Traksy Website

[MrJeeves]

I can't reproduce that in any of my browsers. Is it on an iPad by any chance?

Chrome on Windows!

It doesn't seem to happen now... I did have a really bad connection at the time (and that was causing the autocomplete list to just show "undefined" until the suggestions had been loaded), so maybe that had something to do with it. I've not snooped about in F12 enough yet to have a guess though.

 

[skyhigh]

I've got this fixed on the maps but it will probably not release for a week or so

Thanks, Traksy is a great resource.

 

[MrJeeves]

Okay, well I said that and it happened again...

When I search "Lud", every result is map ID 1. I assume these should probably be filtered out if not found on a map, or "disabled" in some sense so you can tell it doesn't just not recognise what Ludlow is, but that there is no map for it.



I think clicking on any item in the list (which is what I was initially doing) will always select the first item, and as the first item in my list of results was Cardiff Bus it went to map #1.

Also, when there are zero results for a search, it pops up with "undefined" instead Just JS things.



Is the option to pin a train to view gone/broken? I see if you click a train it appears in the top-left with a pin button, but the pin button doesn't seem to do anything. It does let you view the same train multiple times if you pin it, though:



The "Show on map" button which appears if you expand the service info just seems to collapse the service info rather than jumping to the service's position on the map. The same occurs if I tab to the button when it's on screen and press space/enter, too.

I'll lay off the QA for now!

 

[takno]

Also, when there are zero results for a search, it pops up with "undefined" instead Just JS things.

Funnily enough it's almost entirely Rust which decides what goes in that box. I've written enough JS for one lifetime, so I'm moving as much of the model side of things as possible into webassembly.

Is the option to pin a train to view gone/broken? I see if you click a train it appears in the top-left with a pin button, but the pin button doesn't seem to do anything. It does let you view the same train multiple times if you pin it, though:

Sounds like it's working, albeit that it probably shouldn't open up the same train again underneath a pinned train. It's supposed to actually follow through from map to map on the new version, but I've got to admit it's a little patchy whether it works or not at the moment

The "Show on map" button which appears if you expand the service info just seems to collapse the service info rather than jumping to the service's position on the map. The same occurs if I tab to the button when it's on screen and press space/enter, too.

Yes, that button is currently non-functional. It's on the list just as soon as I get through the hellish chunk of work I wish I'd never started around mapping the routes and signal aspects

I'll lay off the QA for now!

The QA is all welcome!

 

[duffield]

One minor issue I've encountered on two different makes/models of Android phone is that when you select a headcode and scroll to the bottom of the pop-up details, the last couple of lines of locations/times are covered by the Android navigation bar. This happens both with Chrome and Firefox.

 

[BRX]

One minor issue I've encountered on two different makes/models of Android phone is that when you select a headcode and scroll to the bottom of the pop-up details, the last couple of lines of locations/times are covered by the Android navigation bar. This happens both with Chrome and Firefox.

I experience similar - and it means I can never press & use the "track train" button when I'm using my phone.

 

[duffield]

I experience similar - and it means I can never press & use the "track train" button when I'm using my phone.

Oh yes! I'd actually forgotten that button even existed since I hardly ever use traksy on anything other than my phone.

 

[takno]

Oh yes! I'd actually forgotten that button even existed since I hardly ever use traksy on anything other than my phone.

This is all interesting stuff. That used to work, and I think it still does on the phone I replaced a couple of months ago, but it doesn't work on my new phone. It is still fine if you install the site to homescreen, so if you're looking for workarounds I'd suggest going with that.

The new version of the site should be a bit better because it's much simpler and doesn't try to get so clever about the height of the page. A lot of the hacks in the old version were because you had to work hard to get Safari on iPhones to give you decent-sized window.

 

[duffield]

This is all interesting stuff. That used to work, and I think it still does on the phone I replaced a couple of months ago, but it doesn't work on my new phone. It is still fine if you install the site to homescreen, so if you're looking for workarounds I'd suggest going with that.

