Apache OpenOffice (AOO) Bugzilla – Issue 41744
Regression, importing .doc file with multi-column TOC hangs
Last modified: 2017-05-20 11:22:06 UTC
I belive this is similar to bugs 40140, 40141, 40147, 40150, etc. Another column bug. Trying to load the document hangs OOo with near 100% of CPU (on SMP system, 100% of 1 CPU only). Behavior in m74 improved over m71 but still hangs.The hang appears in several documents; the common feature seems to be a table of contents in two column layout. If I remove the table of contents from a document, the document loads fine. The documents load fine in 1.1.3 and in Word 2000. Can't use 1.1.x on these because it messes up the formatting too much. I first saw the problem in m51 (the oldest copy I still have), but it is present in m61, m65, m71, and m74. m51 hangs at 0% on the Updating Index progress bar. m6x and m71 hang before displaying the Importing progress bar. m74 hangs when the Importing progress bar goes away after reaching 100%. On all versions, all UI elements are grayed out. I am trying to create a test document. The problem documents are very large (200+ pages each). The issue doesn't seem to show up in short documents, so I am trying to make the shortest possible that still triggers the issue. I'm guessing that the length of the TOC is the trigger.
please attach the offending document to this issue, so that we can reproduce and fix the problem here. You can also send the document directly (mru@openoffice.org) for the case it contains confidential data. Feel free to re-open the issue when you've done. Thanks for supporting us!
Created attachment 22157 [details] Test file (zipped)
I can't submit the problem document, even directly. This is a test document that was created with Word 2000 and contains a two column table of contents, a section break, and a long series of level 1 headings. There is a page number in the footer. This test document doesn't trigger the hang, but it takes a really long time (minutes) to load. The load time increases faster than the length of the document increases. Load time is barely a second without the table of contents. Testing on slow machine, so load times are easy to see. In Word, it's 38 pages long, and the table of contents is 6 pages. In OOo 1.1.3, it loads fine but loses the hyperlinks. In m74, the table of contents is generated with 2 inside of a 2 column section for a total of 4 columns. The hyperlinks are intact, though. Do I need to file a separate bug for this?
Yeah, I can see. The import of this document needs a long time. But it is no hang; thus I lowered Prio to P3. MRU->MMAHER: please have a look, if we can increase the import performance of this here. Maybe it is not import but updating the TOC...
Just tested Scratch.doc with m77. Behavior is worse. Still takes a really long time to load. The changes: Both CPUs on an SMP system @ 100% load. TOC field is present but blank. Text body is now two columns. Two columns should have stopped after last page of TOC (section break).
I filed bug 43113 on the 4 column TOC problem. As far as the original issue goes, m79 still takes forever to load.
mmaher: reassigning to flr. I am no longer responsible for filters.
Problem still exists in 2.0. Load time seems to increase exponentially with the number of hyperlinks in the document. For long documents (over 200 pages), it might as well be a hang. I've let it sit for over 1/2 hour on a fast computer before killing the process. As a workaround, I can use OOo 1.1.x to import these documents, but the formatting suffers greatly. I also notice that 1.1.x tends to clobber hyperlinks, so that may be the reason for the difference.
Request change in status to CLOSED with resolution FIXED. The documents that triggered this issue load with no problem in 2.0.3. First time I've been able to test in a while. Load times are about a minute.
assigning to hbrinkm
Reset assigne to the default "issues@openoffice.apache.org".