Bug 67583 - FILEOPEN: File with 40 images long opening, not responding and RAM increasing - both DOC, DOCX (win only)
Summary: FILEOPEN: File with 40 images long opening, not responding and RAM increasing...
Status: NEW
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Writer (show other bugs)
Version:
(earliest affected)
Inherited From OOo
Hardware: All All
: medium normal
Assignee: Not Assigned
URL:
Whiteboard: BSA NoRepro:4.2.2.1:OSX Repro:4.3.0.0...
Keywords: filter:doc, filter:docx, haveBacktrace, perf
Depends on:
Blocks: Image-Handling Anchor-and-Text-Wrap DOCX-Opening DOC
  Show dependency treegraph
 
Reported: 2013-07-31 13:09 UTC by Slava
Modified: 2023-04-26 10:34 UTC (History)
6 users (show)

See Also:
Crash report or crash signature:


Attachments
Not open file DOC (2.35 MB, application/msword)
2013-07-31 13:09 UTC, Slava
Details
backtrace (13.98 KB, text/plain)
2015-07-25 19:09 UTC, Gordo
Details
Not open file saved as DOCX also doesn't open (2.17 MB, application/vnd.ms-word.document.12)
2016-02-04 10:07 UTC, Timur
Details
load in 6.4.1.2 of docx in odt (2.14 MB, application/vnd.oasis.opendocument.text)
2020-03-10 15:19 UTC, paulystefan
Details
docx new save in excel 2016 (2.39 MB, application/vnd.openxmlformats-officedocument.wordprocessingml.document)
2020-03-10 15:22 UTC, paulystefan
Details
docx new save in excel 2016 and saved in LO6412 in odt (2.14 MB, application/vnd.oasis.opendocument.text)
2020-03-10 15:26 UTC, paulystefan
Details
doc saved in LO6412 in odt (2.15 MB, application/vnd.oasis.opendocument.text)
2020-03-10 15:36 UTC, paulystefan
Details
Flamegraph (240.57 KB, application/x-bzip)
2020-03-11 20:59 UTC, Julien Nabet
Details
reduced to first page odt of doc (73.00 KB, application/vnd.oasis.opendocument.text)
2020-03-14 23:51 UTC, paulystefan
Details
doc reduced to first page in MSO2016 compatibility mode (117.50 KB, application/msword)
2020-03-15 00:04 UTC, paulystefan
Details
export gif of first page from doc in mso 2016 (49.27 KB, image/gif)
2020-03-15 00:22 UTC, paulystefan
Details
import gif of first page of doc in new odt (61.66 KB, application/vnd.oasis.opendocument.text)
2020-03-15 00:25 UTC, paulystefan
Details
Flamegraph (200.84 KB, application/x-bzip)
2022-09-16 19:04 UTC, Julien Nabet
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Slava 2013-07-31 13:09:23 UTC
Created attachment 83363 [details]
Not open file DOC

Problem description: 

Steps to reproduce:
1. Not open file
2. Libreoffice not responding


Current behavior:

Expected behavior:

              
Operating System: Windows 7
Version: 4.1.0.4 release
Comment 1 Joel Madero 2013-08-01 19:15:49 UTC
Seems like this is another rendition of the image caching problem - 

The file will open but you need a lot of RAM and it chews it up (I have 4 gigs and this file just went through it)

Marking as

New
Critical
High

Updated version to match tracker bug - this was inherited from OOo I believe
Comment 2 tommy27 2013-08-11 17:11:32 UTC
test files freezes either LibO 4.1.0.4 or 3.3.4.
changing version to 3.3.4 (but probably even older releases are affected).
Comment 3 retired 2014-03-25 09:45:30 UTC
NoRepro:4.2.2.1:OSX

While this is a rather insane file with lots of images and weired formatting, LO manages to open it without a hang.