The new version of the site should be a bit better because it's much simpler and doesn't try to get so clever about the height of the page. A lot of the hacks in the old version were because you had to work hard to get Safari on iPhones to give you decent-sized window.

Thanks, installing it worked.

 

[takno]

Okay, well I said that and it happened again...

When I search "Lud", every result is map ID 1. I assume these should probably be filtered out if not found on a map, or "disabled" in some sense so you can tell it doesn't just not recognise what Ludlow is, but that there is no map for it.

View attachment 163245

I think clicking on any item in the list (which is what I was initially doing) will always select the first item, and as the first item in my list of results was Cardiff Bus it went to map #1.

Also, when there are zero results for a search, it pops up with "undefined" instead Just JS things.

View attachment 163246

I'll lay off the QA for now!

Search should be working properly now, including excluding locations which aren't on a map, having a smarter error message, being clickable with the mouse, not having duplicates, and having a friendlier "no results" message.

 

[MrJeeves]

Search should be working properly now, including excluding locations which aren't on a map, having a smarter error message, being clickable with the mouse, not having duplicates, and having a friendlier "no results" message.

Much better now! Thanks!

I have just noticed that I'm not seeing any headcodes on the new map, though, in case you weren't aware




Ah, may have been a dodgy old service worker on my browser as a clear of local website data seems to have sorted it for now on my laptop while it's broken on my phone still.

Hm, maybe not. This time it's broken again with this in the console...

Have fun!

wasm.js:460 panicked at src/train_renderer.rs:96:36:
attempt to add with overflow

Stack:

Error
    at https://traksy.uk/wasm/wasm.js:466:21
    at logError (https://traksy.uk/wasm/wasm.js:200:18)
    at imports.wbg.__wbg_new_abda76e883ba8a5f (https://traksy.uk/wasm/wasm.js:465:66)
    at wasm.wasm.console_error_panic_hook::Error::new::hb30632fa4f9b16f4 (https://traksy.uk/wasm/wasm_bg.wasm:wasm-function[2172]:0xef89e)
    at wasm.wasm.console_error_panic_hook::hook_impl::heb2fd7fc466168c1 (https://traksy.uk/wasm/wasm_bg.wasm:wasm-function[558]:0xa1f66)
    at wasm.wasm.console_error_panic_hook::hook::hf15902184706626b (https://traksy.uk/wasm/wasm_bg.wasm:wasm-function[2826]:0xfa4af)
    at wasm.wasm.core::ops::function::Fn::call::h762bd9009ae81c6f (https://traksy.uk/wasm/wasm_bg.wasm:wasm-function[2482]:0xf574a)
    at wasm.wasm.std::panicking::rust_panic_with_hook::h6731baa78621a747 (https://traksy.uk/wasm/wasm_bg.wasm:wasm-function[1111]:0xc9c5c)
    at wasm.wasm.std::panicking::begin_panic_handler::{{closure}}::hb6cd8464ed39ae71 (https://traksy.uk/wasm/wasm_bg.wasm:wasm-function[1475]:0xdab7b)
    at wasm.wasm.std::sys_common::backtrace::__rust_end_short_backtrace::hbdf3ddeb21a1e747 (https://traksy.uk/wasm/wasm_bg.wasm:wasm-function[3231]:0xfde99)


(anonymous) @ wasm.js:460
logError @ wasm.js:200
imports.wbg.__wbg_error_f851667af71bcfc6 @ wasm.js:454
$console_error_panic_hook::error::h480c5d5301323015 @ wasm_bg.wasm:0xe0c16
$console_error_panic_hook::hook_impl::heb2fd7fc466168c1 @ wasm_bg.wasm:0xa2052
$console_error_panic_hook::hook::hf15902184706626b @ wasm_bg.wasm:0xfa4af
$core::ops::function::Fn::call::h762bd9009ae81c6f @ wasm_bg.wasm:0xf574a
$std::panicking::rust_panic_with_hook::h6731baa78621a747 @ wasm_bg.wasm:0xc9c5c
$std::panicking::begin_panic_handler::{{closure}}::hb6cd8464ed39ae71 @ wasm_bg.wasm:0xdab7b
$std::sys_common::backtrace::__rust_end_short_backtrace::hbdf3ddeb21a1e747 @ wasm_bg.wasm:0xfde99
$rust_begin_unwind @ wasm_bg.wasm:0xf70ad
$core::panicking::panic_fmt::h5c7ce52813e94bcd @ wasm_bg.wasm:0xf9dff
$core::panicking::panic_const::panic_const_add_overflow::hde776086e9d58b0f @ wasm_bg.wasm:0xfbb54
$wasm::train_renderer::to_report::h34b487aad595e329 @ wasm_bg.wasm:0x82bd0
$wasm::train_renderer::to_occupier::hd5a5a9ba360397aa @ wasm_bg.wasm:0x3429a
$wasm::map_model::MapModel::broadcast_update::h3e2b324791137a39 @ wasm_bg.wasm:0x36aea
$wasm::map_model::MapModel::handle_update_string::hbdd7dc9ded89ed38 @ wasm_bg.wasm:0x39803
$mapmodel_handle_update_string @ wasm_bg.wasm:0xb37ec
handle_update_string @ wasm.js:358
connect @ traksy-map.js:148
await in connect
runt @ map.js:164
await in runt
run @ map.js:122
(anonymous) @ map.js:195
onLoginChange @ keycloak.js:149
await in onLoginChange
keycloak.onAuthSuccess @ keycloak.js:114
authSuccess @ keycloak.js:1403
req.onreadystatechange @ keycloak.js:1380
XMLHttpRequest.send
processCallback @ keycloak.js:1389
messageCallback @ keycloak.js:814
postMessage
(anonymous) @ VM284 silent-sso.html:2
wasm_bg.wasm:0xfdeaf Uncaught (in promise) RuntimeError: unreachable
    at wasm.wasm.__rust_start_panic (wasm_bg.wasm:0xfdeaf)
    at wasm.wasm.rust_panic (wasm_bg.wasm:0xfdc69)
    at wasm.wasm.std::panicking::rust_panic_with_hook::h6731baa78621a747 (wasm_bg.wasm:0xc9c89)
    at wasm.wasm.std::panicking::begin_panic_handler::{{closure}}::hb6cd8464ed39ae71 (wasm_bg.wasm:0xdab7b)
    at wasm.wasm.std::sys_common::backtrace::__rust_end_short_backtrace::hbdf3ddeb21a1e747 (wasm_bg.wasm:0xfde99)
    at wasm.wasm.rust_begin_unwind (wasm_bg.wasm:0xf70ad)
    at wasm.wasm.core::panicking::panic_fmt::h5c7ce52813e94bcd (wasm_bg.wasm:0xf9dff)
    at wasm.wasm.core::panicking::panic_const::panic_const_add_overflow::hde776086e9d58b0f (wasm_bg.wasm:0xfbb54)
    at wasm.wasm.wasm::train_renderer::to_report::h34b487aad595e329 (wasm_bg.wasm:0x82bd0)
    at wasm.wasm.wasm::train_renderer::to_occupier::hd5a5a9ba360397aa (wasm_bg.wasm:0x3429a)

 

[takno]

Much better now! Thanks!

I have just noticed that I'm not seeing any headcodes on the new map, though, in case you weren't aware

View attachment 163483
View attachment 163484

Ah, may have been a dodgy old service worker on my browser as a clear of local website data seems to have sorted it for now on my laptop while it's broken on my phone still.

Hm, maybe not. This time it's broken again with this in the console...

Have fun!

wasm.js:460 panicked at src/train_renderer.rs:96:36:
attempt to add with overflow

I've deployed a fix for this, as well as some new route mappings around Reading, Bristol and Crewe.

