Typing text in Emacs with company-mode turned on. When I stop at the end of a word company-mode shows available completions. Because the word is already written, it does not need to be auto-completed and I want to type another word, I press space key. company-mode inserts first suggestion instead of canceling auto-completion. Is this normal? How can I make company-mode to cancel auto-completion when I insert space character?
Example:
This file is
Cursor right after "is" and company-mode shows some auto-completion suggestions. After I press space key I get:
This file isearch-adjusted
("isearch-adjusted" was the first suggestion in the tooltip of company)
See the description of company-auto-complete, and set it to nil to disable this behavior.
It's nil by default, though, so it shouldn't have caught you unawares. Are you using someone else's Emacs config?
Related
In EMACS the default seems to be to 'copy' any text highlighted with the mouse. I'm attempting to replicate the functionality of modern text editors, where I can highlight a section of text and press 'paste' to replace it. I have so far added
(delete-selection-mode 1)
to my init.el
The problem is, if I copy something, then highlight to paste in its place, I end up pasting what I had just highlighted, changing nothing.
What do I need to change to fix that behaviour?
The most powerful element of emacs is its introspection features, lets have a look at how we can use them to try and solve this problem. We must use the power of the source.
One of the most important tools for introspection in emacs is the describe-key command which is bound to C-h k. It brings up the documentation of whatever keystroke is called after it. So in our case if we press C-h k and then click and drag we will see the documentation for <down-mouse-1> and more importantly for <drag-mouse-1>. The documentation states that "<drag-mouse-1> at that spot runs the command mouse-set-region". Below it then gives some documentation for this command. It says
Set the region to the text dragged over, and copy to kill ring.
This should be bound to a mouse drag event.
See the ‘mouse-drag-copy-region’ variable to control whether this
command alters the kill ring or not.
Now we know that somehow mouse-drag-copy-region controls whether or not the highlighted text is copied.
If we follow the link of that variable it tells us the default value and some documentation:
If non-nil, copy to kill-ring upon mouse adjustments of the region.
Now all we have to do is set the variable to be nil to get the effect that you want. Place the following code at the end of your init file and you should be all set
(setq mouse-drag-copy-region nil)
I hope that this helps you with this problem and that more importantly it helps you with further issues.
By default, selecting a region with the mouse does not copy the text to the kill ring. If your Emacs does this, you probably have set the variable mouse-drag-copy-region.
In a fresh Emacs (24.5 started using -Q), you can do the following:
Start delete-selection-mode.
Mark a region using the mouse. Copy it using M-w.
Mark a second region. Replace it with the first using C-y.
I see two alternatives, neither of which does exactly what you request. (For both, yes, turn on delete-selection-mode.)
Use the secondary selection for the text to copy, and use the primary selection (the region) for the text to be replaced.
You copy text into the secondary selection using the Meta key plus the mouse - for example, press and hold Meta (the Alt key, usually) while dragging or double-clicking mouse-1.
You paste the secondary selection using Meta plus mouse-2.
Select text with the mouse, then copy it to the kill-ring using M-w. Then select the text to replace with the mouse and use C-y to paste the copied text to replace it.
I am new emacs user and one of the things that irritates me is that when I want to replace current selected text with the one from clipboard I need to delete it first. Every other application that I know replaces pasted text with the current selection by default.
Here's a little bit more detailed description:
Select some block of text
Paste text from clipboard
Emacs just pastes text where the cursor was and previously selected text it is still there. I want that selected text was deleted first.
As artscan wrote in a comment, you can get this functionaly for the normal yank (paste) operations by adding:
(delete-selection-mode 1)
to your configuration.
If you want yank by mouse to also delete the current selection, you can add:
(put 'mouse-yank-primary 'delete-selection 'yank)
in your configuration as well.
I need to use Eclipse to edit a language that doesn't have a good major mode in Emacs. I'm using what it advertises as "Emacs keybinding scheme", but it has several distinctions, some of them just happen too often and are really annoying when they happen.
One such keybinding is the Ctrl+J, or in a more Emacs-friendly format: C-j. This combination is typically bound to insertion of new line and indentation before the caret (unlike Enter or RET, which just inserts a line ending character).
So, I went to the Preferences panel, the keys settings, but I can't find anything that would simply insert a newline character. I.e. I can unbind the C-j combination, but I don't seem to be able to assign it to do the insertion. Is there any way to do it? I only found "Insert Line (Above / Below) Current Line" command, but no analogue to self-insert-command etc.
This was one of the more important keybindings for me from emacs, so I really needed to get this to work.
Go to preferences->keys, choose emacs as the scheme, search for ctrl+j, and unbind it. It's set to incremental search by default, so I changed it to ctrl+s. Once it's unbound, it just works as it should. This has worked for me from eclipse 3.x to Juno.
