Talk:Main Page

From SVR Wiki
Revision as of 09:44, 6 August 2017 by Graham Phillips 110 (talk | contribs) (On this day refresh error: One problem solved)
Jump to: navigation, search

Spambots

Some of you have noticed the nuisance spambots floating about - they don't seem to be anything but a nuisance right now, but they do rather clog up the recent edits list...

I've been modifying the Captcha questions to try and find what works best, and the number at least seems to be reduced.

As an additional measure, I've also enabled blacklisting of known spam sources, as per this section, and have turned on the requirement for email addresses to be verified (I noted that none of the bots had emails registered). I note there's a couple of people who haven't got confirmed email addresses, but most people should be fine - if you need to confirm your address, you can find the option under "Preferences" in the top right corner, then by scrolling down to almost the bottom of the page of options.

--Danny252 (talk) 16:41, 6 February 2015 (UTC)

I've enabled a similar blacklist as well. --Graham Phillips 110 (talk) 22:32, 6 February 2015 (UTC)

We changed the Capcha question to a dynamic one which should have been harder for spambots to crack.
It wasn't working as well as hoped, so I have now reworded the question to see if that's any better.
I believe it's possible to create different user groups with different posting rights, so we could, for example move all the proven genuine contributors to a "trusted" group so that they don't have the inconvenience of having to solve a little puzzle every time they want to make an edit. I'll look in to that later.
--Graham Phillips 110 (talk) 10:13, 2 March 2015 (UTC)

Within an hour of making the above changes, we had a load more spam.
It seems to follow a set pattern of creating a new account and then a User Page, so I have disabled the ability for the default user group, "User", to create pages and manually added all the genuine contributors to a new User Group "Trustworthy".
If I have missed you out, or if you have any other problems, you should still be able to edit this page, so either reply here, email me, or post on the forum or Facebook page and I'll sort it out.
--Graham Phillips 110 (talk) 12:51, 2 March 2015 (UTC)

I've moved all the known genuine contributors to a "Trustworthy" user group and removed page creation rights from all others.
They're still getting through though somehow, so I have just disabled the ability to write directly to the API.
I'm not too sure what this means myself, but I think it's a way to bypass the normal page creation method and, as I can't see any legitimate reason to use it, and there was an option to disable it, that's what I've done.
--Graham Phillips 110 (talk) 17:07, 3 March 2015 (UTC)

I suspect that that is how they were getting through without page creation rights - basically it means sending the command to create a page directly without loading up the "Creating page..." form first.--WillSalt (talk) 22:17, 3 March 2015 (UTC)

I note that the bots seem to be registering, but not actually making any edits currently - am I right to think that? The last time that I can see a spam page *created* is the 27th of Feb. I've also fiddled with the settings a bit more, and it definitely should be only allowing users with confirmed emails to edit/create/whatever now (plus the Trustworthy group, of course).

--Danny252 (talk) 10:02, 4 March 2015 (UTC)

The pattern has changed slightly. The spammers no longer create their own user page, they create a separate page complete with spam title.
I've asked for advice on the Mediawiki support desk and got some interest, but not a solution yet.
It looks like my next step is to work out how to access the access log. I had a quick look earlier, but it's not as simple as I hoped. I'll put a bit more time in to it later.
--Graham Phillips 110 (talk) 10:43, 4 March 2015 (UTC)

Seems the bots may be smart enough to confirm emails as well, so I've removed that change I did, so it should be back to how it is. Regarding the comment on the helpdesk query about alternate access to the database, I've made sure that the old versions of wikis we had installed are now unable to be used at all. They shouldn't have been at all accessible, but now they won't work even if they are. --Danny252 (talk) 11:49, 4 March 2015 (UTC)

Would it be an idea to remove the mediawiki files that aren't being used from the server? There seems to be 3 installations of the wiki, 2 of which are from the early days and no longer needed(?) - and it is still possible to view those pages if you know the url to use. Sharpo (talk) 14:00, 4 March 2015 (UTC)

Just found a way to access logs, could be useful info - once I understand it more!! Sharpo (talk) 17:16, 4 March 2015 (UTC)

Another thing I did earlier today was to update some of the spam blacklists we've got setup. In some discussion with Sharpo over the access logs just now, I can see that a number of the IP addresses weren't being caught before, but should be with the new settings. --Danny252 (talk) 21:55, 4 March 2015 (UTC)

