UIt can cache by deploy id so if you do a new deploy those changes can be picked up rather than still having old cache because its just using the url for the cache
UIt will cache everything but the function requests afaik. And the function requests you can cache yourself with the workers cache api
UYeah you dont need to do that on pages any of the static assets (so in your case everything), will always be cached.
UYes. Rocket loader sometimes can have issues but that depends on the javascript on your site
Uand rocket loader having an issue wouldnt be pages specific either
CIf you go to your previous builds then the deployment logs will be there. I don’t know if the deployment logs expire but I don’t believe they do
KThey'll expire after a year iirc
YHi, currently evaluating Pages. I am getting 522 errors when layering Cloudflare for SaaS on top of Pages. I see that Pages has its own Custom domain support, but it has a limit of 500 domains. Is this a hard limit?
CThere is currently no build caching with pages but it is planned. No ETA though
DAnyone know of a simple way to do SSR with pages without using one of the frameworks? NextJS's edge runtime doesn't seem ready for use yet.
Wframeworks make it easier I'd say but yeah, anything like React has it's ways to do it
WI'd say frameworks are the easy way (and there's lots of them - Remix, Astro, Qwik, SvelteKit, etc.)
JNext.js' edge runtime is fine to use as long as you aren't using a library that depends on unsupported nodejs functionality 
SI am trying to use the access pages middleware gizmo, but the redirect for is bouncing me to
localhostThanks, this is just what I was looking for, will try it out! Are you liking it okay?
WHi everyone - we're exploring deploying our Remix app to CF Pages. We can't do the whole hook-up your repo and let CF do the magic because we're running a mono-repo with various projects in it, so we'd like to push it. I found somewhere that we can use wrangler, and wrangler is deploying the assets to the pages, but I must be missing something, because I see no workers defined and going to the preview URL gives me a 404.
OHi - I got a failure on pages deployment when it opts for v2 of the build system. When I run v1 all is good.
Apologies if someone has already raised this, but it's not immediately obvious why it isn't working beyond saying that webpack is not found
Here is the log error:
Apologies if someone has already raised this, but it's not immediately obvious why it isn't working beyond saying that webpack is not found
Here is the log error:
JBuild image v1 added the local
node_modules/.bin to PATH so this kinda thing worked, but was a little non-standard. Build image v2 no longer does this, so the recommended approach is via an npm run script, or npxIf you add
webpack --config webpack.config.js to an npm run script like build and then use npm run build
OThanks - that was what I tried first as it is added to a build script, but it just kept failing the build
Ohence the reason why it shows the build command as webpack --config webpack.config.js
DThanks, I was having some trouble with nextjs server components, but tried a fresh project and got a basic one working. Will keep exploring it.
MI'm banging my head against the wall trying to get a Pages function to be able to use a service binding locally. Both the pages application and the target bound worker are deployed and I have the service binding setup in the Pages wrangler.toml file as well but the binding is always undefined. Using wrangler v3x, with both the pages and worker running, do I need to pass a flag locally when starting up the pages application in order for the two to be able to communicate?
MI see flags for
pages dev for DO, KV, D1 and R2 but there's nothing in there for worker bindings.JI do not believe that's possible in dev today unfortunately. You can follow https://github.com/cloudflare/workers-sdk/pull/2162 for updates.
MYeah, I'm actually just reading up on that now. So no functional worker bindings from pages locally... ugh. Okay, thank you James.
Jmaybe check out this new HBO Max and Hono combination buy one of the Cloudflare employees.
https://github.com/yusukebe/hono-htmx
https://github.com/yusukebe/hono-htmx
GitHub
Hono+htmx stack. Contribute to yusukebe/hono-htmx development by creating an account on GitHub.
Kseems that pages-help is dead, maybe someone here knows it: what git command does Cloudflare Pages build system uses to clone the repo? I can't get access to the tag from the build script
CWhat build image are you on?
CMy guess is that is does a checkout of the latest commit which is included in the payload.
Kv1
KI'd love to see the exact command line to reproduce it locally and try to dig out the tag
Kgit is like perl - there's always more than one way to do it
MI don't think the exact command is logged anywhere, but based on what is logged, it looks like it's doing a fetch/checkout of the commit hash, and it's probably doing a shallow fetch at that to be efficient
MSo probably smth along the lines of
git fetch origin <hash> --depth 1 && git checkout <hash>MI don't think there's a clone command that'd do that in one shot, given it's doing it by hash instead of branch
MSo if you wanted to access tags etc., you probably just need to do a fetch first?
Kmost probably, yes. I try to repro this. My build in release mode checks the tag and it's not there (the tag is not on the branch being checked out) refuses to build because it's not a release then. If tag is there, it generates a version file which is included in the build and shown on the page.
Kjust to provide the context
Kthanks!
MYeah thats because the build logic isn't checking out a branch, it checks out the commit hash
