Question about contacts format

[jonlachmann]

Hi,

First off, nice project! I am building an email server that will be open source, and I just started adding JMAP support, and having this project to test against has been very useful!

This might be a bit premature, as it is not even merged yet, but I have a question about the contacts format. I thought I would ask to be sure that I am reading the spec correctly.

In the fork by @luna-dj there is support for contacts, but when I try it against my implementation, it complains about the formatting of the request.

Looking here: https://github.com/luna-dj/jmap-webmail/blob/6b1f2582728cd6b97636dfb8584b6ef7da5c491c/test-create-contact.js#L132, we see that the emails of a contact is specified as

emails: {
        'email-0': {
          address: 'test@example.com',
          contexts: ['personal']
        }
      }

but if I look at RFC 9610 in the example in 4.1 I see

"emails": {
           "0": {
             "contexts": {
               "private": true
             },
             "address": "joe.bloggs@example.com"
           }
         }

note the difference in contexts where there is a boolean attached to each value in the RFC. This is also true for RFC 9553.

Is it correct that there is a mistake in the code? Or am I reading the RFC's wrong?

[luna-dj]

This because stalwart has a weird implementation of contacts that i had the pleasure of fighting with over the last week mostly because i can't really read rust and stalwart is written in rust so i asked a AI for help and it make that test script and it works because when the test script creates a contact it also exists on carddav

also take a look at https://github.com/luna-dj/stormbox

[jonlachmann]

Cool, will look into stormbox!

I see, so perhaps the stalwart implementation is not in accordance with the RFC, did you raise this issue with them?

[noelhibbard]

Beware that the repo @luna-dj just linked to has the JMAP server hardcoded to his own stalwart instance. This is a great way to harvest people's credentials.

[jonlachmann]

Beware that the repo @luna-dj just linked to has the JMAP server hardcoded to https://hivepost.nl. This is a great way to harvest people's credentials.

I saw that, and probably sent my test credentials there once before realizing, not that they are useful for anything, but still. After solving this I managed to run it and found it really useful to try out my server implementation.

Another question that came up was that I saw that jmap-webmail appears to have the parameter end instead of duration for calendar events here: https://github.com/luna-dj/jmap-webmail/blob/6b1f2582728cd6b97636dfb8584b6ef7da5c491c/lib/jmap/types.ts#L279. If I read the spec right here https://datatracker.ietf.org/doc/html/rfc8984#name-duration-2 it appears that end should not exist, but duration should.

Please correct me if I am wrong, I still feel very confused trying to decipher the text of various RFCs. Also do let me know if I can help in perhaps creating issues for stalwart if they have implementations which are not following the standard.

[luna-dj]

Beware that the repo @luna-dj just linked to has the JMAP server hardcoded to https://hivepost.nl. This is a great way to harvest people's credentials.

yeah sorry those were not supposed to be there i forgot to change them there were some files that i need to remove from git history and change the passwords force pushed the t my fork to remove the creds