Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Show HN: KeepThisSite (keepthis.site)
1 point by ceritium on Sept 26, 2023 | hide | past | favorite | 5 comments
KeepThisSite is a humble and opinionated bookmarking service. There are several similar services out there, some pretty mature, but only this one is mine.

I am building it because I want to test in a production environment things that I could not try in my daily job. Also, a bookmarking service is something that I would use daily, so I am the first user.

It is free, at least at the moment; probably, I would keep it free forever, at least a basic account. The most expensive feature currently is the full screenshot (which I can optimize) and the MHTL capture.

Some expensive features, like the full screenshot and the MHTML capture, would be only enabled for some users.

I have a long list of features to work on, starting with the UI (browser and edit tags), the web extension (I am already working on it), and boards (private and public), but also other features that are more experimental but exciting in my opinion like allowing the users export and import the account in SQLite, find relations between links, fetch information from other sources about the link (it's already implemented for HackerNews)

If you are interested, you can sign up in less than 10 seconds; there is already a basic system for importing Firefox bookmarks. Let me know if you are interested in other formats.



You haven't done URL validation properly, you are checking if theres "https:" anywhere?


Hi KomoD, why do you think I am not doing URL validation properly? Can you give an example?

Yes, the validation checks if the protocol is http or https.


because I can add "https:testhello" and it turns into "https://keepthis.site/testhello"

and I can also add "eresrsefhttps:sadasdsdfjdfhgkfjd", and "https:/"

hope this clarifies


You are right KomoD, "https:testhello" passes the validation.

The app is in ruby, I am using "URI::DEFAULT_PARSER.make_regexp(%w[http https])" for generating the regex validation, which generates this monstrosity: https://gist.github.com/ceritium/08b34277fd252a2fa7db1f16bec...

I didn't mind being pretty strict regarding the URL validation; I don't even care too much about the fact of allowing "https:testhello", but I don't like that Chrome, Safari, and maybe others are opening <a href="https:testhello"> link </a> as "https://keepthis.site/https:testhello" but Firefox opens it as a new URL without "https://keepthis.site" which is the behavior I would expect.

I am going to re-implement the validation with simpler rules. Many thanks for your feedback.


Wow, yeah that is a monstrosity to say the least, anyway no problem.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: