How to include attachments in a full XAR export (including children)?

Help / Discuss
1

Hi everyone,

we are currently preparing a system migration and trying to clean up our existing XWiki (v10.10) content before moving to a new platform.

Over the years, many attachments (especially images) have been left orphaned because they were deleted in the editor but never removed from the page’s attachments. Our goal is to:

  • Export a complete page/module (with all its children)
  • Include only valid (linked) attachments
  • Exclude unused/orphaned files
  • Get this as a clean .xar file that can be reimported

We’ve tried multiple approaches:

  • Manual .xar export via UI → doesn’t seem to include attachments (at least in v10.10)
  • REST API access → works partially, but doesn’t return child pages reliably or lists empty <pages/>
  • Script-based HTML crawling + attachment download → possible, but not ideal for clean reimport

Main question:
:point_right: Is there a reliable way (via UI, script, extension, or newer version) to export a .xar file that includes all children + their used attachments, without pulling in orphaned files? OR orphaned Files aside, is there a way to export with attachments in older version like ours in the first place?

Any advice from the community would be highly appreciated – especially if there’s an extension or export method we’ve missed.

Thanks in advance! :pray:

2

Hello , :wave:

Attachments are included in the XAR whether you use the Page export feature or the Admin UI Export feature.

Not that I know of. It’s really hard to define an orphaned attachment. It could be referenced in a script for example.

Yes, the XAR export does that. See https://www.xwiki.org/xwiki/bin/view/Documentation/UserGuide/Features/Exports#HXARExport for example

Thx

3

Found this thread from 2019: Finding unused attachments on page (from you hehe :))

4

hi Vincent,

nice to hear from you again.
yeah this topic haunts me for ages!:wink:

So i have a hard time finding the attachments then.
When i look into a .xar file, the largest files i find are the webhome.xml’s.

Are the attachments bound to those?

Otherwise i dont see any attachments, and there is no “in-/exclude” Option in the xar export as well.

5

:slight_smile:

Before replying I tested on a local XWiki instance and exported the Sandbox.WebHome page (which has an attachment) and the attachment is there:

  <attachment>
    <filename>XWikiLogo.png</filename>
    <mimetype>image/png</mimetype>
    <author>XWiki.superadmin</author>
    <date>1751287729000</date>
    <version>1.1</version>
    <comment/>
    <content>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</content>
    <filesize>1390</filesize>
  </attachment>
6

Just to be clear. The Old Version runs on V10.10. Thats a very old version. Sure exporting attachments was included in that area? Were talking min 5 years ago.

Cheers,
T.

7

Indeed that’s old (7 years ago :)). What’s sure is that if you use the Admin UI to perform the export (see https://extensions.xwiki.org/xwiki/bin/view/Extension/Administration%20Application#HExport ) you’ll have the attachments in the XAR export.

8

Ah…we’re onto sth here…that looks drastically different then what i am using. So pls explain to me bcuz i’m dumb. Is that an extra admin extension which i can download or should that be included in the standard 10.10?

9

grafik
grafik1598×884 45.2 KB

thats what i see under export in the administration menu.

and thats what i see when going to the pages export menu
grafik

10

yes that’s the Admin UI export. If you use that, it’s sure you’ll get the attachments in the XAR since that feature exists since the beginning of XWiki.

11

ok now i feel completely stupid?
I dont know what to enter in this mask.
Keep in mind i want to export pages separately, not in a whole?

I have no real export options on that mask.

pls help!=)

12

Export everything and then reimport only what you want. Or unzip the XAR, remove the pages you don’t want (and update the package.xml file), and re-zip.

13

ok, gotcha, thats what i did now. The whole zip is just 2GB while the application on the server is way above 25. Also i cant find anything in the folders. Is there any button i have to click? The options history and backup dont indicate to include attachments. Also, what is the webhome.xml and why is it so big in every folder?

14

what folder?

Because there isn’t. Attachments are always included. Just try it and you’ll see for yourself!

These are the nested pages. See https://www.xwiki.org/xwiki/bin/view/Documentation/UserGuide/Features/ContentOrganization/#HTerminology

15

I made the whole backup file from xar to zip and treid to find the attachments somewhere, especially pictures. I cant find them and by the whole backup size i was assuming, they were not included.

The original plan was to take the modules one by one. Letting them beeing cleaned by chatgpt by comparinng the links to the attachments and getting them reimport. Chatgpt was not able to find any attachments so it assumed its not a feature of this version.

16

Attachments are stored in XAR, they are just a bit hidden as you would probably expect them to be in a separate folder. To find out, you need to check the page in any IDE by opening the XML file and looking for the attachment tag. Also a hint that the exported page contains attachments - is a certain size (i.e my exported Sandbox page WebHome.xml almost ~5MB size.)

<attachment>
<filename>massimiliano-morosinotto-Tw_VQuS3Xio-unsplash.jpg</filename>
<author>XWiki.npetrenko</author>
<date>1748852820000</date>
<version>1.1</version>
<comment/>
<content></content>
<filesize>2315099</filesize>
</attachment>

I made a question for ChatGPT model XWiki Begleiter (so far, it’s developed by some users not from XWiki) and at least it can provide a valid answer

image
image1142×1658 99.2 KB

In an exported XAR file from XWiki (with history and attachments), you can verify that attachments are included by checking both the XML and the file structure.

:white_check_mark: What to check in the XML

Open the document.xml file inside the page’s directory and look for:

<attachment>
17

Hello again. Thanks for your answer so far. I was able to find out how and in which xml’s all the data is stored. Together with my new friend Claude :wink: i wrote a program to find and delete orphaned Data

grafik
grafik896×723 24.2 KB

so far so good, where we still struggle a little is the included History Data. Is there any chance to export a xar of a page (not the whole wiki) without history data? Chatgpt told me about a extension which can delete history. Is that otherwise my only chance?

18

Just look at the export screen… and don’t check the « with history » checkbox.

19

grafik
grafik587×507 20.2 KB

Vincent,where?

20

See the screenshot at: