reply on: SN release: nostr crossposting, symmetric lud18, quote replies, @ autocompletes \ stacker news ~meta

5 sats \ 9 replies \ @WeAreAllSatoshi 5 Oct 2023 \ parent \ on: SN release: nostr crossposting, symmetric lud18, quote replies, @ autocompletes meta

@hampus I see this in the logs when trying to call the invoice request from my dev machine:

{
  statusCode: 500,
  error: 'Internal Server Error',
  message: 'Could not parse query params'
}

with a query param string of ?amount=10000&payerdata=%257B%2522identifier%2522%253A%2522SatsAllDay%2540stacker.news%2522%257D

An issue decoding the URI encoded param, maybe?

ETA: same thing when you send name, as well. Are we doing the payerdata query param encoding incorrectly?

55 sats \ 8 replies \ @WeAreAllSatoshi 5 Oct 2023

@hampus Taking a look at chat.blixtwallet.com's code, I don't see any explicit decoding of the query param: https://github.com/hsjoberg/lnurl-pay-chat-server/commit/a28f998675a675a835ea1a639be2ec1f1fdae0ec#diff-a2a171449d862fe29692ce031981047d7ab755ae7f84c707aef80701b3ea0c80L226

Perhaps we're double encoding when we shouldn't be?

351 sats \ 0 replies \ @hampus 5 Oct 2023

I don't see any explicit decoding of the query param:

It was some time ago I made this server, but I think Fastify is implicitly decoding the GET params. It makes it into an object.

It will naturally not attempt to decode twice, which is why the JSON.parse() errors out.

351 sats \ 2 replies \ @hampus 5 Oct 2023

Perhaps we're double encoding when we shouldn't be?

Yes, as far as I can tell, it seems to be double-encoded when it shouldn't be.

A quick look. Yours require two decodeURIComponents:

// stacker.news
decodeURIComponent(decodeURIComponent("%257B%2522identifier%2522%253A%2522SatsAllDay%2540stacker.news%2522%257D"));
'{"identifier":"SatsAllDay@stacker.news"}'

// blixt wallet
decodeURIComponent("%7B%22name%22%3A%22Hampus%22%2C%22identifier%22%3A%22hampus%40blixtwallet.com%22%7D");
'{"name":"Hampus","identifier":"hampus@blixtwallet.com"}'

351 sats \ 1 reply \ @hampus 5 Oct 2023

Though I'm not sure who's in the wrong here. I'll investigate some more.

798 sats \ 0 replies \ @WeAreAllSatoshi 5 Oct 2023

We’re probably getting URI encoding for free with the native JS API that constructs the URL

5 sats \ 3 replies \ @hampus 5 Oct 2023

Perhaps we're double encoding when we shouldn't be?

Yes, my understanding after reading the specification again is that it should not be encoded twice.