Should the file still not open for you in LO 4.2.2.1 please re-open.
Comment 4 Slava 2014-03-25 10:30:07 UTC
Not open file 
LO 4.2.3.1 Windows 7 x64
Comment 5 V Stuart Foote 2014-03-25 14:04:53 UTC
No this very silly 68 page document of images and cyrillic word art text does not open on 

Version: 4.2.2.1
Build ID: 3be8cda0bddd8e430d8cda1ebfd581265cca5a0f

nor on

Version: 4.3.0.0.alpha0+
Build ID: aeab0183e86fe011d32058864c02b2de4da32dc9
TinderBox: Win-x86@39, Branch:master, Time: 2014-03-24_05:49:26

On same system with MS Office 2007, it does open in Word's 'compatibility mode' but there is considerable shift of the text word art and images as if the page framing is out of sync.

Probably remains a legitimate corner case example for bug 47148
Comment 6 tommy27 2014-05-02 14:32:13 UTC
(In reply to comment #5)
> No this very silly 68 page document of images and cyrillic word art text
> does not open on 
> 
> Version: 4.2.2.1
> Build ID: 3be8cda0bddd8e430d8cda1ebfd581265cca5a0f
> 
> nor on
> 
> Version: 4.3.0.0.alpha0+
> Build ID: aeab0183e86fe011d32058864c02b2de4da32dc9
> TinderBox: Win-x86@39, Branch:master, Time: 2014-03-24_05:49:26

not opens is 4.2.3.3 as well.
I wonder if trying to delete some pages of those 68 in MS Word may help to identify which one causes the bug in order to have a minima test case.

unfortunately I have no MS Word licence to test
Comment 7 tommy27 2014-10-30 09:06:30 UTC
can't still open it in 4.3.2.2 as well (Win7x64)
Comment 8 Gordo 2015-07-25 19:09:28 UTC
Created attachment 117435 [details]
backtrace

Heavens to Betsy!  A backtrace, even.

Windows Vista 64
Version: 5.1.0.0.alpha1+
Build ID: 8cfdd81b70ef37927b40497ffd10034f28335034
TinderBox: Win-x86@39, Branch:master, Time: 2015-07-24_02:47:18
Comment 9 Timur 2016-02-04 10:07:23 UTC
Created attachment 122373 [details]
Not open file saved as DOCX also doesn't open
Comment 10 MM 2016-12-04 20:35:47 UTC
After waiting for some time, the files do open with

Version: 5.2.3.3 (x64)
Build ID: d54a8868f08a7b39642414cf2c8ef2f228f780cf
CPU Threads: 2; OS Version: Windows 6.19; UI Render: default; 
Locale: en-US (en_US); Calc: single

Scrolling / editing is 'a bit slow', but that's another problem I guess.
Comment 11 Justin L 2017-04-04 11:14:36 UTC
With Linux 16.04 it did open for me after a minute, but used 2.7GB of RAM and kept one CPU maxed at 100% even after the document was loaded.
Comment 12 Xavier Van Wijmeersch 2018-07-04 15:26:46 UTC
can open doc file, but it use 4.5 gig off ram and 25% cpu load

Version: 6.2.0.0.alpha0+
Build ID: 3c01b8cc4f15df16b4373855b8797d5dcff59327
CPU threads: 8; OS: Linux 4.14; UI render: default; VCL: kde4; 
Locale: nl-BE (en_US.UTF-8); Calc: group threaded
Comment 13 Julien Nabet 2019-05-28 19:32:28 UTC
On pc Debian x86-64 with master sources updated today, I could reproduce this.

I noticed these logs:
warn:legacy.osl:30696:30696:oox/source/helper/graphichelper.cxx:120: GraphicHelper::GraphicHelper - cannot get target frame
warn:legacy.osl:30696:30696:sw/source/filter/ww8/writerhelper.cxx:586: No NodeIndex in SwFrameFormat ?, suspicious
Comment 14 paulystefan 2020-03-10 15:17:57 UTC
with LO 6.4.1.2 x64 win 10 x64
 
test docx works, but the file has some failures.

saving Docx again in excel 2016 the file is more actual in docx-format.

new opening in LO 6.4.1.2 x64 win 10 x64

the file is again opening and the pictures and text in good shapes in relation to excel 2016.
Comment 15 paulystefan 2020-03-10 15:19:56 UTC
Created attachment 158561 [details]
load in 6.4.1.2 of docx in odt

