BGood to know. Thanks 
BHi,
.js.jsUsing
rclone sync
Rno, or at least that is what I saw when it first launched, you have to invalidate the cache, or else, make immutable js files, depends on your usecase.
BThanks
BYou mean I can invalidate Cloudflare file cache version by calling an API?
SOne thing I don't understand is the pre-signed URLs is it ok to expose them on public apps?
- https://developers.cloudflare.com/r2/buckets/cors/#use-cors-with-a-presigned-url Here it says it's safe to use
Presigned URLs are only used for authentication, which means they are generally safe to distribute publicly without revealing any secrets.
- https://developers.cloudflare.com/r2/api/s3/presigned-urls/#presigned-url-use-cases but here it says that's not safe to share it publicly
It is useful to note that pre-signed URLs will expose your account ID and bucket name to whoever gets a copy of the URL. Public bucket URLs do not contain the account ID or bucket name. Typically, you will not share pre-signed URLs directly with end users or browsers, as pre-signed URLs are used more for internal applications.
Ssounds good thanks
KHello, I have one question, S3 have a class storage called intelligent tiering, that down grad objects class if it isn't accessed, R2 have something like it? I read the documentation and I not found if R2 have storage classes, I found some called classe A and classe B that refers to getObject, listObject, and put object, but nothing about classe storage.
JR2 announced infrequent access tier last week, but as far as I know it’s all manually controlled: https://blog.cloudflare.com/r2-events-gcs-migration-infrequent-access
DHello everyone, I have a quick question, my r2 object is public and can be accessible from anywhere but when I upload to an automated python job it gives me a 403 forbidden error, how can I make it fully accessible? I even tried making a get request with python code and it worked but only it doesn't work in one of the public automations
MCannot access r2 Custom Domains file on iPhone
Y
YWill there be any charges in the future if you don't use TURN?
Y
YWhere is the account and password?
YDoes it seem like there is no place for calls?
Eyou can use #stream or #developer-week-2024
ZThe difference is a bit pedantic, yea. If you compare R2 IA to something like GCP's Nearline, it's more clear why there is a difference. Nearline charges you for both retrieval and egress, while R2 only charges you for retrieval.
ZBut there is indeed still a charge / GB, just like egress charges
DApparently there is a bug in S3 in which the pre-signed PUT URLs ignore SHA256 Hash when uploading. https://github.com/aws/aws-sdk/issues/480
I just tested a pre-signed URL with R2 and see that it too also ignores the hash, and allows upload of arbitrary body payload.
I’m wondering if this was intentional or is a bug?
I just tested a pre-signed URL with R2 and see that it too also ignores the hash, and allows upload of arbitrary body payload.
I’m wondering if this was intentional or is a bug?
LIs it possible to use Sippy with a public bucket, without creds? It looks like that's not supported at the moment - is there a plan to add this? (I'm trying to use Sippy with a public bucket which I don't control. So not exactly the planned use-case but it's soo close.)
ZUse rclone
Lmy goal is for live 'proxying' not a one off copy
CHey if i have an svg string that Im trying to upload to an r2 bucket as a file. when i create a blob from the data and use the put() function from the workers api, it shows as no content type on the cloudflare website
Cis there a way to set this content type?
KWhen you put the object, you can specifiy the
KIt's usually better if you always explictly provide the type
Chow do i do that? i dont see anything in docs
KThe
Conly takes two parameters
Cwait no im wrong
Cmissed that part lol
Cthank you
Bdepends bro, you shouldn't completely share pre-signed url's without very specific constraints, such as, you have to predetermine the key of it, also the content-type, otherwise some mf can just upload an html file with a XSS script in it and if your main website doesn't have cookies your kind of doomed
Bpresigning get requests is crazy
Bnot in a bad way, but like weird cause its very resource intensive when majoriy of the files are supposed to be just GET requestable
Bhow exactly does it work for you? is it like a middleware between the file? like you use s3 ? or whats the architecture exactly
Byeah ok but if you have millions of requests per day its still a lot
Bwdym terribly implemented? you mean the guy that built the cloudflare version?
HMore like S3 itself
HS3 works, but XML? Ugh...
HThe only real big downside is that Pre-Signed URLs aren't cached
HFor unique files, it doesn't matter much
HBut if you are distributing something to a lot of people, but still need auth, then it can be a pain
HWith a header check?