This is why a CMS should just be a folder of markdown
This is why a CMS should just be a folder of markdown
Syep unironically
HBut also, #off-topic (definitely also me doing it)
Sworkers + r2 perfect cms
E
RBringing it back on-topic, this is how I wound up at Astro on Workers forever ago. I basically had been building a KV CMS and then Astro fit the bill on workers/pages
Thi guys, i used to use this command "npx wrangler pages dev -- yarn dev" but its being deprecated and has started having resource locked errors.
New command might be something like "npx wrangler pages dev .output/public --live-reload" but this isn't ideal as I still have to run yarn generate after changes to templates in my nuxt project, as output/plublic is not automatically updated when templates change. This is hugely slowed down workflow so I'm wondeing what the better approach is
New command might be something like "npx wrangler pages dev .output/public --live-reload" but this isn't ideal as I still have to run yarn generate after changes to templates in my nuxt project, as output/plublic is not automatically updated when templates change. This is hugely slowed down workflow so I'm wondeing what the better approach is
IIWorkers Assets is the recommended way going forward.
E
thank you, it might beneficial to have a message on the dash when creating a pages projectSwell it will still work for ages thats why they dont want to do that i think
IYeah that's mainly why I said it's not "officially" deprecated; it's still usable, but likely won't receive any new improvements.
Esort of "end of life"
Stheres barely any features missing other than preview links for branches anyway
RI've given this a lot of thought, because I personally just lost a bunch of time due to the fact that we weren't looking in the /Public directory... and you just saved our butts, by the way.
Just the fact that building new products on it would be a bad idea (tm) is enough to warrant some kind of deprecation notice.
I said somewhere else, Python was deprecating Python 2 for over 10 years. It was important to make the notice, even if support would continue for a long time.
Just the fact that building new products on it would be a bad idea (tm) is enough to warrant some kind of deprecation notice.
I said somewhere else, Python was deprecating Python 2 for over 10 years. It was important to make the notice, even if support would continue for a long time.
RThis unofficial deprecation is equivalent to uncertainty in the platform and that is bad.
SI think a recommendation would be enough. but its a beta.
my guess? static assets stable with the next dev week
my guess? static assets stable with the next dev week
RYeah, realizing that Workers Assets is still beta, I'm guessing that is why there has not been much notice. But they went ahead and recategorized everything and began name changes, and changing the menu around is kind of a hugely alarming notice to almost every user that things are changing, and without canonical word, that is uncertainty and is patently bad.
RLike having to explain this all to investors yesterday morning was not fun.
Tguys any advance on how to adapt this command so that i dont have to manually rebuild output folder each time to reflect changes. Surely its possible
npx wrangler pages dev .output/public
npx wrangler pages dev .output/public
Rlike you want a watch style function or? I guess one option could be to literally use
watch
TWell it doesn't make sense just to watch the public folder because its not rebuild every time you change a template in a nuxt project, only when you run yarn generate.
Formerly you could run a proxy command as a flag on the wrangler command that ran yarn dev, but this isnt allowed in wrangler v4
"npx wrangler pages dev -- yarn dev"
Formerly you could run a proxy command as a flag on the wrangler command that ran yarn dev, but this isnt allowed in wrangler v4
"npx wrangler pages dev -- yarn dev"
Eis it possible to see the log of a deployed pages app?
Ei keep getting a 500 error and I don't know why 
Rmy understanding of the
I'm, personally, still unclear on what you need done that can't be done with a well-placed
--yarn devwrangler pages devI'm, personally, still unclear on what you need done that can't be done with a well-placed
watchCloudflare Docs
Create, develop, and deploy your Cloudflare Workers with Wrangler commands.
HIIRC they removed it because it had bad UX. It still couldn't run Workers, so all you would have is what
yarn devThe new recommendation once it lands in more frameworks is to use the new Vite Environments API, which means that frameworks can now run their dev server in
workerd
Wwrangler pages deployment tail
Si could see d1 replication next month on dev week. containers? maybe not
RWeird question... when making a new Workers project from a github repository, what if I don't want to deploy to workers.dev?
Reason being, it seems that tons of bots just watch for deployments to workers.dev and jump at the opportunity to ping your Worker. :/
Reason being, it seems that tons of bots just watch for deployments to workers.dev and jump at the opportunity to ping your Worker. :/
ZYou can disable your workers.dev domain before you deploy in your
wrangler.toml/jsonc
CBots should only be from the cert. issuance for *.workername.subdomain.workers.dev anyway, if you deploy and snipe off the workers.dev quick off, should avoid it
Cfor functions?
Cthere's frameworks which can emulate that
Cyea, eh, I imagine it's one of those things where the vast majority of people use frameworks anyway
I wonder how they're going to try to make that migrate to workers assets
I wonder how they're going to try to make that migrate to workers assets
Cthey said in the blog post:
We plan to bridge the gaps between Workers and Pages and provide ways to migrate your Pages projects to Workers.
Cidk how you'd migrate file path routing without having something to deal with it
Cmaybe just a compat framework or something lol
pages-legacy-routing
HYou know what would be a real cool and dumb feature for custom domains? Pre-warming. I.e., you activate the domain, issue the cert, but block every request for the first n-minutes
Clol I think you can already do that yourself
CI just delete every cert Workers Custom Domains issue before they get a chance to be issued, and either place them on 1st level or use a *.workers.domain.com wildcard acm cert
AHey like i have hosted my vite react app in cloudflare pages and honobackend in the cf workers and there is an issue like when i login the cookie is not present i chrome cookies but in request header it is availabe and also like when i try to logout the cookie is still on the request header and unable to logout on dev no issues. after some reserch i have found like if the front end and backend have different domains like "pages.dev" and "worker.dev" as browser will not save cookie . so is there is a fix for it like without custom domain on dev side the vite proxy works but on production it will not any fix for it
HI mean the opposite, like in cases where you do want the cert, but don’t want the requests
RI like this idea very much. I do think it could help stave off a lot of spurious traffic. I know that it will eventually just result in this pattern being learned, too, but I think its worth exploring
RI might be misunderstanding but this seems like a CORS issue?
I have setup a system that used a custom domain on Pages as the origin and had a Worker in front of it, but they had the same top level domain. I don't think it would have mattered, though, since IIRC even subdomains are normally considered separate domains unless you use CORS.
I have setup a system that used a custom domain on Pages as the origin and had a Worker in front of it, but they had the same top level domain. I don't think it would have mattered, though, since IIRC even subdomains are normally considered separate domains unless you use CORS.
Ai have fixed the cors issue and still this cookie issues persits as on login on application section of chrome doesn't include cookie but on request header will be having it and when logout nothing is happening
Aand no cors issue is showing on consolelog
NDoes anybody know the best way to send transactional emails from Cloudflare Workers?
NI see MailChannels, but it shows that it stopped sending from Workers last year.
JSendGrid, SES, Postmark, etc. Really anything except MailChannels