Page MenuHomePhabricator

Access issues for Ancestry
Open, In Progress, Needs TriagePublic

Description

While we can access Ancestry via the proxy successfully, the site looks to be having some issues when browsed.

Steps to reproduce

Screenshot 2023-01-23 at 13.47.01.png (2×2 px, 566 KB)

Event Timeline

As a temporary fix, I've found that whenever this happens, I've been able to refresh the page, and it will load the search results from the page URL.

jsn.sherman changed the task status from Open to Stalled.Mar 8 2023, 4:12 PM

I'll resume work on this after we're done using staging for newspapers.com

Nikki reports that the back button in your browser will cause you to stop being recognised as having access. This may be an unrelated issue.

jsn.sherman changed the task status from Stalled to In Progress.Apr 13 2023, 4:33 PM
jsn.sherman moved this task from In Progress to Ready on the Moderator-Tools-Team (Kanban) board.
jsn.sherman subscribed.

Setting this aside to reflect reality.

We should take care of the ancestry login for newspapers.com when we tackle this.

At first glance, the issue looks identical to T322916. My guess is that both sites are configured similarly, given how closely linked they are.
I'm starting with the same cookie change we made for newspapers.com, which is currently waiting on deployment to staging.
https://github.com/WikipediaLibrary/twlight_ezproxy/pull/37

It looks like we'll need to do a little bit more here to address the hostname rewriting that happens on the cdn side of things. Happily, there's a decent breadcrumb for the problem in the other task.

jsn.sherman changed the task status from In Progress to Stalled.May 12 2023, 4:56 PM

@Samwalton9 I believe I have taken this as far as I can on my own. Our web requests are going to the right hosts and the cookies I spot checked looked good to me. Could we bring in some partner technical assistance like we did with newspapers.com?

@Samwalton9 I believe I have taken this as far as I can on my own. Our web requests are going to the right hosts and the cookies I spot checked looked good to me. Could we bring in some partner technical assistance like we did with newspapers.com?

I've asked.

There are several features on Ancestry and Newspapers which don't work via the proxy. For example, in the family-tree viewer the "Media gallery" and "Family group sheet" options send you to a 404 page. More seriously, trying to view an Ancestry user profile also results in a 404 while on Newspapers user profiles are viewable but attempting to Follow a user (Following is apparently necessary to contact registered users through their profiles) brings up a login prompt for an individual Ancestry or Newspapers account. I don't know whether it's worth trying to figure out how to report all this, or if it's all just considered (not)-working-as-intended.

jsn.sherman changed the task status from Stalled to In Progress.Jun 30 2023, 8:05 PM

I've resolved a number of underlying issues in staging, though I haven't solved the paging issue yet. I have a lead on where that problem is happening. See the the previous & next properties of this JavaScript object:

window.ancestry.search.pageModel = {
    "number": 2,
    "cursor": null,
    "size": 20,
    "max": 182750,
    "firstHit": 21,
    "lastHit": 40,
    "allowedPages": {
        "1": "?name=jon_doe&defaultFacets=PRIMARY_YEAR.PRIMARY_NPLACE",
        "2": "?name=jon_doe&pg=2&defaultFacets=PRIMARY_YEAR.PRIMARY_NPLACE",
        "3": "?name=jon_doe&pg=3&defaultFacets=PRIMARY_YEAR.PRIMARY_NPLACE",
        "4": "?name=jon_doe&pg=4&defaultFacets=PRIMARY_YEAR.PRIMARY_NPLACE",
        "5": "?name=jon_doe&pg=5&defaultFacets=PRIMARY_YEAR.PRIMARY_NPLACE",
        "6": "?name=jon_doe&pg=6&defaultFacets=PRIMARY_YEAR.PRIMARY_NPLACE",
        "7": "?name=jon_doe&pg=7&defaultFacets=PRIMARY_YEAR.PRIMARY_NPLACE"
    },
    "prev": "http://www.ancestryinstitution.com.wikipedialibrary.idm.oclc.org:9080/search/?name=jon_doe&defaultFacets=PRIMARY_YEAR.PRIMARY_NPLACE",
    "next": "http://www.ancestryinstitution.com.wikipedialibrary.idm.oclc.org:9080/search/?name=jon_doe&pg=3&defaultFacets=PRIMARY_YEAR.PRIMARY_NPLACE"
};

There's another error I noticed (alongside infinitely-loading search results) that looks to be related. Drop-down lists also infinitely load, despite clearing cache, disabling ad blockers, and using a different browser. Refreshing the page freezes selections in place.

To reproduce:

  1. find a collection of images via "explore by location" (https://www-ancestryinstitution-com.wikipedialibrary.idm.oclc.org/search/)
  2. try to "browse this collection"
  3. observe whether it is possible to make all the selections to continue; I get stuck on the second selection (top image)
  4. refresh the page; the second drop down list is frozen (bottom image)
  5. to unfreeze the list, redo the first selection, and we're back at step 3

image.png (339×379 px, 12 KB)
image.png (355×379 px, 11 KB)

I spent a few hours today trying search & replace for all of the hostname building javascript functions for the search page. I was not having any luck, but I realized why at the very end of my time with it: the search function I was using was deminifying the javascript, making all of my queries for text replacement incorrect. I'll keep that in mind on my next pass at this. Of course, it would go quicker if we had some support from ancestry to just let us know where the properties for the search pageModel are set.

setting this aside for now