Just noticed this, and perhaps it's actually a browser thing - pending further investigation - but it looks like if you're requesting a historical file that doesn't exist, e.g. 1991-01-01.json, the proxy (presumably this is an Apache thing) 'corrects' the URL to 1999-01-01.json
Obviously it should give an error, otherwise you've got inaccurate data.
This started happening when the proxy started downloading, caching and serving the files statically, instead of grabbing them from github each time.
If this is an apache thing intended to prevent 404s, anybody know how to switch it off?
For anyone interested, it's an apache feature - mod.speling (yes, that's one 'L') which needs to be turned off to prevent spelling corrections on URLs.
Just noticed this, and perhaps it's actually a browser thing - pending further investigation - but it looks like if you're requesting a historical file that doesn't exist, e.g.
1991-01-01.json
, the proxy (presumably this is an Apache thing) 'corrects' the URL to1999-01-01.json
Obviously it should give an error, otherwise you've got inaccurate data.
This started happening when the proxy started downloading, caching and serving the files statically, instead of grabbing them from github each time.
If this is an apache thing intended to prevent 404s, anybody know how to switch it off?