poppler issueshttps://gitlab.freedesktop.org/poppler/poppler/-/issues2018-10-25T18:17:41Zhttps://gitlab.freedesktop.org/poppler/poppler/-/issues/318[PATCH] Seccomp sandbox support for pdftotext2018-10-25T18:17:41ZBugzilla Migration User[PATCH] Seccomp sandbox support for pdftotext## Submitted by valo
Assigned to **poppler-bugs**
**[Link to original bug (#100224)](https://bugs.freedesktop.org/show_bug.cgi?id=100224)**
## Description
Created attachment 130253
seccomp support for pdftotext
Since some of the ...## Submitted by valo
Assigned to **poppler-bugs**
**[Link to original bug (#100224)](https://bugs.freedesktop.org/show_bug.cgi?id=100224)**
## Description
Created attachment 130253
seccomp support for pdftotext
Since some of the poopler tools, like pdftotext are used by some file managers to automatically parse pdf files for preview, I thought it might be a good idea to use some sandboxing.
This is a patch that adds seccomp filter to pdftotext. This can also be applied to the other tools that poppler provides, reducing the risk of successful exploitation of poppler (and other used library) vulnerabilities significantly.
I found this quite easy to apply and would be happy to help if you are interested in using this.
This patch can be applied to poppler 0.52.0 without further changes
**Patch 130253**, "seccomp support for pdftotext":
[pdftotext_seccomp.patch](/uploads/92965944b5d4725bc29bd5f741d4c83c/pdftotext_seccomp.patch)https://gitlab.freedesktop.org/poppler/poppler/-/issues/317All characters in this PDF are replaced by boxes2018-10-07T00:34:12ZBugzilla Migration UserAll characters in this PDF are replaced by boxes## Submitted by Michael Fiedler
Assigned to **poppler-bugs**
**[Link to original bug (#91377)](https://bugs.freedesktop.org/show_bug.cgi?id=91377)**
## Description
Created attachment 117205
Screenshot
When displaying the attached...## Submitted by Michael Fiedler
Assigned to **poppler-bugs**
**[Link to original bug (#91377)](https://bugs.freedesktop.org/show_bug.cgi?id=91377)**
## Description
Created attachment 117205
Screenshot
When displaying the attached PDF document using a Poppler-based viewer such as Evince, all characters are replaced by boxes (see screenshot). As other PDF viewers such as Firefox/PDF.js, Google Chrome, MuPDF, Adobe Reader display it correctly, I suppose it is a poppler bug. The bug can also be reproduced by using utils/pdftocairo -pdf and then viewing the output with a non-poppler-based viewer.
The affected PDF file was created by the online interface of the German self-publishing platform epubli.de from my original document, so I do not exactly know which software they used for that.
My test environment:
Debian GNU/Linux 7.8 (wheezy) (x86)
Evince/GNOME Document Viewer 3.4.0
Poppler 0.18.4-6 (from Debian)
Poppler 0.34.0
Poppler Git commit 1aa2f6e8a41a6a86dc02bf7c5cbc62355e780961
All tested poppler versions are affected. I verified it using
utils/pdftocairo -pdf broken.pdf out.pdf
$other_pdf_viewer out.pdf
**Attachment 117205**, "Screenshot":
![screenshot](/uploads/1cce914613a070433be33751b2f59f96/screenshot.png)https://gitlab.freedesktop.org/poppler/poppler/-/issues/316Poppler shows a glyph when it should not2018-08-21T10:40:03ZBugzilla Migration UserPoppler shows a glyph when it should not## Submitted by Marek Kasik `@mkasik`
Assigned to **poppler-bugs**
**[Link to original bug (#106898)](https://bugs.freedesktop.org/show_bug.cgi?id=106898)**
## Description
Cairo backend shows a glyph when rendering a text using Ty...## Submitted by Marek Kasik `@mkasik`
Assigned to **poppler-bugs**
**[Link to original bug (#106898)](https://bugs.freedesktop.org/show_bug.cgi?id=106898)**
## Description
Cairo backend shows a glyph when rendering a text using Type3 font even if there is no glyph for current char code. This was reported to me downstream here together with reproducer:
https://bugzilla.redhat.com/show_bug.cgi?id=1589594
The PDF seems wrong because it does not have glyph for space in its Type3 font.