LO 6.4.1.2 x64 win 10 x64 

save docx file in odt
Comment 16 paulystefan 2020-03-10 15:22:26 UTC
Created attachment 158562 [details]
docx new save in excel 2016

MS Excel 2016 loads the file with message old format.

so i save in excel 2016 format.
Comment 17 paulystefan 2020-03-10 15:26:19 UTC
Created attachment 158563 [details]
docx new save in excel 2016 and saved in LO6412 in odt

docx new save in excel 2016 and saved in LO6412 in odt

better quality of import.

problem of loading also solved.

better handling of Microsoft Office files.

Actual save with actual MS Office solve many problems in Libre Office for Import.
Comment 18 paulystefan 2020-03-10 15:36:57 UTC
Created attachment 158564 [details]
doc saved in LO6412 in odt

doc load and saved in LO6412 in odt

load problem solved

quality of pictures and fonts like excel 2016.
Comment 19 Slava 2020-03-11 07:05:03 UTC
Version: 6.4.2.1 x64
used 4 GB of RAM
Comment 20 Timur 2020-03-11 08:21:40 UTC
paulystefan@web.de: I cannot understand what you did with saving in MSO and as ODT, but that's not the point of this bug.
This is simple Fileopen bug where LO hardly opens original DOC or DOCX (saved in MSO from DOC without change), for a long time and a lot of RAM (I closed at 2,5 GB and counting). 
MSO 2016 opens 68 pages with 140 MB.
Repro 7.0+. I set New again.
Comment 21 Julien Nabet 2020-03-11 20:59:10 UTC
Created attachment 158625 [details]
Flamegraph

Here's a Flamegraph retrieved on pc Debian x86-64 with master sources updated today.
Comment 22 paulystefan 2020-03-14 01:01:48 UTC
LO 6.4.1.2 
i load original files in about 45 sec

doc RAM needed 4.3 GB
docx RAM needed 3.5 GB

but no load crash like first message

so you are right,
there is a ram problem.
Comment 23 paulystefan 2020-03-14 23:51:23 UTC
Created attachment 158681 [details]
reduced to first page odt of doc

LO 6.4.1.2 x64 win 10-64  odt of doc is also with the problem inside.

full odt downloaded before in 158561 needs about 4000 MB Ram or more.

first page here needs big 291 MB Ram. 

full document is with 68 pages.

so the problem is now reduced for more analysis in detail.
Comment 24 paulystefan 2020-03-15 00:04:25 UTC
Created attachment 158682 [details]
doc reduced to first page in MSO2016 compatibility mode

reduced to first page of original doc in mso 2016.

also 291 MB RAM in LO 6.4.1.2 x64 in win-10

so the problem is reduced for more analysis in detail
Comment 25 paulystefan 2020-03-15 00:22:33 UTC
Created attachment 158683 [details]
export gif of first page from doc in mso 2016

gif animation is running in LO 6.4.1.2 in doc and odt
Comment 26 paulystefan 2020-03-15 00:25:34 UTC
Created attachment 158684 [details]
import gif of first page of doc in new odt

LO 6.4.1.2 needs more than 220 MB Ram 

gif-animation runs 

only marked the animation is static frozen
Comment 27 paulystefan 2020-03-15 00:32:08 UTC Comment hidden (obsolete)
Comment 28 paulystefan 2020-03-15 00:33:16 UTC
correction

see bug 104878 

big gif animation is slow in impress

perhaps same problem
Comment 29 paulystefan 2020-03-15 22:10:44 UTC
in mso word a gif with animation ist not permanent running.

it only shows the first picture.
Comment 30 Telesto 2020-05-24 19:29:05 UTC
A horrible bug doc.. I would guess an anchoring problem. mixed with drawing polygons
Comment 31 paulystefan 2021-05-17 13:03:41 UTC
my workarounds:
1.save as DOCX in MSO 2016 or higher

2. option change to no animation in LO 


But there are only 67 pages in LO 7132 with much difference to DOCX and DOC.