 

[Benjwri]

Have been really enjoying the beta website, my one big bit of feedback is the disappearance of being able to select a schedule when it doesn't find one, and also being able to see other schedules under the same head code when it has auto selected one. Especially in the Waterloo area where head codes are commonly repeated it has a tendency to pick up a head code for a different service.

Also a minor bug, not sure if you're aware (Possibly you haven't got to these areas) but the head codes in the berths in the departures on Paddington platforms and across the whole EL CCOS seem to be a bit 'sticky', in that they won't clear after the train is long gone, until another train enters the berth (Or not even that in some Paddington berths) (In the Paddington screenshot the lower numbered trains have all long since arrived at their destination away from London)

 

[takno]

Have been really enjoying the beta website, my one big bit of feedback is the disappearance of being able to select a schedule when it doesn't find one, and also being able to see other schedules under the same head code when it has auto selected one. Especially in the Waterloo area where head codes are commonly repeated it has a tendency to pick up a head code for a different service.

Also a minor bug, not sure if you're aware (Possibly you haven't got to these areas) but the head codes in the berths in the departures on Paddington platforms and across the whole EL CCOS seem to be a bit 'sticky', in that they won't clear after the train is long gone, until another train enters the berth (Or not even that in some Paddington berths) (In the Paddington screenshot the lower numbered trains have all long since arrived at their destination away from London)

Thanks for the feedback. Getting access to the schedules by headcode or schedule id is next on the list, so hopefully I'll be able to add the "pick from a list" function back in at that point.

I was noticing the sticky headcodes earlier. It looks like it's only in places where the berth has a number of sub-berths which may be stepped through quite quickly. The updates sometimes get bunched together, and if the new system gets a bunch of updates all together it will combine them all into a single change, rather than drawing all of them one after another. I suspect when it bunches up the berth steps it's doing it wrong. I'll take a look soon.

I've added a theme picker incidentally in the user menu, which allows you to override your system default light or dark theme. Can't remember who requested it, but it makes my testing a bit easier anyway.

 

[Benjwri]

Thanks for the feedback. Getting access to the schedules by headcode or schedule id is next on the list, so hopefully I'll be able to add the "pick from a list" function back in at that point.

Great, thank you!

I was noticing the sticky headcodes earlier. It looks like it's only in places where the berth has a number of sub-berths which may be stepped through quite quickly. The updates sometimes get bunched together, and if the new system gets a bunch of updates all together it will combine them all into a single change, rather than drawing all of them one after another. I suspect when it bunches up the berth steps it's doing it wrong. I'll take a look soon.

That does make sense, given how Paddington Berths work when a train departs, and the nature of the CCOS.

 

[MrJeeves]

I've added a theme picker incidentally in the user menu, which allows you to override your system default light or dark theme. Can't remember who requested it, but it makes my testing a bit easier anyway.

Oooh, thanks! I didn't request it either, but I like it nonetheless :P

Any chance you could make it toggle the page background, too? For me it only appears to affect the diagrams themselves.

 

[takno]

Oooh, thanks! I didn't request it either, but I like it nonetheless :P

Any chance you could make it toggle the page background, too? For me it only appears to affect the diagrams themselves.

It should be changing almost everything on the page for the new map pages. The only thing I'm unsure about at the moment there is the colour of the search box, which has been left as the system default. I will extend it to the information pages at some point, but pages like subscription management and account management might be a bit harder to do.

I did only do it in about 40 minutes last night though, with no testing on mobile, so if there's something else which isn't changing then please do share a screenshot and I'll try to figure it out.

 

[MrJeeves]

This is the behaviour I see with Chrome on Windows, at least.



It looks like the .dark-override class has background: #fff

 

[ainsworth74]

I take it that these smarter looking maps (not that the existing maps are bad!!) are what you get by subscribing? How widespread are they? For instance is the ECML and North East (where mapping is possible) covered?

 

[Benjwri]

I take it that these smarter looking maps (not that the existing maps are bad!!) are what you get by subscribing? How widespread are they? For instance is the ECML and North East (where mapping is possible) covered?

