Re: GitLab integration take two, light edition

[Eric Abrahamsen <eric@ericabrahamsen.net>]

Adam Sjøgren <asjo@koldfront.dk> writes:

> Eric writes:
>

[...]

> It's just a list of lists of conses, so if only I could remember which
> is which of car and cadr, it would indeed be easy.
>
> This is what I have currently:
>
> (defun adsj-article-treat-gitlab ()
> "Show thumbsup on GitLab comments"
> (when (and gnus-treat-type (string= gnus-treat-type "text/html"))
>;;; (gnus-with-article-buffer
> (let ((project-id (gnus-fetch-field "X-GitLab-Project-Id"))
> (issue-iid (gnus-fetch-field "X-GitLab-Issue-IID"))
> (message-id (gnus-fetch-field "Message-ID")))
> (when project-id
> (string-match "<note_\\([0-9]+\\)" message-id)
> (ignore-errors
> (let ((emojis (ghub-request "GET" (format "/projects/%s/issues/%s/notes/%s/award_emoji" project-id issue-iid (match-string 1 message-id)) nil :auth 'nngitlab :host "gitlab.nzcorp.net/api/v4" :username "adsj" :forge 'gitlab))
> (name-to-emoji '(("thumbsup". "👍"))))
> (widen)
> (message-goto-eoh)
> (dolist (emoji emojis)
> (let ((name (cdr (assoc-string "name" emoji))))
> (insert "GitLab-Emoji: " (cdr (or (assoc-string name name-to-emoji) `("default" . ,name))) "\n"))))))))))
>
>   (add-hook 'gnus-part-display-hook 'adsj-article-treat-gitlab)
>
> The mapping of name-to-emoji needs to be expanded.
>
> Ideally I'd rewrite it to do a map and join over the list, so I'd get a
> single header with a comma-separated list of emojis, if there is more
> than one.

`mapconcat' is what you're looking for there, something like:

(insert "GitLab-Emoji: "
	(mapconcat
	 (lambda (emo)
	   (assoc-string emo name-to-emoji))
	 emojis))

With the appropriate guards and defaults, etc.

> The cherry on top would be if mousing over the emoji showed the name of
> the person awarding the emoji (it's in the data returned). But it's
> getting dark now.

Check out the "help-echo" text property in the manual.

> I tried using a different hook, but then gnus-fetch-field returned nil.
>
> And I cheated on handling the error, just ignoring all errors.
>
>> Is it more complicated than that for some reason?
>
> No, I just don't know (e)lisp, neither the vocabulary of functions nor
> the idioms :-)

Could have fooled me, you do pretty well with bug-hunting on this list!

Eric