uppercase C, used in the left side border text. Every time there is a new font embedded in the document, its first letter disappears in all places where that font and letter is used (case sensitive!)

at the time, i switch to okular and acrobat reader and they show the PDFs fine. the xpdf author said that he would consider working around this problem on the next release... years past and with release 3.03 of xpdf the problem is solved.

So firefox should check what xpdf have done and implement the same workaround. Not fixing this is a big problem, as oracle reports is used in many companies and state/government documents.

i dont know if newer versions of oracle reports fix this problem.

hope this helps and this problem is fixed soon, i'm tired of disable PDF.js from my users browsers.
thanks
higuita
Seems that xpdf dont have cvs/svn/hg/git repository, so its hard to get the code for this change, but the changelog entry i think is this one: 

Tweaked the TrueType font fixup code to deal with an extra zero entry at the end of the loca table.
upstream have a patch for this:

https://github.com/mozilla/pdf.js/issues/3427
Whiteboard: [pdfjs-c-rendering][pdfjs-d-font-conversion] → [pdfjs-c-rendering][pdfjs-d-font-conversion][pdfjs-f-fixed-upstream]https://github.com/mozilla/pdf.js/pull/3433#issuecomment-21750439
Status: UNCONFIRMED → RESOLVED
Closed: 11 years ago
Depends on: 903452
Resolution: --- → FIXED
Target Milestone: --- → Firefox 26
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: