I can't get certain characters, like ’ and ‘, to display properly in vim. They show up like <?>#~X. They work fine on SnowLeopard, but not on Ubuntu 10.10. The vim config is mirrored on the two systems, and contains:
set fileencoding=utf-8
set encoding=utf-8
set termencoding=utf-8
The ’ ‘ signs work in my terminal on ubuntu. Vim on ubuntu is version 7.3 p112, Vim on osx is 7.2 p108.
Any suggestions?
Am I missing anything in my config?
tx,
-aktiv
I have tried an unchanged vim on Ubuntu 10.04 and both of ticks are displayed as expected. The version I have tried was 7.2.330. If it helps, I can try 10.10 later, too.
Some remarks to track the problem down:
Debian/Ubuntu is completely UTF-8 based/aware for some time now. So there is no need to set it on your own, but it should not hurt, too.
For example the back tick (`) is an ASCII character and should "always" be available. It is an important character, e.g. for Shell or Perl programming.
So I do not expect that you are missing something in your configuration. Quite the opposite:
Perhaps you have added something in your configuration that treats the back tick special, e.g. tries to change the color or something like that.
So I would like to suggest, that you remove your custom configuration from the system and tries if the ticks and other characters are working. If this is the case you can add some of the configurations again. At some point the ticks should stop working. Then you have a small number of customizations that make these ugly characters.
Related
I recently switched from bash to zsh and now the tabstops in my Ultisnips-snippets aren't working properly anymore.
I didn't change anything in my snippet files and after I encountered the problem for the first time, I went back to bash, started vim from there and the snippets worked as intended again.
I have these to snippets
snippet ( iA
($1)$0
endsnippet
snippet mk wA
$$1$$0
endsnippet
If I run vim from bash and type mk, it will expand to
$Cursor$
(I will denote the cursor position by Cursor). Then I can type f( and it will expand to
$f(Cursor)$
If I press x and then tab it will take me to
$f(x)Cursor$
(I will call this previous step S). Pressing tab again I will arrive at
$f(x)$Cursor
Running vim from zsh however, I can still do the steps until S but if I then press tab, it will stay at
$f(x)Cursor$
and start inserting tabs after that. If I only do mk however and type something else without brackets, for example ab
$abCursor$
and then press tab, it will work as expected and get me to
$ab$Cursor
Does anybody have an idea what's going on here? What can I do if I want to keep using zsh and Ultisnips? If I run bash -c 'vim file.txt', vim will work really slowly and lag a lot, so that is not really an option.
Any help would be appreciated.
Edit: Apparently the problem has something to do with what version of vim I'm using. I use macOS and the problem I described happens when I use the default vim version. If I use a custom-compiled version of vim, it works as expected. And as it turns out I have some config file that tells bash to use the custom-compiled version, whereas zsh still uses the default one. I also tried it on my PC (running Ubuntu) and with the vim version installed there, it also worked. So I guess, I'll use my custom-compiled vim to avoid this problem...
But does anybody know, what vim setting could be responsible for this? If I'm not mistaken my custom-compiled version is just the huge-feature version of vim.
I'm using Mac OS 10.13.4.
iTerm2 Version is Build 2.1.1
Vim Version is
VIM - Vi IMproved 8.0 (2016 Sep 12, compiled Nov 29 2017 18:37:46)
Included patches: 1-503, 505-680, 682-1283
Compiled by root#apple.com
And I Installed the Vundle plugin
And when I open vim in iTerm2 with oh-my-zsh, the $p appears at the first line.
It doesn't appear with Terminal.app of Apple?
How to solve this problem?
vim is assuming that iTerm2 is xterm (not a good assumption), and attempting to determine the state of the cursorBlink resource by sending an escape sequence containing $p which iTerm2 does not handle properly (see source-code for vim for the escape sequence, and also where it uses the feature). While vim starts with the TERM setting (i.e., "xterm"), it does make a few checks to exclude things like gnome-terminal as noted in the source code comment. But iTerm2 happens to fool vim in this case. So the result goes to the screen.
In XTerm Control Sequences that is documented like this:
CSI ? Ps$ p
Request DEC private mode (DECRQM). For VT300 and up, reply is
CSI ? Ps; Pm$ y
where Ps is the mode number as in DECSET/DECSET, Pm is the
mode value as in the ANSI DECRQM.
Two private modes are read-only (i.e., 1 3 and 1 4 ), pro-
vided only for reporting their values using this control
sequence. They correspond to the resources cursorBlink and
cursorBlinkXOR.
Although vim is fooled here, the problem is due to iTerm2, for which you've probably set the TERM environment variable to xterm-256color, or something like that. It doesn't match iTerm2's behavior very well (the function keys don't match up, etc). ncurses provides a better one. But out-of-the-box MacOS has a terminal database that's about ten years old, and lacks that entry. To get a good terminal database (i.e., one where you could set TERM to iterm2), you could do that with MacPorts or home-brew.
Running infocmp to get a measure of the differences between the (correct) iterm2 entry and linux, xterm-256color shows that
it's actually closer to nsterm-256color (a correct entry for Terminal.app which Apple doesn't provide) with 38 lines of difference,
next closest linux with 76 lines and
still further away from xterm-256color with 94 lines.
The feature that's missing or mis-implemented in iTerm2 isn't in the terminal description: it's a special feature that vim has to guess about.
As mentioned above, this $p is due to vim sending a control sequence for a cursor blink request. In your .vimrc, you can turn off the cursor blink request by clearing the t_RC terminal option. I wrapped this in a conditional which turns it off only for 256 color terminals.
if &term =~ '256color'
set t_RC=
endif
Setting the terminal explicitly to linux solved it for me:
export TERM=linux
The p means print, which is the command the range is being given to, and yes.. it displays to you what is in that range.
In vim you can look at :help :range and :help :print to find out more about how this works. These types of ranges are also used by sed and other editors.
They probably used the 1,$ terminology in the tutorial to be explicit, but note that you can also use % as its equivalent. Thus, %p will also print all the lines in the file.
Context
GNU bash 4.4.12(1)-release
Powerline 2.5.2-1
Powerline installed from arch linux packages, not from pip.
PS1 Script
update-ps() {
export PS1="$(powerline shell aboveleft)"
export PS2="$(powerline shell left)"
}
export PROMPT_COMMAND="update-ps;$PROMPT_COMMAND"
Powerline config
config.json
colors.json
colorschemes/shell/default.json
themes/shell/default.json
Problem
I have the same problem as him, when I write a few chars the line wraps and I start writing in the same line, overwriting what I already wrote (including ps1).
I am aware that this could be a non-printable character problem, but doesn't it means that it's a powerline bug? What other problems can it be?
For me changing in ~.bashrc:
from:
. /usr/local/lib/python2.7/dist-packages/powerline/bindings/shell/powerline.sh
to
. /usr/local/lib/python2.7/dist-packages/powerline/bindings/bash/powerline.sh
resolved the issue
This can occur when the locale is incorrect, or when glibc is built with "C" locale, which does not support unicode characters. Powerline uses non-breaking spaces (2-byte) and also uses 2-byte unicode characters for the triangles. For each one of these characters, the shell (not powerline) thinks that 1 extra character has been printed. If you have 5 spaces and 3 triangles in the prompt, then the prompt will wrap 8 characters before it hits the end of the line. From there everything behaves incorrectly. This only relates to Powerline in the sense that Powerline uses unicode characters in the prompt.
The fix is to correct the locale so the shell understands 2-byte unicode. You can run "locale" to see if the shell is using "C", in which case it needs to be fixed. For my application, we were running Powerline in a Docker image running Centos Linux. We had to fix the locale in /etc/yum.conf to use the lowercase form of utf8 like so:
sed -i 's/UTF-8/utf8/' /etc/yum.conf
Then rebuild glibc-common with the fix:
yum reinstall glibc-common
After that, new terminals behaved and the prompt wrapped correctly. We then modified our Docker image to fix yum.conf before installing glibc. However, we found one of our modules was changing the locale back to C (LANG="C"), so we used the LC_ALL env var to override it:
LC_ALL=en_US.UTF_8
After that, all new terminals behaved correctly.
I'm running zsh with oh-my-zsh on OS X. Every time I use zsh's awesome tab-completion, formatting on the current command line prompt gets really screwed up. For example:
I'll be typing cd fo and try to tab-complete for the 'foo' directory; zsh prompts for completion but changes the command line to cd fo cd fo while it's waiting for me to complete. It's not a big deal but very annoying. Any suggestions?
I had the same issue on PopOS and Arch linux. I tried a bunch of solutions from various places but the only solution that worked for me was this suggestion by romkatv on an issue on the oh-my-zsh github repository.
The solution is to make a copy of the .zsh-theme file of whatever theme you're using in oh-my-zsh and surround all non-ASCII characters (like emojis) with %{%G<CHARACTER>%}
For example, the default oh-my-zsh theme robbyrussel contains 2 non-ASCII characters. The '➜' character in the prompt
PROMPT="%(?:%{$fg_bold[green]%}➜ :%{$fg_bold[red]%}➜ )"
and the '✗' character in the prompt for git directories
ZSH_THEME_GIT_PROMPT_DIRTY="%{$fg[blue]%}) %{$fg[yellow]%}✗"
Using %{%G<character>%} around the 2 non-ASCII characters like this
PROMPT="%(?:%{$fg_bold[green]%}%{%G➜%} :%{$fg_bold[red]%}%{%G➜%} )"
and this
ZSH_THEME_GIT_PROMPT_DIRTY="%{$fg[blue]%}) %{$fg[yellow]%}%{%G✗%}"
solved the issue for me.
I have faced the same problem before, my solution was disabling some zsh plugins. The second probability is that your colour theme may contain a bug which causing this.
# Custom plugins may be added to ~/.oh-my-zsh/custom/plugins/
# Example format: plugins=(rails git textmate ruby lighthouse)
plugins=(git)
This is the final version of my plugin section in the ~/.zshrc file. Any other plugin between parenthesis may be the reason of your situation.
If your problem still continues you need to post your ~/.zshrc to let us check what is in there.
I had the same issue. Interestingly, I saw the problem only iterm2 while the prompt is correctly displayed in the standard terminal of OS X (after reverse-i-search/tab-completion). The reason seems to be that iterm2 defaults to Unicode (UTF-8) default encoding, which however is not correctly interpreted if the corresponding language variable is not set in the shell.
Solution: add the following to your .zshrc
export LC_ALL=en_US.UTF-8
The prompts will be displayed correctly.
I'm on OS X Mountain Lion, running the included ZSH shell (4.3.11) with Oh-My-ZSH installed over the top.
When using tab completion with commands such as homebrew, when ZSH lists the available commands, it is also duplicating the command. For example:
$ brew {tab}
will result in:
$ brew brew
[list of homebrew commands]
I'm unsure what is causing this error, as when I resize the terminal window, the first instance of the command name disappears.
If I hit backspace when the duplicates are displayed, I can only delete the second instance of the command, zsh won't let me backspace any further. Also, if I do remove the duplicate with backspace, zsh then acts as if there is no command typed at all.
My .zshrc along with all my other .configuration files can be found at https://github.com/daviesjamie/dotfiles
UPDATE: I found this post about someone having the same problem on Ubuntu. However, I don't understand the given solution, and I'm not even sure if it applies to my set up?
This effect also could be reproduced if you use any of fancy UTF-8 characters like arrow, "git branch" character and so on.
Just remove this chars from prompt and duplication will not occur.
Also adding
export LC_ALL=en_US.UTF-8
export LANG=en_US.UTF-8
to ~/.profile can help
The problem is likely to arise from misplaced %{ %} brackets that tell zsh that text inside has zero width. The only things that should be enclosed in them are escape sequences that change color or boldness of the text. If you are using new zsh (>=4.3.{unknown version}) I would even suggest to use %F{color}...%f, %K{color}...%k, %B...%b instead of %{${fg[green]}%} or what you have there.
The problem with them is that there is no way to query the terminal with a question like “Hey, I outputted some text. Where is the cursor now?” and zsh has to compute the length of its prompt by itself. When you type some text and ask zsh to complete zsh will say terminal to move cursor to specific location and type completed cmdline there. With misplaced %{%} brackets this specific location is wrong.
If you use iTerm on Mac, be sure to check "Set locale variables automatically" in your profile preferences. I had it unchecked for an SSH connection and it resulted in the same bug and I fixed it by leaving that option checked.
It's an old thread but I faced similar issue in my zsh setup with oh-my-zsh configuration.
Setting export LC_ALL=en_US.UTF-8 fixed the issue.
A lot of answers in a lot of places suggest the export LC_ALL=en_US.UTF-8 solution. This, however did not work for me. I continued to have this issue using oh-my-zsh on both Arch linux and PopOS.
The only solution that worked for me was this suggestion by romkatv on an issue on the oh-my-zsh github repository.
It turns out, at least in my case, that the autocomplete duplication issue would only show up if there was a non-ASCII character somewhere on the line (like an emoji). And ZSH would incorrectly assume that this non-ASCII character needs to take up 2 character spaces instead of 1.
So the solution that worked was to open up the .zsh-theme file of whatever theme you're using, find all non-ASCII characters and use %{%G%} to tell ZSH to only use one character width for that character
For example, the default oh-my-zsh theme robbyrussel contains 2 non-ASCII characters. The '➜' character in the prompt
PROMPT="%(?:%{$fg_bold[green]%}➜ :%{$fg_bold[red]%}➜ )"
and the '✗' character in the prompt for git directories
ZSH_THEME_GIT_PROMPT_DIRTY="%{$fg[blue]%}) %{$fg[yellow]%}✗"
Using %{%G<character>%} around the 2 non-ASCII characters like this
PROMPT="%(?:%{$fg_bold[green]%}%{%G➜%} :%{$fg_bold[red]%}%{%G➜%} )"
and this
ZSH_THEME_GIT_PROMPT_DIRTY="%{$fg[blue]%}) %{$fg[yellow]%}%{%G✗%}"
is what finally fixed the issue for me.
So all you need to do is make a copy of the theme file you want to use and edit all the non-ASCII characters as shown above and you should hopefully never see the duplication issue again.
My solution to make both local and ssh work is something like a combination of #Marc's and #neotohin's answers:
Set export LANG=en_US.UTF-8 (simply uncomment that part in the template .zshrc; exporting LC_ALL, as in #neotohin's answer, instead of LANG may also work, I didn't try)
Uncheck "Set locale environment variables on startup" in the Terminal profile's "Advanced" section (reason: that setting sets LC_CTYPE=UTF-8 instead of en_US.UTF-8, which brakes the locale for me in ssh)