Go to Table of contents, or to the previous page.
Although I have worked hard to make Thessalonica stable, there are some problems which I can’t solve myself, because they depend from bugs present in OpenOffice.org itself. So if you like Thessalonica and want to make it better, please support some bug reports I have created at http://www.openoffice.org (exact references are given below). Detailed instructions for those who are not familiar with Issuezilla can be found at the same site. First, please register (if you have not yet) at http://www.openoffice.org. Registration costs nothing. After registering and logging in go to pages specified below, add your comments or vote for issues. Please remember that only your activity will drew developers’ attention to problems which are really crucial for developing Thessalonica.
See the installation instructions, section Upgrading from earlier versions.
No issues in this block at the present time: all previously described problems had been successfully resolved when OpenOffice.org 2.1 and Thessalonica 3.0 were released.
The algorithm used for conversion is currently much slower than it could be. The problem is that Thessalonica have to test formatting of each text fragment before replacing it. Of course, it could be natural to set formatting attributes for searching text once (exactly as you do in the Search/Replace dialog box) and then perform multiple replacement operations. Thessalonica for Microsoft Word works exactly by this way.
However, searching for formatted text is currently broken in OpenOffice.org. When you use the Search/Replace dialog to find a specific formatting, in most cases it can be found, but sometimes it can’t. However, multiple replacement operations never can be correctly performed, and the most part of text with the specified formatting remains untouched. Of course, it is hard to reproduce this problem, because there are few people using text converters and even less people developing them. That’s why this problem is still not fixed, although I have created issue for it (see http://www.openoffice.org/issues/show_bug.cgi?id=10569) very long ago.
That’s why in early beta versions of Thessalonica I used another algorithm, where text without formatting was searched, and then formatting of all found ocurences was tested. Of course, this worked much slower, but also produced some specific problems, which now (beginning from OpenOffice.org 1.1.0 rc5) can’t be easily resolved. Again, I’ve created an issue for this, see http://www.openoffice.org/issues/show_bug.cgi?id=23552.
So finally I had to completely redesign my converter, which now is based on searching with regular expressions. This algorithm works a bit faster, but it is still similar to the second one, and so I’m unsure that it will never produce similar problems, although right now (with OpenOffice.org 1.1.0) it works almost perfectly. So if you like Thessalonica and want to keep it alive, PLEASE support issues 10569 and 23552.
As it is already explained in the previous answer, converting a large text really takes a long enough time, and, although the conversion process is indicated on the OpenOffice.org status bar, the indicator may stop for several minutes, so it looks like OpenOffice.org is really hanging. However, if you want to monitor the whole conversion process, you may do the following:
change to the directory
user/registry/data/org/openoffice/comp/thessalonica
find the file called Preferences.xcu and open it in a plain text editor;
add the following property to the Converter/Options/ node:
<prop oor:name="LockControllers" oor:type="xs:boolean">
<value>true</value>
</prop>
change true to false;
restart OpenOffice.org.
Of course this will make the whole conversion process even more slower (since updating screen also requires some time), but at least you will be able to see that OpenOffice.org is not hanging, but doing a conversion job for you.
This option is also useful for debugging new sets of conversion rules.
OpenOffice.org may hang or even crash when the user tries to apply Bold or Italic formatting to some fonts which have only regular shape. Of course this is very common situation for Greek or other “exotic” fonts, especially old 8-bit ones. So please be careful when converting your text from Unicode to 8-bit encodings, and always disable Bold and Italic options for 8-bit font you have selected in the Universal Converter options dialog if you know this font family doesn’t support the corresponding shapes.
Go to Table of contents, or to the previous page.