mailing list archives

meli community discussions

⚠️ if something does not work as intended when interracting with the mailing lists,
reach out Github mirror Gitea repo @epilys:matrix.org

Potential IMAP backend rewrite

E-mail headers
From: Manos Pitsidianakis <manos@pitsidianak.is>
To: meli-devel@meli.delivery
Date: Thu, 18 May 2023 22:41:36 +0300
Message-ID: uvdm7.sx3ys5veco6g@pitsidianak.is permalink / raw / eml / mbox 
With the "wisdom" accrued over the years, I feel it's time to completely 
rewrite the IMAP backend. We now have features I didn't have back when I 
started it: better parsers, async, better testing and new IMAP servers.

This crate by @duesee is the most potential candidate for the rewrite: 
<https://github.com/duesee/imap-codec>. Usually I implement stuff myself 
but it'd never reach `imap-codec`'s level.

The motivations for the rewrite are:

- unit test EVERYTHING
- redesign backend to lazily load messages instead of entire mailboxes; 
  therefore improving performance.
- increase compatibility with more IMAP server implementations.

I'd like to hear your thoughts, if there are any.

- epilys
Reply
E-mail headers
From: matzipan@gmail.com
To: meli-devel@meli-email.org
Date: Fri, 19 May 2023 23:27:06 +0200
Message-ID: 3F85FE32-10C4-4621-800D-822645D8CDCB@gmail.com permalink / raw / eml / mbox 
Just realized I replied to Manos directly instead of via the list.

Sounds good! 

I was able to get melib to lazy load messages in my patches, but I was not able to lazy load attachments too.

I would be happy to help out with some integration tests, either with green mail or some other tool.

Andrei

> On 18. May 2023, at 23:23, Manos Pitsidianakis <manos@pitsidianak.is> wrote:
> 
> 
> With the "wisdom" accrued over the years, I feel it's time to completely rewrite the IMAP backend. We now have features I didn't have back when I started it: better parsers, async, better testing and new IMAP servers.
> 
> This crate by @duesee is the most potential candidate for the rewrite: <https://github.com/duesee/imap-codec>. Usually I implement stuff myself but it'd never reach `imap-codec`'s level.
> 
> The motivations for the rewrite are:
> 
> - unit test EVERYTHING
> - redesign backend to lazily load messages instead of entire mailboxes;  therefore improving performance.
> - increase compatibility with more IMAP server implementations.
> 
> I'd like to hear your thoughts, if there are any.
> 
> - epilys
Reply