Could you help me to pass the linting checks? I fixed all the indicated errors but it still gives me the incorrect indentation error and doesn't tell me why. Please tell me what is it so that I can fix it.

Thank you @ema2159 for the hard work! I'll try to take a look at it during this weekend. 👏

I find the problem with the lint. You need to update .dir-locals.el:

((emacs-lisp-mode . ((indent-tabs-mode . nil)
                     (fill-column . 100)
                     (elisp-lint-indent-specs . ((when-let . 1))))))

That would set when-let to only required one level of indentation, which is the one being applied in the macro by default.

After that, you could test it running make lint 😉

@ema2159 One little thing. I really like the changes you propose. I want them to be included in the next 1.4.0 release. But one little thing. Would you mind to split the changes in two PRs? One PR including all all-the-icons integration (it could be this one) and the other one with the new init information.

@ema2159 One little thing. I really like the changes you propose. I want them to be included in the next 1.4.0 release. But one little thing. Would you mind to split the changes in two PRs? One PR including all all-the-icons integration (it could be this one) and the other one with the new init information.

Yes sir, I'll do that as soon as I'm out of the office, probably between today's afternoon and tomorrow.
And just to be clear with the .dir-locals.el, I just need to create a .dir-locals.el in the same directory as the dashboard.el file and write ((emacs-lisp-mode . ((indent-tabs-mode . nil) (fill-column . 100) (elisp-lint-indent-specs . ((when-let . 1))))))
in it?

No rush on the changes 😉

There is a .dir-locals.el already in the repository with configuration for emacs-lisp-mode. You just need to add the element (elisp-lint-indent-specs . ((when-let . 1))) to the alist.

Ok perfect. I also wanted to ask you what do you consider good defaults? I mean, I put all the defcustoms in nil thinking that maybe someone didn't had all-the-icons installed but I'm also thinking that people may miss that the icons option is available.
By the way, I'm pretty sure we both hablamos español a la perfección porque es nuestra lengua madre but we'll keep it in english for standard comunication jaja.

There were two errors failing in the CI. One was the indentation check that is fixed now. The other one is:

In toplevel form:
dashboard-widgets.el:396:49:Error: ‘(path (car (last (split-string (abbreviate-file-name el) " - "))))’ is a malformed function

This happens when trying to byte-compile dashboard-widgets.el. Any idea why?

There were two errors failing in the CI. One was the indentation check that is fixed now. The other one is:

In toplevel form:
dashboard-widgets.el:396:49:Error: ‘(path (car (last (split-string (abbreviate-file-name el) " - "))))’ is a malformed function

This happens when trying to byte-compile dashboard-widgets.el. Any idea why?

Yes, I think Emacs 25 doesn't support the when-let macro. Changing that for a separated let, when expression seems to fix it

Do you know exactly how do I have to include the all the icons package so that it doesn't complain about free variables or functions?

Do you know exactly how do I have to include the all the icons package so that it doesn't complain about free variables or functions?

In .emacs/dependencies.el

For some reason the master test failed to install page break lines and that made that test fail, but the others seem ok.

I was preparing a new PR to update emacs version in the CircleCI configuration. And I found the same problem. Let me work on that one later (WIP #124). Once that one gets merged, we should be good to go!

Seems that we're done! I've separated the PR as you told me and everything seems to be working fine! If there's anything else I can help you with just let me know!

#124 is merged! The problem was a missing coding setting. Feel free to rebase this branch against master or merge master in this branch (as you prefer).

Done, all the checks are passing now. What do I do now with the init info PR?

Nice to see the features of Centaur Emacs are merged! Thank you!

One question: Did you test this scenario: If user doesn't install all-the-icons, but enable dashboard-set-heading-icons and/or dashboard-set-file-icons?

I also add some comments below.

Nice to see the features of Centaur Emacs are merged! Thank you!

One question: Did you test this scenario: If user doesn't install all-the-icons, but enable dashboard-set-heading-icons and/or dashboard-set-file-icons?

I also add some comments below.

No sir, I personally haven't, but I guess it probably just display ugly icons as it usually does when the icons aren't installed, that's why I defaulted the icon's defcustom to nil, but it shouldn't be a bigger problem than that because all the icons was added as a dependency.
Also we're having this issue #128. Any idea why?
By the way, that's a killer configuration you've created.

@ema2159 I was not talking about icons, but about the all-the-icons.el package. If users didn't install this package, undefined all-the-icons-xxx will be called. I use require 'all-the-icons nil t because the package is installed in Centaur Emacs definitely. I think the solution is check if the package is installed before calling the functions, or add all-the-icons to the dependency. I prefer it's an optional package.

Regarding #128 , I have to say it was as designed, since the icons are not widgets/buttons or the part of them, for displaying correctly. The buttons have different face properties which will break the properties of icons. That's why the TAB (next widget) does work but the cursor on icons doesn't.

@ema2159 I was not talking about icons, but about the all-the-icons.el package. If users didn't install this package, undefined all-the-icons-xxx will be called. I use require 'all-the-icons nil t because the package is installed in Centaur Emacs definitely.

Regarding #128 , I have to say it was as designed, since the icons are not widgets/buttons or the part of them, for displaying correctly. The buttons have different face properties which will break the properties of icons. That's why the TAB (next widget) does work but the cursor on icons doesn't.

Yes I understand, but now all the icons is a dependency so it's installation is ensured with use package.

And regarding #128, I understand, so a work around would be that when the shorcut buttons are pressed the cursor fell directly into the text instead of the icon, do you know how to do this?

Okay. Glad you understand the risks.

Aha, and I have a great news, I think I found the solution for #128 . Should I submit a PR?

Of course, feel free to submit a PR and contribute! 😄

Already committed. Please review and merge. Thanks!