Bug 91418 - FILEOPEN DOC anchored images/frames not visible because paragraph char hidden = true
Summary: FILEOPEN DOC anchored images/frames not visible because paragraph char hidden...
Status: NEW
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Writer (show other bugs)
Version:
(earliest affected)
4.1.0.4 release
Hardware: All All
: low normal
Assignee: Not Assigned
URL:
Whiteboard:
Keywords: bibisected, bisected, filter:doc, regression
Depends on:
Blocks: DOC
  Show dependency treegraph
 
Reported: 2015-05-20 23:15 UTC by h8j18k24p4s2
Modified: 2024-05-27 18:36 UTC (History)
7 users (show)

See Also:
Crash report or crash signature:


Attachments
Page with disappearing text boxes and image (45.00 KB, application/msword)
2015-05-21 18:09 UTC, h8j18k24p4s2
Details
bt from a breakpoint (11.74 KB, text/plain)
2015-05-21 21:04 UTC, Julien Nabet
Details

Note You need to log in before you can comment on or make changes to this bug.
Description h8j18k24p4s2 2015-05-20 23:15:31 UTC
This is really getting out of hand. I had another document saved as a .doc file and have been working on it for several hours today. This document also had text boxes and a photo. I saved the document one more time and closed it. Then I checked my e-mail where I saw the "resolution" of the last bug I filed, which said I should save any future documents as .odt files. With the intention of doing just that, I reopened the document immediately, and the whole stinking thing was blank! Not one thing was on the page. Every text box was gone, as was the photo. I didn't have a chance to save it as an .odt file!

WTF is going on with LibreOffice that things keep disappearing? These two documents represent DAYS of work...gone.
Comment 1 MM 2015-05-20 23:56:38 UTC Comment hidden (obsolete)
Comment 2 Adolfo Jayme Barrientos 2015-05-21 01:59:41 UTC Comment hidden (obsolete)
Comment 3 h8j18k24p4s2 2015-05-21 18:09:04 UTC
Created attachment 115782 [details]
Page with disappearing text boxes and image
Comment 4 h8j18k24p4s2 2015-05-21 18:09:23 UTC
Happy to attach it, but I'm sure it won't help - it's a .doc file, as I said, and it's blank. In addition to everything disappearing, it lost the second page (which didn't have anything written on it, but it was there) as well between saving, closing, and opening it again.

Additional note: I tried just now to start working on the same document again, as the page was blank, so I thought I'd start over. I added a text box in the upper left hand corner, typed in text, made it bold and centered it, then tried to move the text box slightly. It completely disappeared. The "Edit" function (undo text box move, I think it said) seemed to think the text box was there somewhere, but nothing reappeared. This was without saving and closing the document.
Comment 5 h8j18k24p4s2 2015-05-21 20:37:23 UTC
New twist: I downloaded OpenOffice and re-saved the version of the document I'd uploaded to Bugzilla as an .odt file. When I opened it in OpenOffice, all the text reappeared! I added a couple of things then looked at it in Page Preview and - nothing. Blank page. Exported it as a .pdf - blank page. Closed it and re-opened the same .odt document in LibreOffice - blank page. Closed and re-opened again in OpenOffice - all there, except not in Page Preview. Took screen shots out of curiosity, pasted them into a new OpenOffice .odt document - shows up just fine, including in Page Preview and when exported as a .pdf. Opened the .odt screenshot in LibreOffice, and all it showed was two boxes with a note in the upper left corners: "graphics1" and "graphics2."

So for the .odt document, LibreOffice shows nothing, OpenOffice shows all text except in Page Preview. Any ideas are welcome, and thanks.
Comment 6 Julien Nabet 2015-05-21 21:03:48 UTC
On pc Debian x86-64 with master sources updated today, I could reproduce this.
I noticed these 2 logs on console:
warn:sot:5471:1:sot/source/sdstor/stgdir.cxx:425: Trying to resize readonly stream by seeking, could be a wrong offset: 7733
warn:sot:5471:1:sot/source/sdstor/stgdir.cxx:425: Trying to resize readonly stream by seeking, could be a wrong offset: 7733
Comment 7 Julien Nabet 2015-05-21 21:04:45 UTC
Created attachment 115784 [details]
bt from a breakpoint
Comment 8 Julien Nabet 2015-05-21 21:08:58 UTC
Michael: sorry to cc you since I suppose you must work on VclPtr related bugs but could you confirm that it's indeed related to tdf#47148 or did I just misunderstand the real problem? (I attached console logs and bt)
Bonus question: do you know if something is planned about tdf#47148 (which seems more like a metabug)?
Comment 9 MM 2015-05-22 12:29:24 UTC
In tdf#47148 the problem was that there are missing images. In this one, everything is gone, including frames and text.

