[WIP] Mimetree and mime part support. Resolve #862.

[ryneeverett]

This is a rebase and continuation of #894.

• rebase
• move pipeto functionality into separate branch/PR
• default keybindings for togglemimetree and togglemimepart
• show current mime part

[pazz]

(moved into review)

[pazz]

Thanks a lot for pushing this further and saving @lucc's old branch from oblivion!

I have a few suggestions and comments we should address before merging this:

1. There should be some default keybindings for these new commands.
2. perhaps we can add a section or paragraph on mime trees in the usage docs?
3. It'd be great if the there was a way to show which mime part is currently displayed.
   For example, check out how envelope buffers do this: by populating the info dict of the buffer, which is then being picked up by the statusline.
   The trouble with this is that the displayed mime part is not a property of the buffer but the message.
4. perhaps it is time to finally introduce a single-message buffer? In any case, we should review any changes proposed here under the assumption that this will come in future, and we should try to write code in a way that miminizes later overhead.
5. I've seen that there is a separate branch dealing with the pipeto command but there are also related commits here. Can we compartementalize and move all those into the later PR?
6. Unit tests and Codeclimate should not fail before we merge

[ryneeverett]

4. perhaps it is time to finally introduce a single-message buffer? In any case, we should review any changes proposed here under the assumption that this will come in future, and we should try to write code in a way that miminizes later overhead.

I implemented a single message buffer in one of my earliest drafts. As I recall I didn't even have to create a custom buffer I just passed the mimepart into the base Buffer class and it worked perfectly. However I didn't find this to be a useful or pleasant UI so rather than add a new Buffer subclass to add a pretty header I switched to the mime part toggling strategy. Long story short this is fairly trivial now that the mimepart is easily accessible.

6. Unit tests and Codeclimate should not fail before we merge

It seems like Codeclimate fails if there are any "issues" and any change to a function that codeclimate doesn't like is considered an issue. One of the failures is for adding the new commands to a list of completion variables in a method codeclimate considers too complex. Another is actually for decreasing the complexity by replacing extract_body (Cognitive Complexity 11) with extract_body_part (Cognitive Complexity 8). I doubt you really want those unrelated methods simplified in this branch.

[lucc]

Thank you @ryneeverett for working on this. I had previously given up because I had hit a wall.

I like the idea of a single message buffer where I can see a mime tree and expand and collaps the individual mime parts just like I can currently see a message tree in a thread buffer and expand and collapse the individual messages.

[ryneeverett]

I like the idea of a single message buffer where I can see a mime tree and expand and collaps the individual mime parts just like I can currently see a message tree in a thread buffer and expand and collapse the individual messages.

I personally don't see any use for this functionality but it would be pretty easy to implement on top of this branch. I've added a reference to the email body part to the mime tree leaves and I've broken out the extract_body_part function to return the displaystring from an individual body part, so the mimetree itself now has all the data needed to construct such a CollapsibleTree.

[ryneeverett]

2. perhaps we can add a section or paragraph on mime trees in the usage docs?

I'm open to suggestions but I'm not sure where I'd add an additional section beyond the automatic documentation in the "Commands in ‘thread’ mode" section page. Adding a top-level usage page on these features strikes me as excessive.

[ryneeverett]

Closing in favor of #1480 and #1481.