As a further update, Sharpo has been keeping an eye on the access log, and has noted that a number of bots were using a security hole which we patched up yesterday (a page which shouldn't have been accessible was). They still seem to be trying quite hard to use it, but no spam has appeared today. --Danny252 (talk) 18:48, 5 March 2015 (UTC)

I meant to say, if you need help interpreting the logs then feel free to ask, seeing as it's the sort of thing I'm supposed to know about profesionally! --WillSalt (talk) 20:05, 5 March 2015 (UTC)

I've manually removed a lot of the spam accounts, which whilst blocked and unable to be used, were still in existence and cluttering up various pages (e.g. the user list, which has gone from 300 to 30). This means that their "contributions" will likely have reverted to show just their IP addresses - don't worry, this doesn't mean that unregistered users have been editing the site! --Danny252 (talk) 09:28, 6 March 2015 (UTC)

Interesting 'feature' is that the Recent Changes now shows your own IP address as the source of the former spam accounts! Gave me a minor heart attack at first! --Robin (talk) 14:29, 6 March 2015 (UTC)

After my last comment on this page, Danny252 and Grahamgave me FTP and CPanel access to the site, and I had a bit of a chat with Danny about what I found and what I thought could do with tidying up (hang on whilst I pop my Sensible Professional IT Consultant hat on...). The end result is: I have just finished doing a big cleanup and deleted just about everything that wasn't needed to run the current install of MediaWiki (I did leave a huge debug log that was hanging about in case anybody wanted it). I did take a backup before I did it, so if you're worried that anything has gone missing, don't be. Please let me know if anything now appears to be weirdly broken, but it shouldn't be. The point of the cleanup is: although everyone else has already done sterling work at disabling the configuration of the old installs, making sure that they're not linked anywhere, and so on, getting rid of extraneous stuff altogether is the best way of reducing the "attack surface" - the number of places people could get in, possibly using hacks we're not ourselves aware of. --WillSalt (talk) 22:28, 8 March 2015 (UTC)

Stourport and Tenbury branches

Would it be more correct to refer to these as the Stourport Line and Wyre Forest Line on the map and alter the links, page titles and redirects to suit?

My inclination would be not to bother (to the casual SVRSevern Valley Railway visitor, both are branches off the line they are on). The former was simply a continuation of the old SVRSevern Valley Railway rather than a line in itself. The latter seems interchangeable between Wyre Forest Line, Tenbury Line and Tenbury Branch; Google favours the former but I'd be interested to dig out some original sources to see what the commonest usage was at the time. I don't think anyone browsing the wiki will get too confused as it stands.--Robin (talk) 18:32, 6 February 2015 (UTC)

OK, I've changed the link on the map to point to Wyre Forest Line, while keeping "Tenbury Branch" on the map itself. As the Wyre Forest Line article says, most Kidderminster to Bewdley trains continued towards Tenbury, so if anything, the Bridgnorth line was the branch. However, going by the fact that the disused line branches away at 90 degrees from the currently used line, it seems more intuitive to refer to it as the branch in a modern context.

Would it be a good idea to have a suitable SVRSevern Valley Railway image or logo on the main page. For a visitor seeing the page for the first time, it would give them some idea what it was all about. Yes, I know it should be obvious, but......

Sharpo (talk) 00:09, 13 January 2015 (UTC)

I've been thinking the same (I just mentioned it in one of the forum threads). I guess the main question is whether the SVRSevern Valley Railway would be happy with us using its logo - it's starting to sound like we'll need to get in contact with the office at some point on a few of these things. If they decide not, does someone feel suitably arty?

--Danny252 (talk) 09:23, 13 January 2015 (UTC)

Perhaps in the meanwhile we should've adopted the SVRSevern Valley Railway Online logo. Also; starting to add headings here to separate topics. I hope you're all OK with this. Boldford (talk) 09:39, 1 February 2015 (UTC)

I emailed marketing@svrlive.com on 16th January to ask permission to use the official SVRSevern Valley Railway logo. I haven't heard back yet, so I've gone and used it anyway. If I get a threatening letter from their solicitors, I'll change it back again. --Graham Phillips 110 (talk) 07:56, 2 February 2015 (UTC)

That logo looks much better! Sharpo (talk) 08:05, 2 February 2015 (UTC)

Location

We'll probably spend a lot of time providing info about the SVRSevern Valley Railway, but we also need to show where it is for people who come across this wiki & decide the railway might be worth a visit. So, for those arty people amongst us, how about finding a suitable size image of the UK that can be used, with a highlighted area showing exactly where the Railway is, then showing an enlargement of the area alongside the UK image with the main stations highlighted and the course of the railway.