I converted the doc to docx with an onlineconverter. Then I could see most of the frames and text again (bit misplaced but...). So it is actually saved, but fails on loading or so.
Comment 10 Michael Meeks 2015-05-22 13:08:04 UTC
odd; if this is a regression - I guess it could be some file-format import crasher fix that causes an exception to be thrown on bad data (if it previously worked) which (I conjecture) stops the filter from continuing its work.

That's not helpful I guess, but I'd not mark it as an image management shambols, but more of an import issue.
HTH.
Comment 11 MM 2015-05-22 23:25:33 UTC
(In reply to Michael Meeks from comment #10)
> odd; if this is a regression - I guess it could be some file-format import
> crasher fix that causes an exception to be thrown on bad data (if it
> previously worked) which (I conjecture) stops the filter from continuing its
> work.

Well, I don't think it is a regression. LO 3.3.4 couldn't read it correctly either. Already put the earliest affected version to 'inherited from Ooo' but someone put it back. 

> That's not helpful I guess, but I'd not mark it as an image management
> shambols, but more of an import issue.
> HTH.

Tried to open the file with wordviewer under win 10 and had no problem opening/reading the file. So yes, it could be an import issue.
Comment 12 Telesto 2016-12-06 13:49:05 UTC
Still reproducible with:
Version: 5.4.0.0.alpha0+
Build ID: 33f5bc54aaa7fe7aa9335726e30f9c349155e04d
CPU Threads: 4; OS Version: Windows 6.2; UI Render: default; 
TinderBox: Win-x86@62-merge-TDF, Branch:MASTER, Time: 2016-12-01_23:21:05
Locale: nl-NL (nl_NL); Calc: CL

File does show an image with:
Versie 4.0.0.3 (Bouw-id: 7545bee9c2a0782548772a21bc84a9dcc583b89)
Comment 13 QA Administrators 2017-12-10 16:42:10 UTC Comment hidden (obsolete, spam)
Comment 14 Telesto 2018-06-25 18:50:20 UTC Comment hidden (obsolete)
Comment 15 Richard Chen 2018-11-29 08:57:13 UTC
Still reproducible with this version:

Version: 6.3.0.0.alpha0+ (x64)
Build ID: 0f25a3c36f27fd51453b9a9115f236b83c143684
CPU threads: 8; OS: Windows 10.0; UI render: GL; VCL: win; 
TinderBox: Win-x86_64@42, Branch:master, Time: 2018-11-27_20:06:55
Locale: zh-TW (zh_TW); UI-Language: en-US
Calc: threaded
Comment 16 Rod 2019-11-23 15:03:08 UTC Comment hidden (off-topic)
Comment 17 Rod 2019-11-24 15:11:13 UTC Comment hidden (off-topic)
Comment 18 Telesto 2020-05-24 10:02:25 UTC
Also in
4.1

but not in
LibreOffice 3.5.7.2 
Build ID: 3215f89-f603614-ab984f2-7348103-1225a5b
Comment 19 Buovjaga 2020-06-14 19:00:10 UTC
Bibisected with Linux 41max to https://git.libreoffice.org/core/+/83ba821c10392c08334f7d8d3775fe3e8d08f8fd%5E!/
Fix #i120928: Import Graphic Bullets of MS Word Document
Comment 20 Timur 2020-09-04 10:05:02 UTC
Normally we need ODT source from LO or DOC/X from MSO. 
This is DOC from LO and usually not valid source, nor we can test filesave bug that made it, but since it opens in MSO, we may call it now fileopen bug with low priority.
Comment 21 QA Administrators 2022-09-05 03:38:14 UTC Comment hidden (obsolete)
Comment 22 Timur 2022-09-05 10:30:55 UTC
Repro 7.5+.
Comment 23 Justin L 2023-05-25 01:17:12 UTC
In a totally bizarre state, I can open Arcmia Product Page Layout.doc perfectly fine on one of my development builds, but the page is totally empty on another build. It is also still empty in bibisect-linux-7.6.
Comment 24 Justin L 2023-08-21 22:14:17 UTC
(In reply to Justin L from comment #23)
> In a totally bizarre state, I can open Arcmia Product Page Layout.doc
> perfectly fine on one of my development builds
That is because when you:
-turn on View - Formatting marks, and
-tools - options - Writer - Formatting Aids - Display Formatting: Hidden Characters
then you can see the hidden content. And since everything is in frames that are anchored to a hidden paragraph, the page is empty.
Comment 25 Justin L 2023-08-21 23:10:39 UTC
already marked as hidden in OOo 3.3.

It doesn't look like a misplaced hidden marker - it comes from an sprmCFVanish toggling hidden on in direct formatting paragraph properties.
<sprm value="0x83c" ispmd="0x3c" fSpec="0x0" sgc="character" spra="0" name="sprmCFVanish" operandSize="1" operand="0x81"/>  #0x81 == 129 contrary to style
Comment 26 Justin L 2024-05-27 18:36:19 UTC
repro 24.8+ - the hidden attribute still hides all anchored frames/images.