Production mode pushRoute issue

I am really stumped. When certain conditions exist, Router.pushRoute runs without doing anything, without any console errors appearing.

All of these are conditions must be met to see the issue:

• Running a Next.js build in production mode. Dev mod is fine in all browsers.
• Cache must be cleared. If you have visited the page you are trying to route to, it will work.
• Firefox or Safari. Chrome is fine.
• For Safari, an additional condition is required: It must be a deployed live site (I’m using Now), all the above conditions work locally.

I am using pushRoute via fridays/next-routes inside an Apollo mutation callback. I have console logged everything ok up until the Next.js router is meant to do it’s thing here. I am pretty confused how to debug from there.

If you manually click a link to the same route in the site nav because you are sick of waiting, it transitions fine.

Seen in 3.0.0-beta10 and 3.0.0-beta11.

[gragland]

@jaydenseric I’ve been having a similar issue which I believe is caused by the same bug. In production if I click a link (created with <Link prefetch>) right after the page loads RouteChangeStart will fire and then nothing happens. The next time I click a link onRouteChangeError fires, just like you describe. Because this is intermittent my guess is that it’s only happening if the prefetch has not yet completed before the link is clicked.

[jaydenseric]

Further clues.

Ruled out it having anything to do with query batching.

Interestingly, when replacing:

Router.pushRoute('user', { userId })

With:

window.location.href = `/users/${userId}`

It works. Something silently breaks within the imperative Router API.