Closed jcs090218 closed 1 year ago
Hi!
Thanks for the effort, but could you explain what exactly this will make easier? js2-mode
has no package dependencies, and the test script is easily invoked through make
. And the diff adds more lines than it deletes.
So what are we getting here?
Hi, thanks for your time to review this PR!
Eask shares the same idea with Cask. I guess their manual page explained better than I do, see Why Cask?. Furthermore from this PR, Eask tries to eliminate the test command like emacs --batch -Q ...
to more human readable command eask compile
, or eask ert
, etc.
And you are correct, it indeed adds more lines than it deletes. I think the biggest addition lines are came from Eask
(+14) and setup-node
and setup-eask
from the test.yml
(+8). I have also added checkdoc
and lint
in the Makefile (sorry, I should've mentioned in the first post). With Eask, you can easily expand/eliminate your needs.
Hope these answered your questions!
To reject this PR is understandable! Since I think the Cask philosophy isn't for everyone, everything should be good! :)
I guess one thing we have from the "no Cask" list is no dependencies. And the existing Makefile
rules, combined with the CI, make it fairly easy for the contributors, I think. But the argument for consistency is not bad. As long as the ecosystem standardizes on this, I guess.
I'll take a look at the added rules and play with Eask when I have the time. In the meantime, let's keep this open, if you don't mind.
Thanks!
This patch does the following:
In
.github/workflow/emacs.yml
jcs090218/setup-emacs
instead ofpurcell/setup-emacs
16
)setup-eask
28.1
, drop Emacs25.1
(Eask must use26.1
or above)Windows
andmacOS
In
Makefile
(newly added)emacs --batch ...
with justeask <command>
make checkdoc
make lint
(but left unused for now)Others
Eask
-fileThis is totally optional, I think the solution may not be ideal since this drops the test on Emacs
25.1
. :(Let me know what you think? Hope this could make CI on Emacs a lot easier! :)