The GPL is not compatible in any way, shape or form with Confluence (or any other commercial environment for that matter). Could you release under BSD or ASL?

Thanks VERY much for this tool. I've verified that my mysql username/password/db info is correct but am getting this error (I created a separate read-only user).

[root@theta mw2cf-translator-0.0.1]# ant exec
Buildfile: build.xml

exec:
     [java] Executing mediawiki to confluence conversion utility
     [java] com.mysql.jdbc.Driver
     [java] Connect failed bailing..

BUILD SUCCESSFUL
Total time: 1 second

Any thoughts?

I will be releasing a new version of this tool this weekend.
The new version includes

• Parsing of all media content from mediawiki (actual files) to local directories ready for attachement to confluence pages
• Various Bugfixes
• Full source disclosure under ASL

SWEET.

Nice tool!
I have some problems with German umlauts. The file names are written in UTF-8 but the text files are not UTF-8 encoded . I can convert the filenames from UTF-8 to Latin-1 using recode. But I can not convert the content with the same tool. Are you using different encodings?

Couple things I noticed:

1:  had to change the value of the distversion property in build.xml to 0.0.2

2:   if the output directory specified in confluence.properties does not exist, the directory is not created, no output is saved, and no error is thrown.  Either create the directory or throw an error that it doesn't exist.   

All I apologise for the lack of stated release over the weekend, had two quite serious issues pop-up in my life outside programming which couldn't wait and had to be given priority.

Now that corrected I should be able to put the latest build up tonight on the intient svn server after I finish work today (hopefully with the above UTF-8 correction).

I have put 0.0.3 up in the attachments section for testing for you all until I get the svn server sorted out later this evening. [|^mw2cf-translator-0.0.3.tar.gz]

Corrections:

• New config option to specify encoding of output file - req by Christian Vollrath
• Distversion corrected in build.xml - Carlton Brown
• Parsing of all media content from mediawiki (actual files) to local directories ready for attachement to confluence pages - this is something which I think is pretty cool
• Various minor bug fixes.

I have been onto atlassian today about a licence to continue development, see what they come back to me with, my hope is if they grant it I will be able to integrate some of the webservices API to make the media and page import fully dynamic from end to end.

Great translator Brad, thanks.
Any idea when the source will be available ?
And has anyone got any experience or thoughts on how to bulk load the converted pages ?  I have about a 500 page site with lots of embedded images that doesn't seem like a good time re-attaching all those images.

Thanks for providing this tool. I have a site with 54 pages and many many images that I would like to transfer to Confluence.  I was able to translate the pages in a matter of seconds. Still have to read more of the threads above and figure out how to get images over. Also have to clean up some odd characters (ie., þÿ) that appear throughout the pages. And there's lot of html that appears. Perhaps I have to enable the html macro so that it is rendered as html and not text. 

This is a great addition, I'm wondering though if Confluence could

1) understand mediawiki syntax by default

2) have a mediawiki mode

(both of these could be incomplete) 

Mediawiki uses [[ and ]], puts images inside [[Image:...]] and uses ''' for bold text.   It seems like most of the basic mediawiki syntax is unique enough that it shouldn't trip up confluence software.   Mediawiki does have a couple of nice features -- the automatic table of contents and the automatic "floating" images.   These would require a little more fidgeting to get right but could increase confluence usability dramatically -- the Confluence-adept user base would grow by 100s of thousands overnight.

It seems that 2) should be easy to implement, just plug this translator into the Confluence editor and add a tab labeled "Wikipedia-style markup" that invokes the translator when saving.

I just installed the converter, configued the two properties files and ran ant and got:

 -bash-2.05b$ ant exec
Buildfile: build.xml

exec:
     [java] Exception in thread "main" java.lang.NoClassDefFoundError: while resolving class: com.intient.wiki.translator.Translator
     [java]    at java.lang.ClassLoader.resolveClass0(java.lang.Class) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at java.lang.Class.initializeClass() (/lib/ssa/libgcj.so.4.0.0)
     [java]    at java.lang.Class.forName(java.lang.String, boolean, java.lang.ClassLoader) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at java.lang.Class.forName(java.lang.String) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at gnu.gcj.runtime.FirstThread.run() (/lib/ssa/libgcj.so.4.0.0)
     [java]    at _Jv_ThreadRun(java.lang.Thread) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at _Jv_RunMain(java.lang.Class, byte const, int, byte const, boolean) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at __libc_start_main (/lib/tls/libc-2.3.2.so)
     [java] Caused by: java.lang.ClassNotFoundException: java.lang.StringBuilder not found in [file:/home/plaverty/test/mw2cf-translator-0.0.3/dist/mw2cf-translator-0.0.3.jar, core:/]
     [java]    at java.net.URLClassLoader.findClass(java.lang.String) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at gnu.gcj.runtime.VMClassLoader.findClass(java.lang.String) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at java.lang.ClassLoader.loadClass(java.lang.String, boolean) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at _Jv_FindClass(_Jv_Utf8Const, java.lang.ClassLoader) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at _Jv_PrepareCompiledClass(java.lang.Class) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at _Jv_WaitForState(java.lang.Class, int) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at java.lang.ClassLoader.linkClass0(java.lang.Class) (/lib/ssa/libgcj.so.4.0.0)
     [java]    at java.lang.ClassLoader.resolveClass0(java.lang.Class) (/lib/ssa/libgcj.so.4.0.0)
     [java]    ...7 more
     [java] Java Result: 1

BUILD SUCCESSFUL
Total time: 0 seconds
-bash-2.05b$

Any guesses on what I'm doing wrong?  Also, we have a lot of MediaWikis, and I don't see a "pages" table in any of their databases.

 Thank you.

Sorry, newbie question here: can I use this converter in a Windows environment, using the Standalone version of Confluence 2.2.2? I get lost right around the "execute '$>ant exec'" part, as when I unzip the files I don't have any executables...

I'm still waiting for a source release... didn't see anything in the attachments. My install of mediaWiki (for some reason) uses different table names. I need the code if I have any hope of doing this conversion.

Table data doesn't seem to convert correctly.  For example, a MW 1.5.7 dump that looks like this:

{|
|CMWO#
|Platform
|Start / End
|Leader
|Conf Bridge
|Notes
|-
|
|
|
|
|
|
|-

ends up geting rendered in Confluence like this:

| CMWO# |
| Platform |
| Start / End |
| Leader |
| Conf Bridge |
| Notes |
| |
| |
| |
| |
| |
| |

Why would this be?  Is there any way I can pre-process the dump so your converter can read it correctly?  Or does the converter need to be updated?

What's the story on the source code release?

We're about to convert 2800 pages, and it works pretty nicely. However, anything we can do to improve the fidelity of the translation will be worthwhile from our perspective. Here are some patterns we observed repeatedly:

1. two extra question-marks ('?') at the beginning of several pages, sometimes fouling up the Wiki markup for the first heading. Some times, this would show up as plain question-marks, other times it would come up as graphic characters:
   • ??h1.
   • ��
2. More issues with tables. This time, the translation of the following source format made a row for each field, and the bold markups came out in plain text:

|-
| style="border-bottom:2px solid black;" |<b>Requirement</b>
| style="border-bottom:2px solid black;" |<b>Status</b>
| style="border-bottom:2px solid black;" |<b>Time Estimate</b>
| style="border-bottom:2px solid black;" |<b>Owner</b>
| style="border-bottom:2px solid black;" |<b>Other Information</b>
|-

|-style="background: #33ff33"
|Prepare for work for this iteration
|100%
|16 Hours
|Charley
|Install New SDK, Study Docs
|-

In any case, we'd like to know one way or the other if the source code will become available. We're already looking at a substantial investment to reparent all of our pages into the appropriate spaces, and reattach our icon files.

All,
Source and binaries released on intient svn server as promised.

http://intient.com/wiki/index.php/Mediawiki_to_Confluence_Translator

Bradley

http://intient.com/wiki/index.php/Mediawiki_to_Confluence_Translator is out of date.

Is there a working link to find the importer anywhere?

thanks