prettier-js-prettify-region does not set region

[mikpom1]

(region-end) and (region-beginning) are the same number (actually beginning of the region).

Is it possible to make the command set them correctly signifying beginning and end of patched region.

[jacksonrayhamilton]

I'm not sure I understand the issue you're having. Would it help if you shared a screen recording of what's not working for you?

I'm wondering if there's some confusion about how prettier-js-prettify-region is supposed to work and be used. In the PR that added this feature, I provided some demos showing how to use it: #92. Is that helpful?

The command expects that you've defined a region already by setting mark and moving point elsewhere. (The concepts of mark/region are explained in the manual - I just reviewed them myself to try and confirm that I understood this properly.) If you haven't defined a region, then the command simply does nothing in an effort to fail gracefully. I don't believe the command is supposed to be responsible for defining the region - rather, you are. Let me know if I'm wrong on this.

[mikpom1]

emacs-prettier.mp4

I tried to demostrate what I mean in this screencast using exchange-point-and-mark command. Please notice how after applying prettier to the region, this command does not do anything. It is becase region-end and region-beginning do not mark the regions resulted in formatting.

PS Actually I need something to apply prettier to a indented region and have a hacky way to do so if the region-end and region-beginning were signifying the resulting region correctly. mark should be set by prettier after formatting regardless of this IMO.

[jacksonrayhamilton]

Oh, I see what you mean now. And I see that you originally said, "patched region," too; my bad, I missed that.

It seems that commands like indent-region, upcase-region, etc will clear the active region afterwards. In the case of indent-region, we find this comment in the implementation:

  ;; In most cases, reindenting modifies the buffer, but it may also
  ;; leave it unmodified, in which case we have to deactivate the mark
  ;; by hand.
  (setq deactivate-mark t))

Seems like the Emacs maintainers judged it might be problematic to maintain the region when a command "modifies the buffer" - which is also the case for prettier-js-prettify-region. For that reason I feel like the current behavior may be preferable, perhaps from a correctness standpoint, and at least for consistency with Emacs internals.

Can you help me understand your use case for wanting to automatically set the region afterwards to be the patched region? (Just curious; I'm sure you have a good reason for it.)

Also, if you were to submit a PR providing this feature in a lightweight and robust way, behind an option, then maybe that could solve your problem. I just think the option, if it were to be added, should be turned off by default.

[mikpom1]

Can you help me understand your use case for wanting to automatically set the region afterwards to be the patched region? (Just curious; I'm sure you have a good reason for it.)

To make a function able to apply prettier-js-prettify-region to an indented block of code.

Also to see after formatting the resulting region, e.g. by using exchange-point-and mark.