Looking at the code, the codeToGID table of CairoFont is initialized with zeros and only existing glyphs get a valid index.
I propose to initialize the table with -1s and propagate this information through getGlyph() so the CairoOutputDev::drawChar() can check whether the index was actually set.https://gitlab.freedesktop.org/poppler/poppler/-/issues/315transparent background2018-10-26T15:11:54ZBugzilla Migration Usertransparent background## Submitted by Benoit Decherf
Assigned to **poppler-bugs**
**[Link to original bug (#35333)](https://bugs.freedesktop.org/show_bug.cgi?id=35333)**
## Description
The splash backend ignore the alpha channel of the paperColor.
I u...## Submitted by Benoit Decherf
Assigned to **poppler-bugs**
**[Link to original bug (#35333)](https://bugs.freedesktop.org/show_bug.cgi?id=35333)**
## Description
The splash backend ignore the alpha channel of the paperColor.
I use poppler-qt4https://gitlab.freedesktop.org/poppler/poppler/-/issues/314Pages from image based PDF do not display2020-11-05T22:00:03ZBugzilla Migration UserPages from image based PDF do not display## Submitted by Carlos Campos `@carlosgc`
Assigned to **poppler-bugs**
**[Link to original bug (#24828)](https://bugs.freedesktop.org/show_bug.cgi?id=24828)**
## Description
Bug forwarded from Evince: https://bugzilla.gnome.org/sh...## Submitted by Carlos Campos `@carlosgc`
Assigned to **poppler-bugs**
**[Link to original bug (#24828)](https://bugs.freedesktop.org/show_bug.cgi?id=24828)**
## Description
Bug forwarded from Evince: https://bugzilla.gnome.org/show_bug.cgi?id=600181
"Evince version: 2.28.1 on Ubuntu 9.10.
Evince cannot display any pages of the document without significant errors
--the PDF consists entirely of image files containing text, and almost all of
the text is invisible. Sometimes multiple pages of the same document are
displayed where the user would expect only one page (again, the text within the
image is missing). This behavior is inconsistent and I don't know how to
consistently reproduce it.
Xpdf displays the document without errors.
Thank you."
"The PDF file can be downloaded from here:
http://damonlynch.net/evince-problem.pdf"
### Depends on
* [Bug 27456](https://bugs.freedesktop.org/show_bug.cgi?id=27456)https://gitlab.freedesktop.org/poppler/poppler/-/issues/313Vertical text is shown as horizontal in output HTML2018-08-21T10:39:52ZBugzilla Migration UserVertical text is shown as horizontal in output HTML## Submitted by Nitesh G.
Assigned to **poppler-bugs**
**[Link to original bug (#65890)](https://bugs.freedesktop.org/show_bug.cgi?id=65890)**
## Description
Created attachment 80994
Input PDF file
Hi,
I have tried to convert th...## Submitted by Nitesh G.
Assigned to **poppler-bugs**
**[Link to original bug (#65890)](https://bugs.freedesktop.org/show_bug.cgi?id=65890)**
## Description
Created attachment 80994
Input PDF file
Hi,
I have tried to convert the attached PDF to HTML(using pdftohtml.exe) and found that the text "Media Services" is shown horizontal instead of vertical
I am attaching the reference PDF.
Thanks,
Nitesh
**Attachment 80994**, "Input PDF file":
[pdftohtml_isues_2.pdf](/uploads/10bf9fc6ff97ad4496b983d956a16e3b/pdftohtml_isues_2.pdf)https://gitlab.freedesktop.org/poppler/poppler/-/issues/312Data overlapping issue in PDF to HTML conversion2018-08-21T10:39:40ZBugzilla Migration UserData overlapping issue in PDF to HTML conversion## Submitted by Nitesh G.
Assigned to **poppler-bugs**
**[Link to original bug (#65888)](https://bugs.freedesktop.org/show_bug.cgi?id=65888)**
## Description
When converting the attached PDF to HTML, I can see data overlapping iss...## Submitted by Nitesh G.
Assigned to **poppler-bugs**
**[Link to original bug (#65888)](https://bugs.freedesktop.org/show_bug.cgi?id=65888)**
## Description
When converting the attached PDF to HTML, I can see data overlapping issues on Page 160 and Page 161 and more.
