Where communities thrive

Join over 1.5M+ people
Join over 100K+ communities
Free without limits
Create your own community

Explore more communities

magit/magit

It's Magit! An Emacs mode for Git.

People

Repo info

Activity

Amanda Cameron

@AmandaCameron

Hullo. I'm starting out with emacs, and I appologise if this is out of scope, but I'm trying to figure out how to better have my windows split. For example, I'd like it so that when a commit is being crafted, the diff doesn't appear in an existing area, but in a new split below, any help would be appreciated.

Trey Harris

@treyharris

@AmandaCameron Trying to understand your question a bit better:

Currently, if you:

Have the magit-status window open as the only window in the frame, and in it you use c c or other such start-commit commands, it splits, such that the magit-status window goes away replaced by the editable commit-message buffer, and the new window contains a newly-created diff to match what you are committing.
Already have two windows open, magit-status and “Buffer B”, c c gives you the same result as No. 1 above: one window with editable commit message, other with the diff.

You’re saying that the No. 1 behavior is okay, but in the No. 2 case, you want the result to be three windows, with “Buffer B” still in the frame?

Amanda Cameron

@AmandaCameron

yes, instead of the diff taking over buffer b

instead of it taking over buffer b's window, that is

Trey Harris

@treyharris

@AmandaCameron Hmm.... there used to be a bunch of magit-.*-function variables you could defadvice to do that (in this case, probably something like magit-commit-create-function), but they were removed during a ~2018 rearchitecting of Magit because the functionality was split between Magit and magit-popup stuff. I’m not sure how you’d do this now

M-x Ag

@iLemming_twitter

@AmandaCameron check the docs for magit-display-buffer-function var. Mine is set to 'magit-display-buffer-same-window-except-diff-v1

Noam Postavsky

@npostavs

split-width-threshold might be relevant as well (that's just a general non-magit specific Emacs setting)

Dieter Komendera

@kommen

since I upgraded my emacs to 27.0.90 it seems magit-save-repository-buffers is ignored, and I always have to confirm to save each modified file. anybody else having this?

Dieter Komendera

@kommen

it looks like this eq returns nil, even though magit-save-repository-buffers is set to dontask https://github.com/magit/magit/blob/12cd9bf1a095f526237ef1cb1b6b3fa54df83abc/lisp/magit-mode.el#L1169

can this be a naming collision issue with the defcustom and the defun both named magit-save-repository-buffers?

Dieter Komendera

@kommen

my elisp foo is not strong enough to figure out what is going on here

Noam Postavsky

@npostavs

@kommen can't be a naming collision, variables and functions have separate namespaces. Have you tried recompiling magit (and perhaps related packages)? There are sometimes incompatibilities across Emacs versions that can get baked into the .elc files and cause strange behaviour.

Dieter Komendera

@kommen

@npostavs yes, I tried recompling and re-installing as well

my last findings where, that using (debug) before the linked eq above, evaluating (type-of magit-save-repository-buffers) says it is a cons, where as doing the same in ielm results in symbol

Noam Postavsky

@npostavs

@kommen What does evaluating magit-save-repository-buffers from the debugger give you?

Dieter Komendera

@kommen

@npostavs 'dontask

but (eq magit-save-repository-buffers 'dontask) is nil

Noam Postavsky

@npostavs

@kommen the ' is included in the evaluation result?

Dieter Komendera

@kommen

yes

(car magit-save-repository-buffers) returns quote

Noam Postavsky

@npostavs

Is that the case at the top-level in ielm as well, or just in the debugger frame?

Dieter Komendera

@kommen

just in the debugger frame

Noam Postavsky

@npostavs

Sounds like something is modifying the value in a strange way. I suggest doing M-x debug-on-variable-change magit-save-repository-buffers RET to catch where that happens.

Dieter Komendera

@kommen

Debugger entered--killing local value of magit-save-repository-buffers in buffer magit: ars: 
  debug--implement-debug-watch(magit-save-repository-buffers nil makunbound #<buffer magit: ars>)
  kill-all-local-variables()
  special-mode()
  magit-section-mode()
  magit-mode()
  magit-status-mode()

Dieter Komendera

@kommen

@npostavs thanks for your help. I don't think it is worth for me right now to continue down that rabbit hole

Noam Postavsky

@npostavs

@kommen that change is probably okay (although I'm not sure why it's local in the first place), so type c to continue from there.

Dieter Komendera

@kommen

but this is the last change to that variable:

Debugger entered--setting magit-save-repository-buffers in buffer magit: ars to 'dontask: 
  debug--implement-debug-watch(magit-save-repository-buffers 'dontask set #<buffer magit: ars>)
  hack-one-local-variable(magit-save-repository-buffers 'dontask)
  hack-local-variables-apply()
  hack-dir-local-variables-non-file-buffer()
  magit-status-mode()
  magit-setup-buffer-internal(magit-status-mode nil ((magit-buffer-diff-args ("--no-ext-diff")) (magit-buffer-diff-files nil) (magit-buffer-log-args ("-n256" "--decorate")) (magit-buffer-log-files nil)))

Noam Postavsky

@npostavs

@kommen ah, looks like you have a wrong setting in a .dir-locals.el file

Dieter Komendera

@kommen

@npostavs damn

so there is (magit-save-repository-buffers . 'dontask) in there, and I guess the quote is wrong

Noam Postavsky

@npostavs

yep

Dieter Komendera

@kommen

thanks! interesting that this now only pops up with emacs 27, it is in there since a couple of months that way...

asymmetric

@asymmetric

can i merge a PR using forge by merging the branch into its upstream and pushing? Will that mark the PR as merged?

found magit/forge#96, answers my question

asymmetric

@asymmetric

is it possible to customize which suffixes are shown?

See magit/magit#4059

e.g. by specifying a threshold

asymmetric

@asymmetric

found out how to do it: activating the option then asks for a level

beautiful implementation! :heart:

Rudi Grinberg

@rgrinberg

Is there a way to create a forge PR and immediately set the reviewers?

s/forge PR/PR created with forge/

Magnus Therning

@magthe

@rgrinberg the only way I know is to use .github/CODEOWNERS

donahchoo

@donahchoo

The org that my GHE repo is in has changed. Is there a way to get magit/forge to use the updated org name without having to reclone the repo?

donahchoo

@donahchoo

I think I did this by renaming origin and then using b u to unset the remote and then set it.

Rudi Grinberg

@rgrinberg

I see. Thanks anyway.

Is there a way to change log settings for the current log buffer?

For example, I'd like to add the diff stat to the current log with -s

Kévin Le Gouguec

@peniblec

Hi! Sorry for the repost, but I still haven't figured this out:

Quick question (hopefully): when global-magit-file-mode is t (the default), does one still need to (require 'magit) in their init file in order to enable magit-file-mode bindings?
My .emacs does not (require 'magit) ATM and C-x g and other shortcuts do not seem to work until I manually call M-x magit-status once. Adding (require 'magit) in the init file fixes this. Is this expected?
The installation node in the manual explicitly says to (require 'magit) when installing from the Git repository, but no such step is mentioned in "Installing from Melpa" nor in "Post-Installation Tasks".
(Tested with Emacs 28)

(My curiosity was renewed by Stefan's recent message on emacs-devel explaining why one shouldn't, in principle, (require 'foo) in their init file)

@rgrinberg

For example, I'd like to add the diff stat to the current log with -s

Butting in without having read the rest; does L -s g in a log buffer do what you want?