January 2nd.
The pages were initially detected as not unique but this has now been rectified with a canonical statement on each of the three ypc pages. this turned out to be due to the fact that the http version was seen as the master, probably because the links to these pages was initially to http from hra.
On Friday (Nov 28th) requesting a Live Test and Indexing on hpps gave the notice: user canonical not a as selected by Google, Google's opinion wins! So it will be interesting to see if indexation requests on both result in the correct canonical, https, being selected.
Saturday, apart from slight change to GSC forfmat of sites list, only minor changes.
Shortly after the above, both became indexed. The canonicals seems to have sorted themselves out, probably as a result of the indexing request on http.
It's interesting that prior to Grok, and the ability one has to ask complex question of Grok, that discovering how to use .htaccess - or even that one can - to prevent indexing of a class of files, would have taken many internet searches, and then would only have revealed the answer if the right things had been looked up.
But with Grok the ability to ask questions such as 'how do I...?' makes arriving at the conslusion and the means by which one might achive the objective much simpler and more direct. However it is not all plain sailing.
In response to one question about generating many copies of a file, each one with a different name, that name selected from a list in a text file, Grok gave a couple of really quite complicated and tortuous answers. Fortunately, having seen the correct answer some weeks before but having forgotten it, Grok was prompted to suggest another asnswer. That too was tricky and not one that was recognised.
It was only on the third attempt that Grok gave the simple answer, which involved a single line entered into Command Prompt. Easy to do, no fiddling around with obscure setting anywhere.
Grok helps greatly but lacks the human common sense to give the simplest and easiest way of doing something. But Grok is improving and learning so Grok will get there before too long no doubt.
Sequence modified for the second time.
Google notifications of reindexing received. Not everything indexed by any means. Indexing requests made for indexed files.
Sequence of pages changed. Order changed again Thursday Dec 4th and again Update: Tuesday 9th.
So all the .doc files - and there are more than 1,500 mof them - were converted and uploaded, and the relevant htms were modified to refer to the .docx files, and all seemed well.
But two days later - possibly when the revised pages were indexed - revenue fell off a cliff edge, down 85%-90% literally overnight. After a couple of days of that the sites were reverted to the status quo ante. But it made no difference, revenue remained on the floor.
Whether the drop was caused by the converstion or was just a conincidence is anyone's guess. But if it was a coincidence it was quite a coincidence.
This page and its siblings, and excluding .docx files from indexing by means of an .htaccess entry are part of the solution. When the .docx were uploaded and referred to there was no Disallow in the robots.txt file for .docx files - there always has been a Disallow for .doc files. So some of the .docx files were crawled and indexed. How that might have caused a problem one cannot say, but in case it did and the problen wasn't a coincidental glitch, the strategy is to get essentially empty versions of all the .docx files to be crawled, then when that has happened to stop further crawling of the .docx files by adding Disallow .docx to the robots.txt files.
Then when the .htaccess is taken note of and .docx files that are currently indexed have all been un-indexed the decks will be clear. The dummy .docx can be deleted which puts things back to where they were, and where thay had been for the past twenty years.
This is a very interesting experiment to see how quickly Google finds new links and shows them on the Pages lists in Google Search Console. So far it's been a slow process and there seems to be a lag between uploading and crawling and a lag between crawling and crawled links appearing in the Pages lists.
Order of hrefs switched around on Nov 29th.
Tuesday December 16th.