Sharpo (talk) 11:39, 13 January 2015 (UTC)

On the subject of maps, I've adapted the map at [1] to only show as much of the SVRSevern Valley Railway as is currently in use. It works OK on my sandbox, but not here http://www.svrwiki.com/Map.
I gather I need to import the template that the map is based on, but I can't find instructions on how to do that.
I was going to duplicate the full GWRGreat Western Railway/SVRSevern Valley Railway map and create a more detailed map, showing more features as well.
Is it worth persevering, or would it just be easier to create the maps some other way? --Graham Phillips 110 (talk) 10:42, 14 January 2015 (UTC)

I was looking on the internet yesterday for suitable maps & saw that very same page, then thought I would do a tracing from a map & then create my own simple map from that - then scan & upload it. It was then I realised I was spending time on this instead of doing other things!! In answer to your question, I don't know how to edit that template to how you want it. May be better to draw a new version? Sharpo (talk) 11:00, 14 January 2015 (UTC)

I think the main issue is that the map requires a standard set of images (about 3000!) that exist on Wikipedia, which you can tile together to show the layout (e.g. plain track, station, disused station, and so on), as well as all the templates it pulls in. It might be possible to copy it over, but I don't know if it would be a good idea to pick and choose the bits we want, or alternatively to copy the whole lot over with a lot of redundant bits, but with the flexibility to make more in future.

--Danny252 (talk) 12:15, 14 January 2015 (UTC)

I was hoping it would be possible to import the template and just those icons used as it's a handy way of linking directly to stations on the line directly from the map. I was then going to create a more detailed map showing all the bridges etc. with a link to each. Maybe I'm asking too much and it would be best to stick to a normal image file of a map. --Graham Phillips 110 (talk) 12:55, 14 January 2015 (UTC)

I remember years ago using some free software that allowed me to add links from a word or words pasted into an image on my site. There may be something similar for a wiki if you have time to try & search for it. This was the sort of thing I just found after a search, but not within a wiki - http://www.onextrapixel.com/examples/image-map/

Back again, just found an example on a wiki, there is an image where you can click on the different people you can see. http://www.mediawiki.org/wiki/Extension:ImageMapSharpo (talk) 15:09, 14 January 2015 (UTC)

Wow, took me a few hours to understand it, but I have learnt how to create more than one link from within an image within a wiki. My own wiki test included 2 locos in a photo & moving the cursor to either of the locos would give a link for that loco. Sharpo (talk) 21:40, 14 January 2015 (UTC)

I've been spending far too long on this as well, but I've finally got there.
Click on the town names on the Map .
It takes a bit of trial and error to get the box coordinates in the right place, but I think this is a better than the standard route map as it allows linking of nearby features of interest, such as the Engine House. --Graham Phillips 110 (talk) 01:46, 15 January 2015 (UTC)

Some good work there. The main thing now is to get a suitable map - the OSOrdnance Survey maps will presumably somewhat awkward, given that the "Kidderminster" label is nearer to Bewdley Station than Kidder! --Danny252 (talk) 10:14, 15 January 2015 (UTC)

I like the map on the front page. Would it be worth putting on each of the station pages to show their positioning on the SVRSevern Valley Railway? --Brick60000 (talk) 16:17, 16 January 2015 (UTC)

Making changes

UpIn reference to the direction of travel means towards the major terminus (i.e. towards Kidderminster on the present day SVR) to now, everyone's been making changes without consulting anyone else. There haven't been any disputes yet, but should we be putting things to a vote first?
Minor edits of text should be OK, but what about more conspicuous changes like adding the maps and logo? I just went ahead and did it. Should we have discussed it first? I can easily undo it.
Should we have a picture on the front page? How do we pick one?
--Graham Phillips 110 (talk) 08:43, 31 January 2015 (UTC)

If you decide to have a picture on the front page, keep it interesting by changing it every week. Otherwise, once this wiki nears completion (if that is ever likely to happen) people aren't likely to look at it very often. Sharpo (talk) 13:46, 31 January 2015 (UTC)

