Viewing UK reports page with 0 reports directly (via than ajax) can give error

[dracos]

e.g. Bristol closed reports. Due to default map co-ord (0,0) then trying to be converted to easting/northing somewhere.

[ludovic-tc]

Please could I have some repro steps for this bug @dracos?

[dracos]

tl;dr Looks like this was also reported at #2617 and already fixed there.

This bug was three years old, so I didn't know/remember anything about it :) For interest, here's the steps I took to discover the above fact, might be some useful tips in here:

1. I said Bristol closed reports, so I visited the Bristol cobrand, all reports, picked closed and ended up with the URL https://fixmystreet.bristol.gov.uk/reports/Bristol?sort=updated-desc&status=closed (I removed any lat/lon/zoom as otherwise those are taken as defaults) - it displayed okay. But it had reports, whereas this ticket said a page with 0 reports so I assumed back then it didn't have any.
2. I tried to find a URL that would have zero reports on load (as per the ticket, not via ajax), and luckily restricting to the first category in the filter list on the All Reports page worked: https://fixmystreet.bristol.gov.uk/reports/Bristol?sort=updated-desc&status=closed&filter_category=Abandoned%20vehicles - but again, the page still loaded fine.
3. I checked the source of that page and saw it did have e.g. data-latitude=0 so it had got the map co-ordinate mentioned. But it was not showing an error.
4. The ticket said the error was caused by lat/lon -> e/n conversion, so I found the function that does that conversion, convert_latlon_to_en, and looked for where that was used. Most uses were on a per-report basis, or during reporting, not relevant, but there were uses under FixMyStreet::Map::.
5. Two of them were CheshireEast and Northamptonshire, so not relevant to Bristol, but the other was FixMyStreet::Map::UKCouncilWMTS, and I saw that FixMyStreet::Map::Bristol was a subclass of that.
6. Looking for the use of the function in UKCouncilWMTS.pm, I saw the line above had special (0,0) handling!
7. I ran git blame to see when the fix was added, and tracked down the relevant GitHub ticket from that :)