It's not that convenient when you do a typo during an incremental search and the search string receives the wrongly typed character. Is there a way to prevent this. As if control-g was pressed automatically on error.
For example we have the following text:
keywords
keys
Default emacs behavior:
We start incremental search and search for "keyz"
The "keyz" is displayed in the search echo area and the "key" part in "keywords" is higlighted
We press s
"keys" won't be found, the cursor stays on the "keywords" line, search echo area displays "keyzs", which is not convenient
Needed behavior:
We start incremental search and search for "keyz"
The "key" is displayed in the search echo area and the "key" part in "keywords" is higlighted
We press s
"keys" is found and highlited
You could try something like
(defadvice isearch-printing-char (before drop-mismatches-on-next-char activate)
(while (or (not isearch-success) isearch-error)
(isearch-pop-state)))
Emacs keeps the incorrect part, because it happens very often that you search for a string and it is not found, but not because it's incorrect, only the search string is found before the cursor. In this case it is very convenient that you can press C-s and the search starts from the beginning of the file.
It is very useful behavior and it happens to me more often than mistyping the search string. If there is indeed an error in the search string then you can simply press C-g to go back to last good search string.
I think the problem is that you're not thinking about searching in a way that's congruent with the way isearch has been designed to work, and so your question doesn't really make sense within the context of isearch as it currently exists.
Isearch does already give you exactly the feature what you want, but you have to tell it that you want it to happen by typing that C-g you seem so vehemently opposed to typing. If you don't tell isearch what you want to do, and when you want it to do it, how is it supposed to know what to do?
As #Tom tried to explain, the default way isearch starts from the current position in the buffer, and can restart at the beginning of the buffer if you've typed some failed characters and then press C-s, is a very valuable feature. I'm sure many people rely on this behaviour. Your method of using a macro to always start an isearch at the beginning of the buffer would confound and confuse many of us, though of course it's not a bad thing for someone such as yourself who is accustomed to it. It does mean though that the rest of us are quite confused by your dislike for having to press C-g to delete the non-matching text.
Think also for a moment about what a second C-s does if you press it immediately after starting isearch (any time but the first time in a session) (i.e. before you type any other character). Note in particular what happens if your previous search string would only partly match something in the current buffer, and then you press C-g (and also note how the failed search string is presented, regardless of whether it would partly match something in the current buffer or not).
Think also about how your feature might adversely affect the use of multi-isearch-next-buffer-function.
Claiming that other editors can do what you want isearch to do in emacs doesn't really help your case much.
I think what you really want is some slightly different type of search function which only allows you to search for text that it is possible to find in the current buffer, instead of isearch's ability to search for anything whether that text happens to exist in the current buffer or not.
Perhaps isearch-mode could be adapted to do what you want it to do, but one way or another I think you'll have to write some elisp code. Perhaps you could implement your new search mode as an option within isearch-mode that can be toggled on and off in the same way case sensitivity can be toggled on and off; and that can be set by default, again in the same way that case sensitivity can be turned on or off by default.
If you make a typo during incremental search all you need to do it press backspace to correct the typo.
We start incremental search and search for "keyz"
The "keyz" is displayed in the search echo area and the "key" part in "keywords" is higlighted
We press backspace
We press s
"keys" is found and highlited
As with most of the other answers, I'm just pointing out another feature which helps mitigate the described problem. This one is particularly useful if you've continued to type several would-have-been-good characters after the bad one.
M-e is the binding for editing the isearch string, and in the case where there are no matches for the current string, it rather helpfully places point at the first non-matching character.
So if you have made a small typo, you can quickly type M-e, fix the mistake, and type RET to return to isearch using the corrected search string.
If I have a multiline sentence in Emacs it naturally overflows onto the the following lines. Now, if my cursor is at the beginning of such a sentence and I press the DOWN ARROW key, the cursor is placed at the beginning of the next sentence (which might be at 4-5 lines down), rather than on the next line itself (which other editors do). Same is the behavior of the END and HOME keys.
Is there a way in which I can change this behavior and get the cursor on the next line instead of the next sentence?
I haven't yet tried it myself, but I think what you are asking for is the default behavior for emacs 23. What version are you running?
You might want to check out the page Move By Visible Lines page on the emacswiki.
You might want to try auto-fill-mode or longlines-mode. To get either use M-X then type the command you want. Toggle them off the same way.
If that doesn't work you may want to examine the binding that is being applied to your down arrow. Type C-h k then hit the down arrow key.
It sounds as though the text is wrapping, so by definition (a line being a group of characters separated by a carriage return), it is moving down to the next line.
I agree it is a pain, however a lot of other editors also have this behaviour.
One way is to turn off wrapping:
M-x toggle-truncate-lines
You wont be able to see all of the text in the editor, however it will move down to the next line.