How do I recover my content after a failed publication?

[mmulich]

This is possible two issues, but I'm more concerned with recovering my content.

The issue is that I've attempted to put an link to another page in a page within my book. And the content ends up in a "Failed/Error" state that I can't fix as a user.

| Book With Issue (BWI)
|- Page One (PO)
|- Page Two (PT)

To reproduce,

1. Create a book with two pages (as shown above).
2. Publish the book.
3. Copy the link to PO.
4. Edit PT to include the link to PO
   *. (For good measure add a keyword to the book so that it is edited as well.)
5. Publish

The workspace now shows the state as "Failed/Error". If you grab the publication id and look at the publishing response you'll see the error is an InvalidReference. For example:

{
state: "Failed/Error",
messages: [
{
xpath: "/div/p[5]/span/a",
code: 20,
publication_id: 748,
value: "http://dev.cnx.org/contents/f12b4c7a-73b5-472e-88db-3c7cd026c942@1/Gomvr",
epub_filename: null,
pending_document_id: 1114,
pending_ident_hash: "583344b8-c7f6-47c7-a9ac-a66e6b50e4bd@2",
type: "InvalidReference",
message: "Invalid reference at '/div/p[5]/span/a'."
}
],
publication: "748"
}

This error is probably valid since I don't think publishing knows about any other domains besides cnx.org and legacy.cnx.org. That will require further investigation within cnx-publishing and the reference resolution code.

This issue needs to answer the questions How do I recover my content? now that it is in a broken state.

[karenc]

I don't think this is a problem in authoring. See this test, if there's a publication error, you can save again and authoring will reset the status to "draft" and then you can try publishing again.

[rich-hart]

I think that the updated reference to Karen's test is this link

[mmulich]

I tried using this method of PUT to recover. It didn't work out for me. But I'm likely doing something wrong.

>>> cookies = dict(session="<session-info-from-request>")
>>> id = '3ad7cf08-a3eb-4f95-b32a-b9faf25141e9'
>>> url = 'http://localhost:8080/contents/{}@draft.json'.format(id)
>>> import requests, json
>>> resp = requests.get(url, cookies=cookies)
>>> resp
<Response [200]>
>>> data = resp.json()
>>> resp = requests.put(url, data=json.dumps(data), cookies=cookies)
>>> resp
<Response [200]>

Upon returning to the content in the browser, it strangely still wants to pull from the published version, which doesn't exist.