The way that confluence builds URLs is not the same as MoinMoin - here are some URL mappings that could be set up (via a redirector installed with the old wiki URL) that would make some existing links/bookmarks to continue working. I'm not thinking that all will work but at least direct links to pages would be nice.
/display/<SPACE>/<TITLE>
This works like a regular wiki (though there may be some differences in URL encoding to be checked) - however note that I've seen one example where a title includes the "+" character and Confluence seems to choose to switch to the /pages/viewpage.action approach (see below).
The example:
Title: How to Install MM3a7+REST+Django+APP
URL: http://wiki.list.org/pages/viewpage.action?pageId=11960560
Whereas it seems that most pages use this human-readable type link - for example:
Title: New Menu Structure_Layout
URL: http://wiki.list.org/display/DEV/New+Menu+Structure_Layout
Note that in order for a mapping to work we'd need to store SPACE + TITLE information for all existing pages.
Where the SPACE/TITLE page naming is preserved (which is by far the most convenient option), a redirect would just remove the /display prefix.
/pages/viewpage.action?pageId=<PAGEID>
Example: http://wiki.list.org/pages/viewpage.action?pageId=11960560
As per comments above (in the /display/ section) it seems that Confluence chooses to switch to an ID-based URL instead of encoding characters like + in a human-readable URL.
In this case the PAGEID is the inner-text value of the first id element within a page object element - for example the page starting with this XML in the dump:
<object class="Page" package="com.atlassian.confluence.pages"> <id name="id">7962730</id>
Is accessed with either of:
As per this example - note that both approaches can be used for most pages. I have not found a url encoding that allows the How to Install MM3a7+REST+Django+APP to be accessed via the /display/ handler.
/x/<TINYID>
Example: http://wiki.list.org/x/AgA3
Here, the AgA3 "tiny URL" identifier, TINYID, is converted to an integer (in this case it is 3604482) using a simple algorithm and this is then used by Confluence to look up the actual page (as described above).
/pages/tinyurl.action?urlIdentifier=<TINYID>
Example: http://wiki.list.org/pages/tinyurl.action?urlIdentifier=AgA3
This is an alternate form of the <TINYID> reference. The <TINYID> is exactly the same as in the much more comonly seen /x/<TINYID> URL.
/dosearchsite.action?searchQuery.queryString=<TERM>&searchQuery.spaceKey=<SPACE>
This is effectively equivalent to the result of the following macro:
<<FullSearch(<TERM> title:regex:^<SPACE>/)>>
This can be converted to a URL of the following form:
http://wiki.list.org/?action=fullsearch&context=180&value=<TERM>+title%3Aregex%3A^<SPACE>%2F&fullsearch=Text
/spaces/space-bookmarks.action?spaceKey=<SPACE>
This appears in the Mailman Confluence content, but does not seem to be supported any more.
/spaces/exportspacexml.action?key=<SPACE>
This exports the given SPACE as XML, and provides the basis of the conversion process. Moin provides the PackagePages action, but that takes a list of page names as one of its parameters, and unlike the Confluence exporter it does not provide all versions of a page. There are other ways of packaging a wiki using the moin program, but Moin doesn't seem to really support through-the-Web archiving.
/pages/doexportpage.action?pageId=<PAGEID>&type=<TYPE>
This is an obsolete method of exporting pages. Here, with a TYPE of TYPE_PDF a page is exported as PDF. Instead of this, the following URL path should be used...
/spaces/flyingpdf/pdfpageexport.action?pageId=<PAGEID>
ActionMarket/RenderAsPDF provides an action for PDF export and would be accessed using a URL of the following form:
http://wiki.list.org/PageName?action=RenderAsPDF
/dashboard.action
Confluence displays a "dashboard" showing some page content and various panels containing update information (wiki changes, blog posts) and a list of the wiki spaces. The intention is to show such content on the front page of a Moin wiki, so this URL path would just redirect to the root of the wiki.