Apache OpenOffice (AOO) Bugzilla – Issue 9539
Open Office hangs up giving 100% CPU utilization when it tries to open a word document containing one large string of integers..
Last modified: 2013-08-07 14:43:23 UTC
Summary Open Office hangs up giving 100% CPU utilization when it tries to open a word document containing one large string of integers.. System Configuration Platform: PC OS: WinXP, NT 5.01 Version: 2600 Steps to replicate bug 1. Open this word document using File->Open 2. Application hangs up trying to open this file Notes I am also attaching the file used. The reason I feel that this is an important bug to fix is because the application does not handle an error like this gracefully. The application should not hang up while opening the document. This is because the user can have documents which are large in size and would least expect the application to fail at this point of time. Besides the file used for this test case is just 401 kb.
Created attachment 3764 [details] word file used for testing
Replicated on Platform: PC OS: Win2000, XP Version: 2195, 2600 Notes: Sometimes the word file above does not open properly in word. Word sometimes displays the text as a picture. Therefore in order to see the file you may have to open the file in Wordpad. The application crashes both with the original steps and also when the Open With... command (choosing OpenOffice 643c) is used to open the file. When the file is opened using OpenOffice 1.0.1 however, the application does not crash. The application is able to indicate that it cannot open the file and closes the application down. I made modifications to the file provided above, by breaking the text down into more managable portions, instead of just one long string. However the program still crashes. Although I don't know if it is common for anyone to produce a file with just one long text string, many large text strings are more common. I hope that this matter will be looked into as soon as possible.
Sujit and Bharath, thank you for using and supporting OOo. For formalities sake, duplicated on Win2k Server, OOo 643C. Another bounds check from the team. Expect more to come.
Changing to new and assigning to CMC.
cmc->aidan: This is one for you :-), ww8par.cxx ReadText adds text to a string. It needs to handle when the case of the txtnode containing STRING_LEN better. If the len of the string hits string_len it needs to force the parser to act as if 0xd a para break was encountered so as to read in huge text like this as seperate paragraphs for each maximum len chunk.
Aidan: This has been fixed for the next release. The initial problem is to do with the paragraphs being greater than 65K in size. There is still a problem with the test document that Caolan thinks may be due to the lack of spaces in the document. Inserting a space after every 10 pages or so, makes the document load eventually. See second attachment.
Created attachment 4506 [details] The original attachment with spaces inserted after every 10 pages
Aidan: Reopening for assignment to QA
Aidan->Mru: Can you verify this. The MS Word importer can now handle docs with paragraphs greater than 65K in size. There is still a problem with performance (possibly layout), but this cannot be fixed in the ww8 filter.
JA: re-prioritized according to new priority guide lines. JA->MRU: please check this. I changed the priority to P2 because this scenario describes a hangup...
Looks fixed with internal Milestone 644m4s4.
Closed. Fix will be available with OpenOffice 1.1 Beta (will be released soon).