It’s access to the beta of the new maps you get by subscribing. I haven’t found anywhere not covered, but some maps just have minor weird behaviours or a few berths out of place, presumably where they haven’t been updated. ECML is definitely covered and seems to work fine.

 

[ainsworth74]

It’s access to the beta of the new maps you get by subscribing. I haven’t found anywhere not covered, but some maps just have minor weird behaviours or a few berths out of place, presumably where they haven’t been updated. ECML is definitely covered and seems to work fine.

Thanks

 

[MrJeeves]

I take it that these smarter looking maps (not that the existing maps are bad!!) are what you get by subscribing? How widespread are they? For instance is the ECML and North East (where mapping is possible) covered?

Takno previously said every map is available in the new format — my assumption is that it's the same signalling berth map, but with much more accurate mapping of routes and signalling data.

 

[takno]

I take it that these smarter looking maps (not that the existing maps are bad!!) are what you get by subscribing? How widespread are they? For instance is the ECML and North East (where mapping is possible) covered?

It's the same range of maps - they're generated from the same source files. The difference is that the new maps can contain more "things", like signals pointing diagonally, single/double slips, track names and the like.

The information is mostly still shown on the old maps, but in a messier way, with diagonal signals pointing off to one side, or slips represented by thin text dashes.

The route data is better, partly because the old system has a bug I can't fix in it which leads to random routes being shown, and partly because I've developed a better way of creating route tables which has allowed me to add in a few more areas which weren't really covered before.

 

[Benjwri]

By the way do you prefer bugs being reported here or by email? Have sent a few specific map issues by email recently.

 

[OliverH68]

I've noticed on the Chiltern line map some platform numbers need swapping round.

Gerrards Cross, Haddenham & Thame, Bicester North & Village, Kings Sutton need to be swapped round.

Heyford also doesn't appear to have any platform numbers shown.

 

[takno]

By the way do you prefer bugs being reported here or by email? Have sent a few specific map issues by email recently.

I've noticed on the Chiltern line map some platform numbers need swapping round.

Gerrards Cross, Haddenham & Thame, Bicester North & Village, Kings Sutton need to be swapped round.

Heyford also doesn't appear to have any platform numbers shown.

In general, if you email it it is more likely to get fixed at some point, rather than missed, but if you post it here and it's something that appeals to be then it's slightly more likely to grab my attention.

I do care about platform numbers and small inaccuracies, but they aren't very interesting to deal with, so I'd rather have them by email so that I can just build up a backlog and do a whole bunch at once.

 

[ainsworth74]

It's the same range of maps - they're generated from the same source files. The difference is that the new maps can contain more "things", like signals pointing diagonally, single/double slips, track names and the like.

The information is mostly still shown on the old maps, but in a messier way, with diagonal signals pointing off to one side, or slips represented by thin text dashes.

The route data is better, partly because the old system has a bug I can't fix in it which leads to random routes being shown, and partly because I've developed a better way of creating route tables which has allowed me to add in a few more areas which weren't really covered before.

Ah thanks for that explanation

 

[takno]

This is the behaviour I see with Chrome on Windows, at least.

It looks like the .dark-override class has background: #fff

View attachment 163528

Ahh, yes, had this backwards. It should be fine now, although you might need to do a hard refresh to get the new stylesheet

 

[OliverH68]

In general, if you email it it is more likely to get fixed at some point, rather than missed, but if you post it here and it's something that appeals to be then it's slightly more likely to grab my attention.

I do care about platform numbers and small inaccuracies, but they aren't very interesting to deal with, so I'd rather have them by email so that I can just build up a backlog and do a whole bunch at once.

No worries - I will email in

 

[JLX5]

I just subscribed to give the beta site a go, however cannot seem to get it to work. When putting in a station in the search bar and clicking on it, nothing happens even though the URL in the browser does change.
Am I doing something wrong?