Observations:-
1. The issue is only reproducible if I open the HTML in IE Explorer and not reproducible if I open it in Mozilla and Firefox. Given that IE is the standard go to OOTB browser in Windows (really!), this is a valid bug.
2. If I extract pages 160 and 161 from PDF and then try to convert to HTML, then I cannot reproduce the issue.
Thanks,
Niteshhttps://gitlab.freedesktop.org/poppler/poppler/-/issues/311250KB PDF converted to 50MB PDF2018-08-21T10:39:37ZBugzilla Migration User250KB PDF converted to 50MB PDF## Submitted by Ofir
Assigned to **poppler-bugs**
**[Link to original bug (#90440)](https://bugs.freedesktop.org/show_bug.cgi?id=90440)**
## Description
Created attachment 115751
PDF to convert
Tested with poppler-utils 0.30.0 Ub...## Submitted by Ofir
Assigned to **poppler-bugs**
**[Link to original bug (#90440)](https://bugs.freedesktop.org/show_bug.cgi?id=90440)**
## Description
Created attachment 115751
PDF to convert
Tested with poppler-utils 0.30.0 Ubuntu 15.04 64bit.
Converting the attached PDF to SVG with pdftocairo give me a 50MB SVG which crashes viewers.
It takes about a minute for Evince to display the PDF compared to 1 second with Foxit Reader.
**Attachment 115751**, "PDF to convert":
[Flow-background.pdf](/uploads/85cff448f7d860696648282d61c873f6/Flow-background.pdf)https://gitlab.freedesktop.org/poppler/poppler/-/issues/310Quotation mark is selected at the wrong position2018-08-21T10:39:32ZBugzilla Migration UserQuotation mark is selected at the wrong position## Submitted by Germán Poo-Caamaño
Assigned to **poppler-bugs**
**[Link to original bug (#106777)](https://bugs.freedesktop.org/show_bug.cgi?id=106777)**
## Description
Created attachment 139960
PDF test case
The issue was origin...## Submitted by Germán Poo-Caamaño
Assigned to **poppler-bugs**
**[Link to original bug (#106777)](https://bugs.freedesktop.org/show_bug.cgi?id=106777)**
## Description
Created attachment 139960
PDF test case
The issue was originally reported in https://gitlab.gnome.org/GNOME/evince/issues/248
In the attached PDF, page 1, paragraph 1, when selecting the first paragraph, the quotation mark at the beginning cannot be selected, therefore, I cannot be copied. However, if one try to select until right before the next paragraph, then the quotation mark is selected, and copied at the end of the paragraph.
The text:
„Kein Stein wird auf dem anderen bleiben!" Das ist die Quintessenz Roland Ber-
gers über die Perspektiven der Informationsgesellschaft in seinem Gutachten für
die Enquete-Kommission: „Zehn Thesen zur Ökonomie der Informationsgesell-
schaft am Standort Deutschland im 21. Jahrhundert". 1 )
is copied as:
Kein Stein wird auf dem anderen bleiben!" Das ist die Quintessenz Roland Ber-
gers über die Perspektiven der Informationsgesellschaft in seinem Gutachten für
die Enquete-Kommission: „Zehn Thesen zur Ökonomie der Informationsgesell-
schaft am Standort Deutschland im 21. Jahrhundert". 1 )
„
It can be reproduced wit poppler-glib-demo.
**Attachment 139960**, "PDF test case":
[1311004.pdf](/uploads/0c0655498d292e763366d30745e3ffc3/1311004.pdf)https://gitlab.freedesktop.org/poppler/poppler/-/issues/309Add image names to pdftohtml dump in xml mode2018-10-27T15:01:13ZBugzilla Migration UserAdd image names to pdftohtml dump in xml mode## Submitted by Raphaël Monrouzeau
Assigned to **poppler-bugs**
**[Link to original bug (#29480)](https://bugs.freedesktop.org/show_bug.cgi?id=29480)**
## Description
Created attachment 37764
First patch: Add image names to pdftoh...## Submitted by Raphaël Monrouzeau
Assigned to **poppler-bugs**
**[Link to original bug (#29480)](https://bugs.freedesktop.org/show_bug.cgi?id=29480)**
## Description
Created attachment 37764
First patch: Add image names to pdftohtml dump in xml mode
I wanted pdftohtml to dump information about images in a page in xml mode.