My 10p worth is so long as you think it is an improvement, carry on and add/edit as necessary which is the idea of a Wiki; stopping for votes just delays things. Worth adding a note in the summary if comment is required, eg if I had a better logo (which I don’t!) I would upload it and say “shall we try this instead?”.--Robin (talk) 17:32, 31 January 2015 (UTC)

I've "protected" the front page, logo and maps to prevent any malicious edits. We haven't had much spam yet, but there was the possibility that someone could alter the front page to something unwanted. If anyone wants to make a genuine edit, they can put it here on the Talk page and one of the administrators can do it.
I created the logo in Paint on my PC and thought I'd got a reasonably good match for the chocolate & cream. The cream looks a bit beige now on my chromebook. If anyone's got the correct colour codes, I'll do a new one.
I don't know if a picture as well as the two maps would be too much though?
There's a random image extension for wikimedia which can be installed.
One idea I had was a long thin picture of a full train in profile, taking up the full width of the page, but only a couple of lines worth of depth. The picture could be selected at random from one of each carriage set. Worth a try?
--Graham Phillips 110 (talk) 09:27, 1 February 2015 (UTC)

I decided to tweak the wording a bit - please revert it if you don't like it! --Robin (talk) 16:14, 16 December 2016 (UTC)

Image of the month

A couple of thoughts on pictures, following a revisit to Festipedia.

They have a picture of the month on the front page. Would that be a reasonable frequency to adopt, especially as the mods would presumably have to change it?
They also have a lot of historical photos 'from the collection', especially on the loco pages. Does the SVRSevern Valley Railway have a similar collection (Kidderminster Railway Museum?) and if so, is anyone on speaking terms with the owner(s) who could broach the subject of adding some here? Particularly the SVRSevern Valley Railway in GWRGreat Western Railway days and SVRSevern Valley Railway locos in original service (did any of our current locos definitely work on the original SVRSevern Valley Railway?), and also early days of the SVRSevern Valley Railway in preservation.

--Robin (talk) 12:52, 2 February 2015 (UTC)

I'm planning to go down to the KRM this weekend for their signalling exhibition, and I'll have a go at mentioning this. I don't know whether they've got any plan to digitise their collection, but if so it would likely be a good source of images. --Danny252 (talk) 12:57, 11 June 2015 (UTC)

Using images with unknown copyright

I've avoided doing so until now, but uncredited photos keep turning up on Facebook and it seems a shame it miss out on them, so I've added two to the Dowles Bridge page with what I hope is a suitable "Get out of jail free" disclaimer. Even with an 1868 photo, there's no guarantee that the copyright has expired under the 70 year rule. Any thoughts on the legality of this ?--Graham Phillips 110 (talk) 07:27, 30 January 2016 (UTC)

What I've not yet worked out is whether the copyright is (or at least can be) "extended" if the work is sold to someone else - if I buy a collection of old photos, do I get any rights, or is it still all linked to the original author (unless they specifically give some permission/waiver)? Equally, I'm not sure if digitising a piece of work is ever able to extend copyright (e.g. if I scan an old photo, do I get rights over the scan of somebody else's work?). If neither of those apply, then it's 70 years from the production of the work if the author is not known, or 70 years after their death if it is (so for an 1868 photo, the author would have had to survive until 1946). --Danny252 (talk) 16:30, 30 January 2016 (UTC)
Regarding whether you can extend copyright by digitising a piece of work, the Ordnance Survey relinquishes crown copyright after 50 years as per the link I posted here. However the National Library of Scotland claims copyright over their scans of OSOrdnance Survey maps which are older than that, hence it appears you can do so (and I have updated all the OSOrdnance Survey extracts accordingly).--Robin (talk) 17:32, 12 July 2016 (UTC)

Changes needed please

A section to note any amendments needed on this page for the attention of the administrators. Suggest we delete items when done to keep the list current.