Perhaps there are some improvements in LO 7.2.0 with the better importer.
Comment 32 paulystefan 2021-08-05 21:06:30 UTC
See no signifikant improvement in LO 7202

Version: 7.2.0.2 (x64) / LibreOffice Community
Build ID: 614be4f5c67816389257027dc5e56c801a547089
CPU threads: 8; OS: Windows 10.0 Build 19043; UI render: Skia/Raster; VCL: win
Locale: de-DE (de_DE); UI: de-DE
Calc: CL

A partial workaround: new saving in MSO201x.
Comment 33 Telesto 2021-09-21 17:29:11 UTC
Does open.. however with some delay - 25 seconds or so
Version: 7.3.0.0.alpha0+ (x64) / LibreOffice Community
Build ID: d5e55d204b71710eb5eb5d2c683dd6698626df3c
CPU threads: 4; OS: Windows 6.3 Build 9600; UI render: Skia/Raster; VCL: win
Locale: nl-NL (nl_NL); UI: en-US
Calc: CL
Comment 34 paulystefan 2022-06-16 13:19:41 UTC
Without animation option


Does open in 

Version: 7.3.4.2 (x64) / LibreOffice Community
Build ID: 728fec16bd5f605073805c3c9e7c4212a0120dc5
CPU threads: 8; OS: Windows 10.0 Build 19044; UI render: Skia/Raster; VCL: win
Locale: de-DE (de_DE); UI: de-DE
Calc: CL
Comment 35 Julien Nabet 2022-09-15 19:20:13 UTC
I gave a new try with master sources updated today on initial doc, it's quite quick to open, idem with LO Debian package 7.4.1

Could someone give a try with 7.4.1 or a daily master build?
Comment 36 Julien Nabet 2022-09-16 07:14:59 UTC
Ilmari: would you have some time to give it a try?
Comment 37 Buovjaga 2022-09-16 07:22:56 UTC
(In reply to Julien Nabet from comment #36)
> Ilmari: would you have some time to give it a try?

Testing with attachment 83363 [details], stopwatch time to responsive is about 25 secs for me, still takes some processing to be *fully* responsive after that. So seems to be the same results as in 2021. Not sure what the goal would be, but let's lower the severity.

Version: 7.5.0.0.alpha0+ (x64) / LibreOffice Community
Build ID: 922b79a0f5a9151a6870ba395abcac5b54055275
CPU threads: 2; OS: Windows 10.0 Build 19044; UI render: default; VCL: win
Locale: fi-FI (fi_FI); UI: en-US
Calc: threaded Jumbo
Comment 38 Julien Nabet 2022-09-16 19:04:52 UTC
Created attachment 182503 [details]
Flamegraph

On pc Debian x86-64 with master sources updated today + gen rendering, I retrieved a Flamegraph during opening of https://bugs.documentfoundation.org/attachment.cgi?id=83363, if it can help.
Comment 39 Roman Kuznetsov 2023-04-26 10:34:17 UTC
(In reply to Buovjaga from comment #37)
> Testing with attachment 83363 [details], stopwatch time to responsive is
> about 25 secs for me, still takes some processing to be *fully* responsive
> after that. So seems to be the same results as in 2021. Not sure what the
> goal would be, but let's lower the severity.
> 
> Version: 7.5.0.0.alpha0+ (x64) / LibreOffice Community
> Build ID: 922b79a0f5a9151a6870ba395abcac5b54055275
> CPU threads: 2; OS: Windows 10.0 Build 19044; UI render: default; VCL: win
> Locale: fi-FI (fi_FI); UI: en-US
> Calc: threaded Jumbo

Still the same in

Version: 7.6.0.0.alpha0+ (X86_64) / LibreOffice Community
Build ID: 0ee9501c0b7dc1a291715fff9c1934b1c08cb654
CPU threads: 8; OS: Windows 10.0 Build 19043; UI render: Skia/Vulkan; VCL: win
Locale: ru-RU (ru_RU); UI: ru-RU
Calc: threaded