I've just arranged for non-encrypted (HTTP) requests to the Strathspey sites to be redirected to HTTPS, which has been working on Strathspey for a long time but had to be addressed explicitly. I've also enabled ”HTTP Strict Transport Security” which should cause browsers to make any
http://… requests to Strathspey-hosted domains to
https://… instead without even trying HTTP in the first place, once they've accessed the site in the reasonably recent past. (For the time being the holding time for HSTS is only 10 minutes; this will be increased to six months or a year once I'm satisfied it all works, because it's a hassle to undo if anything goes wrong.) This should bring the Strathspey sites up to the standard that is expected in 2019.
Note that this does not apply to the RSCDS Vienna Branch site as for technical reasons I can't obtain an X.509 certificate which covers that site. This will hopefully be fixed in due course.
· · Posted by Anselm Lingnau · 6 December 2019 2:13
It's time once more to crack down on spam-type user accounts on the my.strathspey system, which are annoying and seem to outnumber the real accounts.
From looking at the database, it seems that most if not all of these unwanted accounts either have both their first and last names set to the username of the account (which in real life would be weird) or else have no first and last names set at all. In theory we require first and last names, so any account that doesn't have any at all is technically breaking the rules of the game.
What I'm going to do a week from now is delete all accounts that fit this description where the user in question hasn't logged in for six months. If you want to make sure your account isn't inadvertently deleted, you should:
Ensure that it has a first and last name associated with it (it would sure be nice if it was your actual first and last name but if you call yourself “Hamish McGonigle” nobody is going to be the wiser), and/or
Log into your account. You can log right out again but we want to know that you're there.
Accounts that have dance lists, ratings, or collections in the dance database will also be exempt.
After these suspect accounts have been removed, all remaining accounts that have no first and last name associated with them will have their first name set to “Anonymous” and their last name to “Dancer”. The users in question will of course be free to set them to something else in their profile (except the empty string or any name consisting of all spaces).
· · Posted by Anselm Lingnau · 18 January 2018 1:15 (last edited on 18 January 2018 1:16)
Just secured the domain name “scd.club”. This means that I'm now in a position to offer free domain registrations of the form “XYZ.scd.club” if anyone is interested.
I'm probably going to add “frankfurt.scd.club” as an alias of “www.frankfurt-scd-club.org” just for kicks.
· · Posted by Anselm Lingnau · 4 February 2015 12:10
There appears to have been an outage of the Strathspey (web) Server earlier today, which I am unfortunately not in a position to investigate in detail (being in Portugal and all that). I restarted the web server and it appears to run, and we should all keep our fingers crossed that it will continue for the next few days because I don't really know if and when I'm going to be online.
(Before you ask: I used to have an automatic check that would advise me if the server was down, but my mobile phone data plan doesn't work outside Germany, and the infrastructure in question wasn't quite carried across to the new server machine in the first place. I'm going to look into crash-proofing the server better when I'm home after next Sunday.)
· · Posted by Anselm Lingnau · 8 December 2014 1:38
The update seems to have gone through OK – it was a bigger job than originally planned but the system is now in as good a shape as it ever was.
We're also just half an hour of programming and a 2-hour import run of nearly 70.000 messages away from getting the Strathspey Archive back on track. There are a few loose ends to be tied up (e.g., dealing with the address rewriting that the nasty AOL/Yahoo! DMARC business inflicted on us earlier this year), but I'm quite positive it won't be long now. This goes hand in hand with an updated version of www.strathspey.org, which means I get to finally get rid of the previous version of the framework. Yay!
As always, if you find any anomalies or have other questions, suggestions, or criticism then do let me know.
· · Posted by Anselm Lingnau · 20 November 2014 1:59
The my.strathspey status updates now include news from the “News“ page of the RSCDS web site at www.rscds.org. These are “scraped“ from the actual site at regular intervals as the Society does not publish an RSS or Atom feed (which would be the usual method).
Unfortunately, news items on the RSCDS site do not carry a date, which is why we get a whole bunch of them at one go (the Strathspey system needs to catch up, and we can't easily back-date them to where they ought to belong in the stream). From now on only new items should show up, though.
This is a side benefit of work that I'm doing in order to move the www.strathspey.org site over to the current platform. Devotees of the Strathspey Archive will be pleased to know that the new archive implementation seems to work well on my machine (including PostgreSQL-based full-text search in past postings), and this will also be available on the public server in due course.
· · Posted by Anselm Lingnau · 14 November 2014 15:56
I'm currently working on the Strathspey Archive with a view to not only reinstating timely updates, but also improving the user experience with regards to threading etc. This is proving more time-consuming than I thought because I need to re-index all 65.000+ past Strathspey messages, and the early parts of the Archive are in a fairly terrible state when it comes to metadata.
In particular, many messages have the same message ID (which theoretically isn't supposed to happen at all), and rather a lot of messages either refer to one of those duplicates (which needs resolving by hand) or else have something weird in their »In-Reply-To« header (which could be resolved by hand but I don't have the time just now).
The very early parts of the Archive have moved from one machine to the next for a very long time and it seems that the archiving software of the early days had a few nasty bugs. I hope that when I'm finished with the problems within the first few thousand messages things will improve. (Right now I'm in the 1800s.)
The strategic importance of this work is that I'm trying to move full-text search into the Django implementation of the mailing list archive (and in this process can manage to get rid of the external Sphinx search engine), and that we'll also get archives for the other mailing lists hosted on the system, most notably dancedata-friends.
· · Posted by Anselm Lingnau · 5 May 2014 11:38