Article/page count

I've removed the line in the introduction that gave the current number of articles/images on the wiki. I had recently noticed the main page was taking much longer to access than any other page, and that other pages would not load until the main page was loaded, implying it was somehow locking up a large number of resources. Removing the page/image counts, which were probably re-counted each time the page was opened, appears to have significantly improved the speed! I suspect this is a combination of the actual article counting process being slow for some reason, as well as the main page having to be generated from scratch each time it was viewed (in case the counts had changed) which is quite a slow process... --Danny252 (talk) 08:47, 12 July 2016 (UTC)

Does seem quicker - no great loss not having the information. --Robin (talk) 17:32, 12 July 2016 (UTC)

Wiki content license

It was realised earlier today that it was never made clear what license was applied to content on the Wiki, with the exception of uploaded files (where a license is picked on uploading).

I've applied the Creative Commons Attribution-ShareAlike 3.0 license, the same as is used on Wikipedia - this license does not apply if a different license is specified (e.g. uploaded images).

If you have any complaints, speak now, or forever hold your peace! --Danny252 (talk) 22:34, 19 November 2016 (UTC)

No complaints, but could you tweak the site settings to include the CC icon in the button at the foot of the page where currently is just shows 'Creative'. Found this page of instructions which may help. --Robin (talk) 16:45, 24 November 2016 (UTC)
Like so? --Danny252 (talk) 21:04, 24 November 2016 (UTC)
That looks better, thanks!--Robin (talk) 21:11, 24 November 2016 (UTC)

SVRSevern Valley Railway photographic guide

This comes up periodically. There is no published guide of publicly available photo locations. There is an old thread at http://railways.national-preservation.com/threads/svr-photo-guide.5590/page-2 with quite a lot of info and a thread around PacificLocomotive with a 4-6-2 wheel configuration Power. I'm not a photographer but a guide would be useful. It would need some thought though, so as initial suggestions what it might include: Work north to South Location name GPS OSOrdnance Survey reference Means of access Particular restrictions Tips - better in the afternoon than morning type thing Example photo

Also a does and don'ts section (trespass etc.)

Thoughts?--Patrick Hearn (talk) 10:55, 19 December 2016 (UTC)

I'd erred away from doing one in the past, in case the SVRSevern Valley Railway considered it to be encouraging people to photograph from lineside (even in legitimate viewing locations) rather than paying to visit the railway. However the SVRSevern Valley Railway itself gave a number of possible viewing locations for PacificLocomotive with a 4-6-2 wheel configuration Power, so maybe not considered an issue? If we do add one, your suggestions would be a good start.--Robin (talk) 15:59, 29 December 2016 (UTC)

Official response from Lewis Maddox: "Great idea, if you need any photos let me know." So, going back to the original post, what's the best way of setting it up? Does it need a table with the headings I suggested. Is there a better way? This might be one where a little thought and input from photographers is worthwhile...--Patrick Hearn (talk) 13:06, 4 January 2017 (UTC)

Just while thinking about it, on the Current events on the Severn Valley Railway page, I put a comment relating to SVRSevern Valley Railway Live that "Information which is publicly accessible for enthusiasts includes Traffic Notices, Working Timetables, lineside photography policy and application forms, and so on." However I can't see photography policy on the new version, or on the SVRSevern Valley Railway website either. Any idea where they would be found these days?--Robin (talk) 16:50, 4 January 2017 (UTC)
When I looked earlier this year, the lineside photography information available on the old site seemed to be horribly out of date, and upon enquiring directly with the SVRSevern Valley Railway on Facebook, the response I got gave me the impression they were only interested in direct inquiries, rather than advertising it online? --Danny252 (talk) 09:52, 5 January 2017 (UTC)
In terms of setting it up, the unofficial WSR site WSR.ORG.UK has a photography section here. There may be some good ideas to be had from it.--Robin (talk) 16:55, 4 January 2017 (UTC)
Thank you. A job for when I have some time :)--Patrick Hearn (talk) 12:28, 18 January 2017 (UTC)

Locomotives hired to or visiting other railways

We have sections for locos coming in, are there equivalents for SVRSevern Valley Railway-resident locos going the other way? I can't find any. TIA--Patrick Hearn (talk) 12:28, 18 January 2017 (UTC)

Off the top of my head, no section exists yet. The exception would be for main line excursions, which are listed on the linked locomotive pages, but not many of those happen these days! --Danny252 (talk) 13:09, 18 January 2017 (UTC)
Some individual loco pages also mention longer term hires to other railways, but you're right in that there is no equivalent section for gala appearances elsewhere. No reason we couldn't have such a page, but it would be a major exercise to backfill the history.--Robin (talk) 15:50, 18 January 2017 (UTC)
Thanks both. Do you think we should we start one with 2017 galas out as an incomplete page, and allow people to backfill? Or let it ride?--Patrick Hearn (talk) 02:43, 19 January 2017 (UTC)
If you're eager, go ahead - just look how the gala visitors page started! I have to admit I don't really know where to find historical information; the SVRSevern Valley Railway News sometimes mentions locos going off on hire, but I don't know if this is comprehensive? Then again, I'm sure Robin will find all the information about 7714's visit to the Little Piddling and Nowhere Railway's 1976 gala... --Danny252 (talk) 10:44, 19 January 2017 (UTC)
It must have been there as a static exhibit - according to the SVRSevern Valley Railway Wiki it wasn't steamed until 1992. --Robin (talk) 13:10, 19 January 2017 (UTC)

Bug reports

Navbox error

I'm seeing {{#invoke:Navbox|navbox}} instead of Navboxes appearing on the carriage and loco pages. Anyone else? Can one of the administrators have a look please? --Robin (talk) 16:08, 16 May 2017 (UTC)

It was something I did without realising by altering parser settings while trying to convert page names to dates and vice versa. It should be working again now. --Graham Phillips 110 (talk) 16:48, 16 May 2017 (UTC)
Yes fixed. Belated thanks.--Robin (talk) 16:18, 24 May 2017 (UTC)

On this day refresh error

Another oddity. When I first access the home page without being logged in, the "On this day" link shows May 18. As soon as I log in, it refreshes correctly. Any ideas? Is it some sort of caching issue? --Robin (talk) 16:18, 24 May 2017 (UTC)

I was getting that as well. It was caused by {{#time:}} using "the time the page was last rendered into HTML" rather than the current time. For some reason, it would always default to May 18, even immediately after logging out after displaying the correct date while logged in. I've changed it now to read {{CURRENTMONTHNAME}} and {{CURRENTDAY}} which do exactly what their names suggest. It seems to have solved the problem on the wiki itself, although I don't know if there may still be a caching problem for individual users. --Graham Phillips 110 (talk) 00:56, 25 May 2017 (UTC)
I'm aware that it is still not working as it should. Editing the page also updates the displayed date, so the error doesn't show up until a day later, which makes it a slow job fixing it by trial and error. I've asked for help on the Mediawiki Project Support Desk--Graham Phillips 110 (talk) 08:58, 27 May 2017 (UTC)
I think it's working now. If you want an explanation...The various Time and Date functions use the time the page was last rendered in to HTML. For a logged in user, that will be their current session, for a non-logged in user, it is the time the page was last modified, as shown at the bottom of every page. By linking to the internal On This Day page as if it was an external link, it seems to be creating the link using the real current time, even when logged out.--Graham Phillips 110 (talk) 22:52, 28 May 2017 (UTC)
Another minor bug fixed. As the wiki was seeing the On This Day page as an external link, it was opening in a new tab, rather than the same tab like all other internal links. I couldn't fix it properly, so the external link is now displayed as a single full stop right at the bottom of the page, I couldn't make it completely invisible. Just having the link somewhere on the page forces all dates to work correctly, so now the internal link works as it should. There's another minor bug in that the wiki is set to GMT and is not adjusting for BST. I might look in to that one at some point as well. --Graham Phillips 110 (talk) 07:25, 29 May 2017 (UTC)
I really think I've got it this time. What was confusing me was the random image on the Main Page Test page which was causing that page to refresh every time it was viewed, whether logged in or not, which the proper Main Page wasn't doing. I've now set up the Main Page to automatically refresh every time. I thought this might slow down loading, but it doesn't seem to make any difference. What was slowing everything down was fetching nine individual images from Wikimedia Commons for the schematic map. I've now duplicated these locally, so not only does the page load a lot faster, logging in is a lot quicker as well.--Graham Phillips 110 (talk) 14:55, 2 June 2017 (UTC)
Thanks for spending the time on it! It would be an idea to update the pre-1963 map page to use local icons as well; it can be very hit and miss loading at times. (Not meant as a request, happy to do it myself when time permits).
Probably a bit late to ask this as were are closing in on 100 individual Day pages, but is there anything in them which could be better handled in a template called to each page, rather than the block of text we copy and paste each time?--Robin (talk) 15:32, 2 June 2017 (UTC)
I'd already started on the other maps when you replied. Consider it done. Yes, now I know how easy it is to create and call a template, that would be a better way of getting the forward and back dates. I did wonder if it might be better if the links just said "Previous" and "Next" rather than the date, but that would mean going back and altering all the existing ones for consistency.--Graham Phillips 110 (talk) 16:49, 2 June 2017 (UTC)
I thought it had been working fine, but today it is linking to June 20 when not logged in, even though a page exists for June 25. Not sure if it is a coincidence, but the home page was last edited on June 20. Has anything changed in the back end, apart from switching Lingo back on? --Robin (talk) 19:58, 25 June 2017 (UTC)
Another oddity with 'On this day in history'. When logged out, the next and previous page links on 'August 4' and other days works fine. When I log in they stop working, replaced by a grey box with a dash in. Not sure how long it has been happening - anyone else getting the same and if so, any idea why? --Robin (talk) 14:37, 4 August 2017 (UTC)
It took a bit of trial and error, but I've finally got the Previous & Next links working. It turns out it needs spaces either side of the + & - signs when searching for Pagename + 1 day etc. I've got no idea why it worked when logged out, but not when logged in, or why it used to work and then stopped working, but it works now. I'm still stuck on the out of date On This Day link though. It looks like we need a bot to automatically edit the page at 00:01 every day, but creating a bot is beyond me at the moment.--Graham Phillips 110 (talk) 09:44, 6 August 2017 (UTC)

Display error using references

There appears to be a display error in our implementation of Wikimedia, as per the testing I did on my user page. Normal Wikipedia practice is that a reference should follow punctuation marks (and if you do otherwise it will be corrected by a patrolling Bot!). However with an indented paragraph (and I think also a bulleted one), the next line does not display correctly when you save the page unless the punctuation follows the reference (it is fine in preview mode which is a right pain!). As a work round I will therefore do this only where necessary.--Robin (talk) 16:45, 6 June 2017 (UTC)

In short, I have no clue as to the cause... I've thrown it over to the Mediawiki support desk. --Danny252 (talk) 20:22, 6 June 2017 (UTC)
The smart people over there identified this as being caused by a conflict with another extension. A bit of trial-and-error reveals it's the Lingo extension, which is what we use to show the glossary info (the little explanations for BRBritish Rail or British Railways, GWRGreat Western Railway, etc.). I've disabled that extension for the time being. As for a full solution with both enabled, I think the only option is to finally upgrade from version 1.24, as updated versions of Lingo require higher versions of Mediawiki... --Danny252 (talk) 13:34, 7 June 2017 (UTC)
Since the Lingo extension was disabled, I'm finding a lot of pages with the glossary text rendered as per my earlier note below, such as 'OW&WOxford Worcester and Wolverhampton Railway' on The Severn Valley Railway under GWR/BR ownership page for example. I can reset individual pages by adding ?action=purge to the URL, but is there a way to refresh all pages? --Robin (talk) 21:22, 7 June 2017 (UTC)
Yes, there's a function in the back end to invalidate all cached pages - hopefully they should all reset when next viewed. --Danny252 (talk) 11:18, 8 June 2017 (UTC)

Mobile support

I've added the extension suggested to me on my talk page to add much better support for mobile. At a quick glance it appears to work, but keep an eye out for anything which doesn't behave.

I've not touched the extra features. Perhaps someone wants to investigate the features like the "Nearby" location functionality, so that they can pull up the Arley page when sat at Arley - it sounds like the sort of thing Graham and Robin would enjoy messing around with! Link to documentation. --Danny252 (talk) 22:02, 6 June 2017 (UTC)

