nostr:npub12vkcxr0luzwp8e673v29eqjhrr7p9vqq8asav85swaepclllj09sylpugg, you appear to be accidentally swapping out entire URLs that contain npub addresses, rather than only when they're listed alone or with the nostr: prefix. It's probably an error in your regex, but this makes it hard for your users to find out about other apps by following hyperlinks they see in their Primal feed. Also, I often do this to share some particular information displayed on the other app, and if you swap out the links, then the readers can't see what I intended.

Here, two examples from Jumble and Lumina.

https://primal.net/e/nevent1qvzqqqqqqypzplfq3m5v3u5r0q9f255fdeyz8nyac6lagssx8zy4wugxjs8ajf7pqqsyqnj79symsak0gmlllpe0rfl6c3cu65u6fuq59zg2wty05gtpwts2lk9d8

https://primal.net/e/nevent1qvzqqqqqqypzplfq3m5v3u5r0q9f255fdeyz8nyac6lagssx8zy4wugxjs8ajf7pqydhwumn8ghj7argv4nx7un9wd6zumn0wd68yvfwvdhk6tcprdmhxue69uhhg6r9vehhyetnwshxummnw3erztnrdakj7qpqdcysz0ekxta0vve7dclen9pvtn67z78s9u87qsghx5c5s3akwx7qg3mynk

Reply to this note

Please Login to reply.

Discussion

Man! I hate regex. By brain shortcircuits any time I come within 100 meters of one.

Same, I spent a week on a markdown regex and then we found a bunch of errors in it, and I took a week off, to work on other things, because I was like

NOT TODAY, REGEX!

I've also noticed that AI is even worse at regex than humans are, which says a lot.

AI is an average rite*/*ya amIwrong? 4 mass app. th@ IZ*/*ya

For real? Wow. I thought AI would be good at it. I wonder if it's for similar reasons that it couldn't count the 'r's in "strawberry"?

Regex is more common-sense and personal experience than algorithm.

This is also weird UX, where I use a partial address as an example of the format, and you try to swap it out and return an error. If you can't find an address, just leave it as-is, don't alter it, please.

https://primal.net/e/nevent1qqsftgw79d576kf8qhv8q3s25ug4ltqxmj3xczprs4z6fe9wwxzwztgfcddr0

https://njump.me/nevent1qqsftgw79d576kf8qhv8q3s25ug4ltqxmj3xczprs4z6fe9wwxzwztgpp4mhxue69uhkummn9ekx7mqzyrwkvn27gqtyxw5v660sqkhpfqyqgdgh3x6emed0qcnkmejkx0f3j64nfzc

They do all kinds of dumb shit at primal because of their stupid cached relay idea.

All in the name of PeRfOrMAncE

so much for microapps and user choice

Apparently, they were pressured to do this. 🤷🏻‍♀️

But it's a bug, not a feature, to me.

Bug indeed.

Get pressured into to stupid ideas, get stupid results.

Garbage in, garbage out.

“Primal” - web, android, or iOS?

Web

Thanks will investigate web link behavior.

To be clear, you are saying notprimal http links are replaced in primal web?

Yes, I gave two concrete examples in this thread.

nostr:npub1l5sga6xg72phsz5422ykujprejwud075ggrr3z2hwyrfgr7eylqstegx9z I addressed the http nostr links in frame of sharing nostr native links best practices as described in the nostrability issue.

There are recommendations for all three apps you mentioned - jumble.social, lumina.rocks, and primal web.

As for discovery of nostr apps, this is a great, and bigger outstanding question.

Here is the NIP-89 tracker https://github.com/nostrability/nostrability/issues/7

In there is a ‘client tag’ column, which could be used to signal nostr app usage cross apps.

Thank you!

have you written your lexicon?

What lexicon?

You mean my Markdown parser? Yes, I wrote it.

I meant nip-89 for Alexandria

Ah, I haven't thought of that. I should do that.

it's probably not necessary, actually only for the headache