2016 — 26 February: Friday
This afternoon's ceremony will mark my third attendance at the Wessex Vale crematorium.1 I was wondering if I still remember how to tie a tie?
Meanwhile...
... I have the makings of my next crockpot for assembly and thermal agitation this morning, with an experimental variation in the "gravy" department. And a puzzling bug (?) unearthed during the Kodification of "Henry IV Part 2" (of all things). It seems to have been triggered by an inexplicable change of case of one character in a generated string during the Python processing, but what do I know? It also explains the absence of that title from yesterday's screen shot of recently-added titles.
A phonecall...
... has saved me from libelling Kodi. It turns out a function buried within a call to a Python Unicode normalisation routine has decided to treat "IV" as a Roman numeral (fair enough) but to return it as "Iv" which is the root cause of some of the subsequent misbehaviour.
And, lest we become complacent at that point, there is another pitfall just waiting in the wings for its chance to trip us up. Kodi treats these two parts of poor King Henry as one entity, and "stacks" the entry for Part 2 in such a way (relative to Part 1) that the 'molehole' SHTML web page generator (until modified) would fail to parse the title "properly" on to two adjacent lines of my web list.
You could say King Henry is a Royal pain, in fact.
As funerals go...
... today's went pretty well, I think. Still far from my preferred social function. And it was damnably cold.
In between times...
... and now that I've finished the "Kodification" of all my film titles, I've been pecking away at what to do with titles that online DBs don't know about. Take, for example, this recent acquisition, for which there's no online "hit":
One workaround is to KodiPhy it, "accept" whatever title is a reasonable match, let it generate a dummy film stub file, populate the DB, and then go back into Kodi to tinker with the generated data. I don't really care what Kodi "makes" of titles it can't pinpoint. I'm just opportunistically using Kodi to populate its SQLite DB with my material. Once it's all in the DB, Brian's cunning Python can squeeze it back out into 'molehole' web pages for me, as this extract shows:
| The Fifth Element | C027 | 2009/03/13 | £14.12 | Blu-ray |
| Fifty Shades of Grey | B180 | 2015/06/26 | £14.99 | Blu-ray |
| Film-Fest #2 (1999) | R211 | 2016/02/08 | £4.55 | DVD |
| Final | L158 | 2006/12/29 | £2.25 | DVD |
| The Final Countdown | C033 | 2014/12/01 | £10.21 | Blu-ray |
Unlike the titles either side of it, there's no hyperlink to a corresponding IMDB entry for this "Film-Fest" DVD.
It would be nice to have a stub-generator that could be asked not to try to find a match, but merely to enter the data that interests me. My present generator will work if there's a "sort of" match that I can work on. Meanwhile, Len has been telling me of his suggested "start from an .nfo file" approach (rather than from a dummy film title "stub") for those cases where there's no online DB match. Local .nfo files can obviate any need for online DB scraping.