I really like the simplicity of this and the approachability for students. If I were still teaching CS I’d be using it as part of the curriculum.

Do you support conditional GET AND PUT (i.e. eTags)?

Hey! Thanks for checking me out.

I do not currently support anything like that, however I think you bring up a very good point, and it could be very beneficial. One could however add their own lastModified variable as a part of the json blob they push up and implement the support themselves. From your teaching perspective of course, that may not be a great experience for the students.

Yeah, there are very few APIs that support conditional requests but they’re super important topics and it’d be great to have a way to demonstrate them.

Appreciate the insight, I'll consider it!

Security people: What are the odds this becomes part of the command and control chain for a botnet in the next 30 days?

Botnet maybe not, but as a place to store phished creds etc., quite likely.

Services like requestbin have had a lot of trouble with this sort of thing, historically https://web.archive.org/web/20190711192140/https://github.co...

> We have discontinued the publicly hosted version of RequestBin due to ongoing abuse that made it very difficult to keep the site up reliably. Please see instructions below for setting up your own self-hosted instance.

(note, that message is old, requestbin did live on despite it, iiuc)

So rate limiting on GET, then, too? How would you prevent abuse while keeping it simple?

That's the problem, we can't have nice things on the internet anymore.

Either you could try adding limitations that would make it less appealing to shady folks, e.g. amount of storage in the value.

Or most likely adding user accounts and depending on how far you need to go captchas or offload it by e.g. using GitHub oauth (and coupling yourself and your users to a 3P).

Neither of them are great. While the tool is fairly unpopular there probably isn't an issue.

Unfortunately, you spoke too soon, exactly what you said has happened lol.

I mean, there's a whole industry dedicated to preventing abuse of computing resources exposed to the internet. You basically opened your MongoDB instance to the world - you are lucky they haven't already stored anything horribly illegal in there.

Anything is certainly possible on the internet, if or when that happens I'll intervene.

Implementing authentication and or authorization is not beyond my ability, and could have been setup, but it would have severely tarnished ease of access to the platform, and hindered the amount of people who already tried out the service.

The hubris of software engineers never ceases to amaze me. I'm not trying to scare you here, I am just gently suggesting that you might not have the capacity to provide this as a public service. Keep using it for your projects if it works for you, maybe wrap it in a docker container for other people to deploy on their own if they so want.

Side note, I just tested a curl command with

  -d '{"key": {"$regex": ".*"}, "value": "..."}'

and it went through, updating the already existing first key in the collection, so you're not even sanitizing your inputs - there's likely some Mongo lookup code that will enumerate the existing collections, or overwrite someone else's data, or possibly read all of it, I am not going to continue with the probing (are you sanitizing the api key?). You can either assume good faith from all your users, or open the API to the internet, but definitely not both.

O0f. I could try putting up a no robots sign, maybe that will stop them

So what? I could C2 a botnet right now with base64'ed hacker news comments. You don't need a captcha to register an account.

HN has a system for reporting and responding to abuse.

Yeah I feel you I'm just trying to point out the fact that there are a lot of ways to read and write arbitrary data pseudo-anonymously.

As a one-off there are many ways, but if you want to scale things up then you run out of options pretty fast. OP potentially just made themselves the new lowest hanging fruit.

Good point.

I think it’s pretty cool. It’s got that HomeDepot tool aisle vibe of being appealing even if I don’t have an immediate use for it.

For a second I thought it would be nice to not auto append a UUID to the hash, but I can imagine it’s simpler not having to check for or manage collisions.

Agreed, I think a fuller version could require authentication and then you'd be able to have whatever name you want maybe?

Wouldn't it be more semantically meaningful if you would use "...api/store/<hash>" with GET and PUT? That way you would also get better caching.

I like your idea of switching the endpoint specifically for caching, would be a great improvement for the service. This does run on Next.js which offers some sort of caching out of the box. When it becomes a problem, it will definitely be necessary. Thanks for the suggestion!

Couldn’t this just as easily be a front end to an object store like s3 - without the need for mongodb?

Yes! This could be done in many different ways with different technologies.

Annnnnnnnd someone just infinitely created a bunch of empty hashmaps and maxed out my collections. :)

Thanks HN <3

Deleted 255 collections. Investigating throttling of some sort...

You probably should add some rate limiting on the API.

I added a simple rate limit last night, thanks for your suggestion! :)

I would love some feedback on the imagative ways the service has been abused since you created it.

So far the worst thing is that someone is trying to actively stop my service from being functional? Idk, I guess it's more fun than other things?

Anyways they mass created hashmaps using UUIDs. Someone lower in the comments made a great point on this, unfortunately this is why people have to crack down on things and require accounts and stuff like that before you can use a service.

:) Service is up and running, but we'll see for how long.

Great idea putting all actions behind a token. This removes most types of abuse.

Still, a great place to backup my video library!

I once thought about doing this using RDF/Tuple based data where most elements are forced to be UUID-4's (whose purpose and meaning are invisible) except for values. I imagined a public service without a token.

I would like to see being usable from a browser (with the key, value and token as query parameters):

https://hashmap.me?key=key1&value=value1&token=tony-test-d87...

Thanks so much for the suggestion and checking me out.

I think for Read it makes a ton of sense to be able to provide a key to only get that specific key from your dataset.

I could add support for allowing you to use your token in the url to avoid headers, but to be very clear that token should NEVER be exposed to other users or to the public, that token isn't changeable, and if another user has it they can modify your dataset. Maybe I could add read only tokens at some point?

With your exact URL above, are you proposing the ability to Write data from the browser? I see you're providing a value so it lead me to believe you're suggesting writing a key/value pair with your token from the browser

Hmmm, it looks like the post is flagged. I had to look up what that meant. I'm guessing I may be breaking the rule of including the website name in the title of the post. :(

How can I retrieve a single key? As far as I can see, the only read api available returns all keys.

Hey chaz! Currently not an option, and as I write that out, you're right that it should be an option.

Even if it was something like "/api/read?key=getOnlyThisKey" which returns your value or perhaps a 404 or maybe a 204, implying your response was succesful, but there was no key with the specified requested key.

[Nick Haflinger has entered the chat..]