Gmail threads vs himalaya emails

[echo-bravo-yahoo]

Hey!

I'm trying to write a query for Gmail emails that are unseen and have a label, e.g., himalaya envelope list -f "MyCustomLabel" "not flag Seen". This works pretty well for most situations, but I've found one situation where Gmail's thread & label concepts seem to differ a lot from himalaya's.

Consider this situation:

• A first email, read, with the label MyCustomLabel but not the label Inbox.
• A second email, in a thread with the first email (the sender sent to you twice, replying to the first), unread, with the label MyCustomLabel and the label Inbox.

A query for himalaya envelope list -f "MyCustomLabel" "not flag Seen" returns neither, because the first is read, and the second is in the Inbox folder (despite having the MyCustomLabel label in Gmail). A query for himalaya envelope list -f "Inbox" "not flag Seen and flag MyCustomLabel" returns neither because the first is read, and the second is not labelled with MyCustomLabel by himalaya even though it is in Gmail.

Here's an example of the second, unread email that's currently in the Inbox folder but has the MyCustomLabel label:

> himalaya envelope list -f "Inbox" -s 200 -o json
  {
    "id": "87595",
    "flags": [],
    "subject": "Re: Sr SWE, Web Apps [Remote] @ Machinify, Machine Learning for healthcare - Worth a chat!",
    "from": {
      "name": "REDACTED",
      "addr": "REDACTED"
    },
    "to": {
      "name": null,
      "addr": "REDACTED"
    },
    "date": "2025-03-13 20:50-05:00",
    "has_attachment": false
  }

And here's a screenshot of the email thread from the Gmail interface (it has 3 emails in the thread, but the 3rd is functionally the same as the 2nd - unread, labelled, in inbox):

[soywod]

I see what you mean. It's part of a bigger problem, where every action is tieded up to a folder or mailbox. IMAP works this way be design. This makes the thread feature bearely usable, because messages are usually in different folders. Adding a folder filter inside the search query could fix the issue, but it challenges the whole structure. I definitely need to find a better solution to that problem. Thank you for initiating the point!