Bug 131381 - docx rendering problems
Summary: docx rendering problems
Status: RESOLVED INVALID
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Writer (show other bugs)
Version:
(earliest affected)
6.3.4.2 release
Hardware: All All
: medium normal
Assignee: Not Assigned
URL:
Whiteboard:
Keywords: filter:docx
Depends on:
Blocks:
 
Reported: 2020-03-16 19:11 UTC by msdobrescu@gmail.com
Modified: 2020-10-13 10:46 UTC (History)
1 user (show)

See Also:
Crash report or crash signature:


Attachments
samples (documents 1 to 3) (4.39 MB, application/x-7z-compressed)
2020-03-16 19:11 UTC, msdobrescu@gmail.com
Details
Sample DOCX 1 (291.39 KB, application/vnd.openxmlformats-officedocument.wordprocessingml.document)
2020-09-15 06:02 UTC, Timur
Details
Sample DOCX 1 compared (589.54 KB, image/png)
2020-09-15 06:04 UTC, Timur
Details
Sample DOCX 2 (893.84 KB, application/vnd.openxmlformats-officedocument.wordprocessingml.document)
2020-09-15 07:36 UTC, Timur
Details
Sample DOCX 3 (296.42 KB, application/vnd.openxmlformats-officedocument.wordprocessingml.document)
2020-09-15 07:44 UTC, Timur
Details
Sample DOCX 3 compared (244.91 KB, image/png)
2020-09-15 07:47 UTC, Timur
Details
Document 1 compared (1.16 MB, image/png)
2020-10-13 07:37 UTC, msdobrescu@gmail.com
Details
Second document in LO 6.4 (496.51 KB, image/png)
2020-10-13 07:56 UTC, msdobrescu@gmail.com
Details
Third document in LO (254.36 KB, image/png)
2020-10-13 07:59 UTC, msdobrescu@gmail.com
Details
Document #4 heavily scrambled (1.07 MB, application/vnd.openxmlformats-officedocument.wordprocessingml.document)
2020-10-13 08:09 UTC, msdobrescu@gmail.com
Details
Document #4 in LO (506.18 KB, image/png)
2020-10-13 08:10 UTC, msdobrescu@gmail.com
Details
Document #4 exported from MSO as PDF (411.12 KB, application/pdf)
2020-10-13 08:11 UTC, msdobrescu@gmail.com
Details
Document #4 compared MSO LO (346.50 KB, image/png)
2020-10-13 09:00 UTC, Timur
Details

Note You need to log in before you can comment on or make changes to this bug.
Description msdobrescu@gmail.com 2020-03-16 19:11:57 UTC
Created attachment 158728 [details]
samples (documents 1 to 3)

Hello, I have some docx files that look scrambled in LibreOffice Wirter (6.3.4.2.0+ under Sabayon Linux and Windows 10) by comparison to the original Microsoft Word. This applies from characters to elements positioning.

I have attached three documents, each:
- in original form as .docx
- exported from Microsoft Word as .pdf
- screenshot as they appear in this version

Sample 1. The most messed up to illegibility. I can't even guess what should look like, I can't select the rounded areas/drawings/images. The first line of the title seems to have the characters replaced in other codepage.

Sample 2. An area moved from the page 3 to page 2, under some images.

Sample 3. There are mathematical operations. In our school convention, when disposed vertically, after the operands a horizontal line is drawn. In this case seems to strike some operands. Also rectangles appear around some texts.
Comment 1 Timur 2020-03-18 07:12:52 UTC
Bugzilla is not "document based", like "this document doesn't display nice". 
Bugzilla is "issue based", so a single issue must be pointed at, after a search for not being a duplicate. 
It's highly unlikely that any bugs of type "multiple problems with this file/bad rendering/this should look like", will be fixed. 
Each issue (section break, paragraph break, text box size, picture position...) should be analyzed and checked for already reported bugs.
If bugs don't exist, they should be reported separately, even if they happen with the same file. 
Example file should be reduced to minimum test case for a specific problem, with clear file history or steps to reproduce from scratch.
Also, no zipped examples, just single file, please remove 7z and attach.
And give exact description of a single bug, after you searched for existing ones.
Comment 2 QA Administrators 2020-09-15 03:59:47 UTC Comment hidden (obsolete)
Comment 3 Timur 2020-09-15 06:02:20 UTC
Created attachment 165503 [details]
Sample DOCX 1

These are MSO created DOCX. DOCX 1 is old format, but similar if resaved in MSO.
Comment 4 Timur 2020-09-15 06:04:29 UTC
Created attachment 165504 [details]
Sample DOCX 1 compared

(In reply to msdobrescu@gmail.com from comment #0)
> Sample 1. The most messed up to illegibility. I can't even guess what should
> look like, I can't select the rounded areas/drawings/images. The first line
> of the title seems to have the characters replaced in other codepage.

Wrong read of EMF image in Sample 1. 

This does not mean bug should be confirmed, should be searched.
Comment 5 Timur 2020-09-15 07:19:40 UTC
(In reply to Timur from comment #4)
> Created attachment 165504 [details]
> Sample DOCX 1 compared
> Wrong read of EMF image in Sample 1. 

Didn't open in LO 3.5. Similar from 3.6 to 7.1+. 
There are more EMF bugs and question is if this is a duplicate.
Comment 6 Timur 2020-09-15 07:36:47 UTC
Created attachment 165506 [details]
Sample DOCX 2

No repro for DOCX 2 (original older and newer resaved in MSO). 
LO 7.1+ and earlier open 3 pages as MSO (unlike reporter's screenshot with pages 2 and 3 overlapping). 
There shouldn't be difference Win/Lin, I tested both.
Comment 7 Timur 2020-09-15 07:44:56 UTC
Created attachment 165507 [details]
Sample DOCX 3

2 pages DOCX, of which reporter attached screenshot of page 2. 
No repro, both pages fine. 

I even tested DOCX 2 and DOCX 3 in LO 6.3.0 with and without OpenGL, no repro, not clear how this was reported.
Comment 8 Timur 2020-09-15 07:45:40 UTC
Finally, I mark as Invalid report.
Comment 9 Timur 2020-09-15 07:47:25 UTC
Created attachment 165508 [details]
Sample DOCX 3 compared
Comment 10 msdobrescu@gmail.com 2020-10-13 07:37:54 UTC
Created attachment 166332 [details]
Document 1 compared

The images are broken, but the overall document is messed up, there are lines, ellipses, text misplaced, or of different size. How to determine the reason?
Comment 11 msdobrescu@gmail.com 2020-10-13 07:38:24 UTC
Hi, I was sick, now I come back and slowly take care of the issues again. 

It still reproduces for me on:

Version: 6.4.6.2
Build ID: Gentoo official package
CPU threads: 8; OS: Linux 5.7; UI render: default; VCL: gtk3; 
Locale: ro-RO (ro_RO.UTF-8); UI-Language: en-US
Calc: threaded
Comment 12 msdobrescu@gmail.com 2020-10-13 07:56:20 UTC
Created attachment 166333 [details]
Second document in LO 6.4

The second document, besides the bubbles in MSO vs rectangles in LO, which is not that problematic, the text area goes under the images.
Comment 13 msdobrescu@gmail.com 2020-10-13 07:59:15 UTC
Created attachment 166334 [details]
Third document in LO

In the third document, I guess the problem is the font size.
Comment 14 msdobrescu@gmail.com 2020-10-13 08:09:41 UTC
Created attachment 166335 [details]
Document #4 heavily scrambled

Here, the drawing seems to be the issue. This is impossible to figure out how it should look. MSO exports it in odf, also messed up even when reopening it in MSO.
Comment 15 msdobrescu@gmail.com 2020-10-13 08:10:38 UTC
Created attachment 166336 [details]
Document #4 in LO

Rendering of the fourth document in LO
Comment 16 msdobrescu@gmail.com 2020-10-13 08:11:58 UTC
Created attachment 166337 [details]
Document #4 exported from MSO as PDF

For reference, if no MSO available.
Comment 17 msdobrescu@gmail.com 2020-10-13 08:13:15 UTC
Reopened, for review of the new samples.
Comment 18 Timur 2020-10-13 08:30:16 UTC
You didn't read or understand Comment 1. 
There are rules in bug reporting https://wiki.documentfoundation.org/QA/BugReport and this is Invalid report. 
Surely is status Reopened wrong, https://bugs.documentfoundation.org/page.cgi?id=fields.html#bug_status.

To determine bug status, relevant is daily master from https://dev-builds.libreoffice.org/daily/master/current.html .
So after I showed Document 1 with LO 7.1+, it's wrong to say "it's not good with 6.4". 

Since you obviously didn't follow advice to search for EMF problem with Sample DOCX 1, I'll open a new bug on that.
Comment 19 Timur 2020-10-13 08:43:55 UTC
I opened a bug for Sample 1 and wrote there that we see it differently. 
Still, install daily master to check.
Comment 20 msdobrescu@gmail.com 2020-10-13 08:57:09 UTC
Hello,

I don't accept your tone, so I ignore such messages.
Sorry for not being clear to me the reason of a problem.
I advice you to have more patience and accept the fact that other people have a different knowledge and can't understand the reasons from your technical perspective.
I understand that you believe I have created those, but I didn't, I had no idea of the EMF existence in these documents at all.
On other hand, I can't build LO, as developer I would expect not to from my users, I would rather advice to test in a provided set of binaries.
As developer, I would take a report and break it in several reports, if necessary, put the things in order and plan them, refer them in the old invalid report, as the reporter might not see the things as yo do.
I have tried some AppImages, I see Fresh is the 7.0.x, not 7.1.
I confirm the first document is still the same, probably due to the EMF.
The second ones are solved.
The new one is still broken.

Have a nice day.
Comment 21 Timur 2020-10-13 09:00:30 UTC
Created attachment 166340 [details]
Document #4 compared MSO LO

Here is Document #4 compared in MSO 2016 and LO 7.1+.
Graphics is good. It looks wrong because of manual enters to move 2nd graphics to 2nd page. Rendering of size or Times New Roman font is not exactly the same and it looks wrong. Simple enter corrects that. 
Not sure if that would be a bug. Because we have many bugs on those very small differences. Solution is to use page break. 
So I will not report a separate bug. 

Another general problem is that reporter renders all differently so it looks wrong to him. 
Can't say why. General advice in https://bugs.documentfoundation.org/enter_bug.cgi - which he manages to skip in bug report -  is to reset/delete user profile. 
Another advice is one I gave: install daily master and see all those docs again. 
It will install separately to working LO. 

Again, do not reopen this bug. So far, track bug 137440 to see how others will see it. Write there how you see it in daily master.
Comment 22 msdobrescu@gmail.com 2020-10-13 10:46:13 UTC
Clearing the user config did not change anything. Maybe Sabayon uses some old library in some regard to this.