I looked at the mobile home page briefly this morning on my android tablet. One obvious error was that it was rendering the glossary info within the text, so SVRSevern Valley Railway appeared as SVRSevern Valley Railway for example. Maybe the same version conflict you mention on the reference error above? Also the map of the line had 'breaks' in it, but I've seen that on the mobile version of Wikipedia too.
Any pages I linked to were still in the Desktop layout, so only the home page was in mobile mode. Is the intention that all pages should be mobile enabled? --Robin (talk) 15:42, 7 June 2017 (UTC)
I did see the breaks in the map, which I think would be non-trivial to fix. The glossary info is likely the other item you mentioned, which should now be fixed.
I've noticed that linked pages are showing as non-mobile, but I'd had no luck working out why. However, the act of invalidating the cache and forcing pages to be regenerated to fix the glossary problem also seems to have fixed that! --Danny252 (talk) 11:22, 8 June 2017 (UTC)
I think the breaks in the line were caused by variations in the way text and images get resized. At certain sizes, the height of the images was less than the height of a line of text. I've created a single image of the whole map, including text, now. Images of text don't always resize well, but it seems OK and readable at all sizes to me, on PC and mobile. Let me know if it's not for you. I'll do the other maps soon.--Graham Phillips 110 (talk) 22:51, 9 June 2017 (UTC)
Browsing with a desktop PC (Win 8.1 + Chrome) while not logged in, an increasing number of pages are appearing in mobile mode, with text collapsed below heading level. Logging in automatically resets them to desktop mode, logging out reverts them to collapsed mobile mode. There is no pattern as to which pages that I can see; I wondered if they have been cached according to the last user's preferences? Are there any settings in the mobile extension to control how pages should load depending on the OSOrdnance Survey/browser, and/or on linking from page to page, or is it random? Anyone else seeing the same issues? --Robin (talk) 22:24, 13 June 2017 (UTC)
I've set $wgInvalidateCacheOnLocalSettingsChange in Local Settings to True which forces all pages to be regenerated and seems to have fixed it for now, although it may come back as pages get cached if it is caused by the last user's preference.--Graham Phillips 110 (talk) 22:27, 14 June 2017 (UTC)

Database error

Just made a couple of edits which appear to have updated, but each time I have arrived at a page saying "Database error. A database query error has occurred. This may indicate a bug in the software." --Robin (talk) 13:50, 9 June 2017 (UTC)

Looks like Graham forgot to do something when installing the geolocation whatsit: "Fri Jun 9 14:21:36 UTC 2017 [...] Table 'mw_geo_tags' doesn't exist (127.0.0.1)". I've turned it off for now... --Danny252 (talk) 14:29, 9 June 2017 (UTC)
Yes, message gone thanks. --Robin (talk) 15:39, 9 June 2017 (UTC)

MediaWiki updated

The Wiki has now been updated to version 1.28.2. The update should mostly be transparent, with few changes to use. The main changes I am aware of are:

  • Password requirements are now tighter (min 8 characters).
  • Live Preview when editing is possible.
  • Recent Changes has been spruced up, with features such as grouping multiple edits together.
  • The mobile skin now appears to be working correctly, with pages generating the correct version.
  • For those tinkering with the backend, note that the directory and database for the Wiki have both changed.

Consult the changes for versions 1.25, 1.26, 1.27 and 1.28 for more detailed info.

Also, please grumble loudly if something is broken! --Danny252 (talk) 11:02, 19 June 2017 (UTC)

Thanks. Can we try enabling the Lingo extension again to see if it now works without breaking formatting (see above)? --Robin (talk) 13:07, 19 June 2017 (UTC)
The Lingo issue still exists, so I've put it in as a bug report for the Lingo extension. --Danny252 (talk) 20:17, 20 June 2017 (UTC)
I'm seeing an intermittent issue with media viewer when not logged in. For example on the Bridgnorth page, clicking the Bridgnorth Platform 1969 gallery image gives an error message Sorry, the file cannot be displayed. There seems to be a technical issue. You can retry or report the issue. Error: could not load image from http://www.svrwiki.com/mediawiki-1.28.2/images/8/82/Bridgnorth_Platform_1969.jpg. At the same time, Chrome is flagging up 'This page is trying to load scripts from unauthenticated sources'. Is it because we have changed to a secure site but the viewer is trying to load an unsecured page? --Robin (talk) 10:10, 21 June 2017 (UTC)
In short, I think yes - the HTTPS link to the image works, but the HTTP does not. That seems to be an issue with both Mediawiki (I should configure it to give HTTPS links), but also I don't understand why the image link redirects improperly. Will investigate... --Danny252 (talk) 13:57, 21 June 2017 (UTC)
Actually, forcing the HTTPS link generation might have been a simple one line change - does it behave now? --Danny252 (talk) 14:01, 21 June 2017 (UTC)
I'll keep an eye on it but from a quick look it seems to be OK. Thanks for sorting it. --Robin (talk) 15:27, 21 June 2017 (UTC)