The first patch below makes pdftohtml generate images as without the -xml switch
and dumps its name in the xml file; please review it, I'm open to suggestions, style requests and everything required.
Here is the description of the patch:
The -c (complex) and -xml modes are not linked anymore.
The -c switch has no real effect on -xml mode (as before).
However the -i switch is now looked at in -xml mode.
Without it images are now generated and image tags
do reference their name. The DTD has been updated.
**Patch 37764**, "First patch: Add image names to pdftohtml dump in xml mode":
[a.patch](/uploads/e027720d19156c3bd7f8557b13f26eab/a.patch)
### Blocking
* [Bug 29482](https://bugs.freedesktop.org/show_bug.cgi?id=29482)https://gitlab.freedesktop.org/poppler/poppler/-/issues/308Support for writing Encrypted data2018-10-27T15:09:04ZBugzilla Migration UserSupport for writing Encrypted data## Submitted by Carlos Campos `@carlosgc`
Assigned to **poppler-bugs**
**[Link to original bug (#17901)](https://bugs.freedesktop.org/show_bug.cgi?id=17901)**
## Description
Bug forwarded from Evince: http://bugzilla.gnome.org/sho...## Submitted by Carlos Campos `@carlosgc`
Assigned to **poppler-bugs**
**[Link to original bug (#17901)](https://bugs.freedesktop.org/show_bug.cgi?id=17901)**
## Description
Bug forwarded from Evince: http://bugzilla.gnome.org/show_bug.cgi?id=552875
"Please describe the problem:
Saving a copy of a PDF document works fine. But as soon as you enter something
into a form field and then save a copy, only the form fields including their
data are saved into the copy. The whole background is omitted however.
This might work for printing only the entered data but not for keeping the
document including the data. Without the background the document is useless
Steps to reproduce:
1. Open a PDF with forms
2. Enter something
3. Save a copy
Actual results:
Evince only saves form fields and data.
Expected results:
Evince should also save the background.
Does this happen every time?
Yes"
A test case can be found attached to the original bug report.https://gitlab.freedesktop.org/poppler/poppler/-/issues/307The Chinese Pdf can not be displayed correctly and selected on evince 2.20.1 ...2018-10-26T15:14:49ZBugzilla Migration UserThe Chinese Pdf can not be displayed correctly and selected on evince 2.20.1 and popler0.6.2.## Submitted by zhao wei
Assigned to **poppler-bugs**
**[Link to original bug (#16305)](https://bugs.freedesktop.org/show_bug.cgi?id=16305)**
## Description
Created attachment 17052
the sample that can not be displayed correctly
...## Submitted by zhao wei
Assigned to **poppler-bugs**
**[Link to original bug (#16305)](https://bugs.freedesktop.org/show_bug.cgi?id=16305)**
## Description
Created attachment 17052
the sample that can not be displayed correctly
Some English letters will be displayed under other letters.
When you select some words (the words include English and Chinese at the same time), the words that were selected can not be displayed correctly. But, the result of the copy is right.
**Attachment 17052**, "the sample that can not be displayed correctly":
[xzl998.pdf](/uploads/8ec57848f1f8be8ad58cc0a7753b55d4/xzl998.pdf)https://gitlab.freedesktop.org/poppler/poppler/-/issues/306poppler: oss-fuzz integration2018-10-25T09:15:17ZBugzilla Migration Userpoppler: oss-fuzz integration## Submitted by pdknsk
Assigned to **poppler-bugs**
**[Link to original bug (#106554)](https://bugs.freedesktop.org/show_bug.cgi?id=106554)**
## Description
Created attachment 139615
cpp
I'm interested if you're interested in hav...## Submitted by pdknsk
Assigned to **poppler-bugs**
**[Link to original bug (#106554)](https://bugs.freedesktop.org/show_bug.cgi?id=106554)**
## Description
Created attachment 139615
cpp
I'm interested if you're interested in having poppler integrated into oss-fuzz. In case you're not familiar with it, it's a Google-sponsored project for continuous fuzzing of OSS.
https://github.com/google/oss-fuzz
You may have noticed that I reported a few bugs recently, which were found in preparation for possible oss-fuzz integration.
