Changed location of count.min.js?

[betaveros]

It looks like the old site code used a script at //static.goatcounter.com/count.min.js, but when I check my site code now it uses //gc.zgo.at/count.min.js and the old script URL is 404ing. Was this intentional? Was there some kind of notification about the change? I think a bunch of recent stats are missing from my sites because I didn't realize the site code had changed and didn't update my static sites...

[arp242]

The intended behaviour is that both work; I’m not sure why the old one stopped working. Unfortunately my laptop broke last weekend and I’m currently in the process of getting a new one, so can’t investigate much further at the moment 😕

…

On Mon, Dec 2, 2019, at 12:04, Brian Chen wrote: It looks like the old site code used a script at `//static.goatcounter.com/count.min.js`, but when I check my site code now it uses `//gc.zgo.at/count.min.js` and the old script URL is 404ing. Was this intentional? Was there some kind of notification about the change? I think a bunch of recent stats are missing from my sites because I didn't realize the site code had changed and didn't update my static sites... — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#71>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAH4D5AJWYGSEF7YXDYP2FTQWSCLLANCNFSM4JTP3XUQ>.

[betaveros]

Weird, it's back up now... I'll leave this open in case it's something you can investigate later or something somebody else notices, but I guess it's resolved.

[arp242]

Yeah, I fixed it yesterday: 16af9c8 – I had a lot of stuff going on and didn't have time to update the issue.

The problem was that I updated the static host to gc.zgo.at in the config, and that caused it to stop serving the old static.goatcounter.com :-( You should probably use the new domain btw, since it'll be faster for visitors.

Getting better testing for these kind of end-to-end issues is one of the major things I'd like to do before "1.0" and removing the beta label.

[arp242]

Oh, and the reason this wasn't noticed immediately is because the Varnish cache has an expiry of 30 days, so it got served by that for a month before it started erroring out. That's also why I thought it was working.

I set up some monitoring with uptimerobot today, so errors like this should be caught immediately in the future.

[CanRau]

Is it intended that gc.zgo.at/count.min.js redirects to un-minified gc.zgo.at/count.js ?

[arp242]

Yeah, that's intentional. Minifying isn't really worth it in bytes saved after gzip. It's just a compatibility thing.

[CanRau]

Okay got it 👍 just found the answer in the Telegram group 😁