Issue 109954 - PDF Export: Link to www target with special character in domain name becomes link to file on local HDD
Summary: PDF Export: Link to www target with special character in domain name becomes ...
Status: CONFIRMED
Alias: None
Product: General
Classification: Code
Component: code (show other issues)
Version: 3.3.0 or older (OOo)
Hardware: All All
: P3 Normal (vote)
Target Milestone: ---
Assignee: AOO issues mailing list
QA Contact:
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-03-08 19:42 UTC by bbyron
Modified: 2014-05-29 06:14 UTC (History)
3 users (show)

See Also:
Issue Type: DEFECT
Latest Confirmation in: 4.2.0-dev
Developer Difficulty: ---


Attachments
Open Office Doc (4.94 MB, application/vnd.oasis.opendocument.text)
2010-03-08 19:45 UTC, bbyron
no flags Details
PDF (1.72 MB, application/pdf)
2010-03-08 19:46 UTC, bbyron
no flags Details
Test Kit (81.50 KB, application/x-zip-compressed)
2014-05-29 06:01 UTC, Rainer Bielefeld
no flags Details

Note You need to log in before you can comment on or make changes to this issue.
Description bbyron 2010-03-08 19:42:18 UTC
I created the following document in OpenOffice Writer:

https://www.yousendit.com/download/THE3aXRFdkcwZ214dnc9PQ

This document includes multiple links to websites. The first link goes to:
http://affiliate_landmine.s3.amazonaws.com/mp3/the_other_side_of_lisbuilding_wit
h_russell_brunson.html

I set up all links in this document the same and they all work when I 
ctrl+click. 

The problem comes when converting the document to PDF. I clicked File > Export 
as PDF > Export (leaving all PDF settings at default) to get this PDF:

https://www.yousendit.com/download/THE3aXRHcWZGR0d4dnc9PQ

All links remain the same except the first, which has been somehow changed to: 

file:///C:/Users/brent/Desktop/http://affiliate_landmine.s3.amazonaws.com/mp3/th
e_other_side_of_lisbuilding_with_russell_brunson.html (the file location of the 
new PDF has been added to the beginning of the link).

I have tried everything I can think of to get around this with no luck. I have 
also tried linking to other similar URLs, including:

http://affiliate_landmine.s3.amazonaws.com/mp3/all.html

and

http://affiliate_landmine.s3.amazonaws.com/mp3/how_to_thrive_in_a_down_economy_w
ith_russell_brunson.html

These links have the same result of the added file location each time the 
OpenOffice document is converted to PDF. 

I want to emphasize that this change only occurs on the first link in my 
document. The others work fine after the file is converted to PDF. 

I have also tested these same links with other PDF writers and they all work 
perfect every time.
Comment 1 bbyron 2010-03-08 19:45:05 UTC
Created attachment 68228 [details]
Open Office Doc
Comment 2 bbyron 2010-03-08 19:46:19 UTC
Created attachment 68229 [details]
PDF
Comment 3 Rainer Bielefeld 2010-03-09 07:43:34 UTC
I can reproduce the efefct with "01OtherSideOfListbuilding.odt" and with
"Ooo-Dev 3.3.0 multilingual version English UI WIN XP: [DEV300m73 (Build 9456)]"!
I can not reproduce that with self created documents.
When I copy the transparent image covering "Click here to play Audio Book" to an
own document, the problem remains, although links in other images will work fine.

@bbyron:
Please provide a step by step instruction how to create such a defective link
"from the scratch"; please also contribute information concerning all your
relevant settings.
Comment 4 michael.ruess 2010-03-11 14:39:00 UTC
Reproducible with attached sample. The document path is added to the hyperlink
on the graphic.
Comment 5 Rob Weir 2013-07-30 02:17:14 UTC
Reset assignee on issues not touched by assignee in more than 2000 days.
Comment 6 Joe Smith 2014-05-28 21:50:49 UTC
I'm not able to reproduce this testing with AOO 4.1 on Fedora Linux with a new document.

With the attached test document, I see the same results as mentioned in comment 3 above.

If I open the test document 01OtherSideOfListbuilding.odt and edit the link from the first page:
Right-click > Picture > Hyperlink, erase the original URL and type in
http://www.openoffice.org
then export to pdf, I get a working link.

If I go back and paste in the original URL
http://affiliate_landmine.s3.amazonaws.com/mp3/the_other_side_of_lisbuilding_with_russell_brunson.html
then export to pdf again, I get a broken link.

For this link, the pdf contains the code
<</Type/Action/S/URI/URI(./http:%2F%2Faffiliate_landmine.s3.amazonaws.com%2Fmp3%2Fthe_other_side_of_lisbuilding_with_russell_brunson.html)>>

It seems that OO treats this is a local file link, so it modifies the link and quotes the /s.

If I go back again and paste in this link address:
http://affiliate-landmine.s3.amazonaws.com/mp3/the-other-side-of-lisbuilding-with-russell-brunson.html
(all underscores changed to hyphens), then export to pdf, I get a correct link in the pdf.

It seems that the presence of the underscores is causing the problem.

Trying again with http://www.openoffice_test.org leads to this code in the pdf:
<</Type/Action/S/URI/URI(./http:%2F%2Fwww.openoffice_test.org)>>

So, one underscore is enough to cause the damage.
Comment 7 Rainer Bielefeld 2014-05-29 06:01:52 UTC
Created attachment 83480 [details]
Test Kit

Still Reproducible with "AOO 4.2.0-Dev – German UI / German locale [AOO420m1(Build:9800)  -   Rev. 1597166  2014-05-20 1]" on German WIN7 Home Premium (64bit)", “historic” 4. User Profile used for all predecessor versions

In odf documents of attached simplified test kit you will find 3 Links:
* Link 1 has an underscore in html document name: link in PDF export looks
   fine, leads to WWW
* Link 2  has an underscore in domain name: link in PDF export is to 
  local file, not to www :-(
* Link 3 has underscore in domain name replaced by encoded blank: link in 
  PDF export is to local file, not to www   :-(

All links to not existing targets, you can see in Tooltip whether it leads to local file or www
Comment 8 Rainer Bielefeld 2014-05-29 06:10:12 UTC
Additional Info:
(a) As new test kit shows problem is not limited to Writer, same in Calc
(b) Problem seems limited to special characters in domain name
(c) additional to results of Comment 6: also other special characters in
    domain name can cause the problem
(d) an additional test (not included in test kit) with underscore in
    path name (http://de.wikipedia.org/wi_ki/Canon_Pellix.html)
    showed that this does not cause a problem as the reported
(e) Test shows that problem has nothing to do with graphic, also in text links
Comment 9 Rainer Bielefeld 2014-05-29 06:14:19 UTC
(f) Was still ok with OOo 2.0.2
(g) already broken in OOo 3.1.1