You don't really have to do anything, other than give an email address (or multiple) with are then CC'd to bug reports found by the fuzzing system. (Only project owners can be CC'd.)
oss-fuzz reports the bugs on its own bug tracker, rather than the project's bug tracker. When a bug is fixed, oss-fuzz detects this automatically and closes the bug report.
https://bugs.chromium.org/p/oss-fuzz/issues/list
Possible optional future steps would be to host the fuzz target directly in the poppler repo (so that you can modify it yourself directly, rather than submitting PRs through GitHub). Or the addition of more fuzz targets. I only went with the cpp API so far, as it seems the easiest to integrate.
A fuzz target is just a single source file. As an example, I'm attaching the fuzz target for the cpp API I've been using.
**Attachment 139615**, "cpp":
[target.cc](/uploads/fe3fd56ad59ae4e5b178279fd006e3ba/target.cc)https://gitlab.freedesktop.org/poppler/poppler/-/issues/305override font to use in pdf documents2018-08-21T10:38:48ZBugzilla Migration Useroverride font to use in pdf documents## Submitted by Dotan Cohen
Assigned to **poppler-bugs**
**[Link to original bug (#23546)](https://bugs.freedesktop.org/show_bug.cgi?id=23546)**
## Description
PDF documents specify the font to use to render the document (or even ...## Submitted by Dotan Cohen
Assigned to **poppler-bugs**
**[Link to original bug (#23546)](https://bugs.freedesktop.org/show_bug.cgi?id=23546)**
## Description
PDF documents specify the font to use to render the document (or even embed the font if it is not one of the PDF defaults). Please let the user override this font with a font of his selection. This is a usability and an accessibility issue.
Thanks.https://gitlab.freedesktop.org/poppler/poppler/-/issues/304splash backend renders poorlier than cairo's2018-08-21T10:38:40ZBugzilla Migration Usersplash backend renders poorlier than cairo's## Submitted by nanmus
Assigned to **poppler-bugs**
**[Link to original bug (#18379)](https://bugs.freedesktop.org/show_bug.cgi?id=18379)**
## Description
Created attachment 20056
Above is Evince, below is Okular.
Please refer to...## Submitted by nanmus
Assigned to **poppler-bugs**
**[Link to original bug (#18379)](https://bugs.freedesktop.org/show_bug.cgi?id=18379)**
## Description
Created attachment 20056
Above is Evince, below is Okular.
Please refer to the attached picture.
**Attachment 20056**, "Above is Evince, below is Okular.":
![kde-pdf](/uploads/9d265656d6344e8af69647b754dc2ce8/kde-pdf.png)https://gitlab.freedesktop.org/poppler/poppler/-/issues/303Incorrect handling of odd length ASCIIHexStream data2018-10-05T22:49:20ZBugzilla Migration UserIncorrect handling of odd length ASCIIHexStream data## Submitted by Ilya Gorenbein
Assigned to **poppler-bugs**
**[Link to original bug (#37784)](https://bugs.freedesktop.org/show_bug.cgi?id=37784)**
## Description
Created attachment 47352
Sample file
The number of symbols at ASCI...## Submitted by Ilya Gorenbein
Assigned to **poppler-bugs**
**[Link to original bug (#37784)](https://bugs.freedesktop.org/show_bug.cgi?id=37784)**
## Description
Created attachment 47352
Sample file
The number of symbols at ASCIIHexStream stream should be even. If the number of bytes is odd the last byte should be eliminated. Attached is a sample file. Adobe is able to run it, but alerts that there was a problem.
**Attachment 47352**, "Sample file":
[dbg.pdf](/uploads/e412499ece8686879b6e9cacdd250031/dbg.pdf)https://gitlab.freedesktop.org/poppler/poppler/-/issues/302Poppler should not give compile warnings2018-10-05T22:49:56ZBugzilla Migration UserPoppler should not give compile warnings## Submitted by Hib Eris
Assigned to **poppler-bugs**
**[Link to original bug (#28897)](https://bugs.freedesktop.org/show_bug.cgi?id=28897)**
## Description
Hi,
I am on a mission again to remove compiler warnings. I will submit a...## Submitted by Hib Eris
Assigned to **poppler-bugs**
**[Link to original bug (#28897)](https://bugs.freedesktop.org/show_bug.cgi?id=28897)**
## Description
Hi,
I am on a mission again to remove compiler warnings. I will submit a few patches for it.https://gitlab.freedesktop.org/poppler/poppler/-/issues/301Vulnerability discovered in Poppler PDF2018-10-05T22:50:26ZBugzilla Migration UserVulnerability discovered in Poppler PDF## Submitted by reg..@..re.com
Assigned to **poppler-bugs**
**[Link to original bug (#101062)](https://bugs.freedesktop.org/show_bug.cgi?id=101062)**
## Description
Hello,
The Cisco Talos team found a security vulnerability impac...## Submitted by reg..@..re.com
Assigned to **poppler-bugs**
**[Link to original bug (#101062)](https://bugs.freedesktop.org/show_bug.cgi?id=101062)**
## Description
Hello,
The Cisco Talos team found a security vulnerability impacting Poppler PDF customers. As this is a sensitive security issue, this report is to request a PGP key for further communication. Please acknowledge receipt of this email so we can confirm we have the right system for reporting security issues with Poppler PDF.https://gitlab.freedesktop.org/poppler/poppler/-/issues/300Noncompliance with Standard-14 fonts requirement in PDF specs2021-11-01T11:24:52ZBugzilla Migration UserNoncompliance with Standard-14 fonts requirement in PDF specs## Submitted by Stefan Löffler
Assigned to **poppler-bugs**
**[Link to original bug (#49037)](https://bugs.freedesktop.org/show_bug.cgi?id=49037)**
## Description
Created attachment 60413
Patch for handling of Standard-14 fonts on...## Submitted by Stefan Löffler
Assigned to **poppler-bugs**
**[Link to original bug (#49037)](https://bugs.freedesktop.org/show_bug.cgi?id=49037)**
## Description
Created attachment 60413
Patch for handling of Standard-14 fonts on Windows
According to the PDF specs, every PDF viewer must be able to handle 14 standard fonts even if they are not embedded.
On Windows, poppler does not supply suitable substitution fonts for Symbol and ZapfDingbats. "symbol.ttf" could be a fallback, but has a quite different appearance from the type1 font Symbol and has some missing/different glyphs (e.g., for the letters 'f', 'j', 'v'). ZapfDingbats is totally different from "wingding.ttf".
GlobalParams::setupBaseFonts(char * dir) [poppler/GlobalParamsWin.cc] does support passing a custom directory to it, but it is only called once without custom directory by GlobalParams::getDisplayFont(GfxFont *font) [poppler/GlobalParamsWin.cc].
In addition, if one manually changes the directory passed to setupBaseFonts(), the application crashes because it tries to access a NULL pointer.
Attached is a series of patches with the following changes:
1) Remove "wingding.ttf" as it's not a suitable replacement. For a testcase pdf file, see http://texworks.googlecode.com/svn/trunk/testcases/base14-fonts.pdf and the expected output at http://texworks.googlecode.com/svn/trunk/testcases/base14-fonts.png)
2) Fix the crash when supplying a custom directory to setupBaseFonts() by checking for valid paths
3) Provide the directory "share/fonts/type1/gsfonts" relative to the library/application to setupBaseFonts() by default. Ideally, this should probably be encapsulated into a configuration variable similar to POPPLER_DATADIR, but I don't know the poppler sources well enough to code that.
~~**Patch 60413**~~, "Patch for handling of Standard-14 fonts on Windows":
[win32-font-handling.patch](/uploads/41cb181edcfdf1340c4982a08be584b3/win32-font-handling.patch)https://gitlab.freedesktop.org/poppler/poppler/-/issues/299save pdf's without password encryption2018-10-05T23:23:34ZBugzilla Migration Usersave pdf's without password encryption## Submitted by Mike Frysinger `@vapier`
Assigned to **poppler-bugs**
**[Link to original bug (#18440)](https://bugs.freedesktop.org/show_bug.cgi?id=18440)**
## Description
i have a pdf that is password encrypted that i can open a...## Submitted by Mike Frysinger `@vapier`
Assigned to **poppler-bugs**
**[Link to original bug (#18440)](https://bugs.freedesktop.org/show_bug.cgi?id=18440)**
## Description
i have a pdf that is password encrypted that i can open and read just fine. however i'd like to now save it without the password encryption. the okular guys indicate that poppler itself needs to support this.