`i_CTRL-O` may change the cursor position in an unexpected way.
For example, when `autoindent` is set, the user will expect that
`i asdf<CR><C-O>` will place the cursor right below `a`.
However, `<C-O>` moves the cursor to the first column of line 2.
Expected:
```
asdf
█
```
Actual:
```
asdf
█
```
Therefore, it's desirable to use `i_CTRL-\_CTRL-O`, the variant of
`i_CTRL-O` that does not move the cursor.
Otherwise if the user has set a `git config clone.defaultRemoteName
foo`, then vim-plug will fail to detect the latest upstream changes as
the remote will be incorrect, and will repeatedly state that the plugin
repo needs to be cleaned.
Previously, the highlight of the `------------` line below the `Last Update:`
changed from `plugH2` on the whole line to `plugDash` on the first `-`
when the cursor moved over it and went below it. This commit updates the
`syn match` commands a bit to correct that issue.
Close#1027
* Make git operation faster
When using many plugins, vim-plug may spawn many git processes for them.
* get revision
* get branch
* get remote.origin.url
This is too heavy. especially on Windows. This change get revision, branch,
remote origin url directly from .git directory.
This idea is borrowed from @k-takata's commit for minpac.
Executing external programs is slow especially on Windows.
Read the information directly from .git directory.
* Copied from devel branch of minpac
* Avoid errors
* Show errors
* Use empty()
* Use empty string instead of v:null
* Check spec.branch is empty
* Use branch
* Fix branch and revision
* Remove l: and use s:trim
* Fix and simplify s:git_get_remote_origin_url
* Do not cut off commit hash for correctness
Co-authored-by: Junegunn Choi <junegunn.c@gmail.com>
Use build stages to group related jobs (Vim, Neovim, Vim 7.4).
Use "silent" to avoid hit-enter prompt when redirecting output.
Always run async and sync tests to debug runtime errors.
Vim 7.4.0052 (Ubuntu Trusty) does not allow dynamic keys in inline dictionary.
https://docs.travis-ci.com/user/build-stages/
":Plug" performs a quick type check for most options so that the user can check which plugin has invalid configuration on startup. This does not prevent errors, resulting from modiying "g:plugs" after running "plug#end()". Plugin repo is added to the error message for convenience. Most users should expect no noticeable difference in startup time.
Close: #930
Related: #936
This allows Neovim to bypass the shell and run git directly.
Vim still needs the shell because of how commands are collapsed on Windows and because setting the job's working directory via "cwd" does not work.
Refactored s:clone_opt to a list to make this possible.
* s:system supports list type for command
Objective is to reduce batchfiles on Windows.
List type gives more flexibility on s:system()
on how to pass the shell command to the builtin system().
If system() supports list type for command
and there is no working directory, run it directly on system().
Targets Neovim only.
Else, convert the list to an escaped command
so that the user's shell can execute it.
Neovim's system() does not support working directory system()
so consider refactoring s:system to use a synchronous job.
* Do not escape simple shell arguments
Regexp taken from vim-fugitive s:shellesc().
* Set shellredir on Windows
Prep to use list type for command passed to s:system() within s:spawn()
* Internalize shellredir for s:spawn
s:spawn needs to redirect stderr to stdout for jobs callbacks
but s:system (for old Vim versions) sets shellredir if needed.
* Leverage job api for cwd and stderr
Vim/Neovim support stderr redirection and support error callbacks.
Vim 8 and Neovim can set a job's working directory via 'cwd' key
but it cannot be used as is on Vim because CI fails for the Vim release in Ubuntu Bionic and the latest Vim release.
The existing error message printed when plug#end() is called without
calling plug#begin() doesn't make the dependence on plug#begin()
obvious; I had to go digging in the vim-plug code to discover what I'd
done wrong. This attempts to clarify the error a bit, to make it more
obvious to a user.
Close#926
This fix shouldn't be necessary
because vim-plug's buffer should always have 4 lines
but a buffer can be modified in some cases
before nvim 0.4.0 and vim v8.1.1360.
Changing chcp breaks cmd.exe if switching from multi-byte to 65001.
cmd.exe depends on codepage to parse batchfile
so batchfile cannot have unicode characters.
Target powershell if unicode support is required.
User should fix their terminal font to display unicode characters.
Terminal programs can only use Wide String APIs.
For Vim, this requires +multi_byte feature and `set encoding=utf-8`
in the user's vimrc.
Neovim always defaults to `set encoding=utf-8`.
https://dev.to/mattn/please-stop-hack-chcp-65001-27db
git v2.10.0 adds "log.showSignature" config
so that "git log --show-signature" runs by default.
Changing the commit format via "--pretty" does not prevent the signature
from appearing.
Only "--no-show-signature" prevents this.
Close#728
Warn user if shellslash and shell are incompatible.
Set/unset shellslash for file functions on Windows. Based on https://github.com/junegunn/fzf/blob/16fc6862a89eef0f02d32ab8b365887522719da8/plugin/fzf.vim#L30-L107
Support shellescape for git-bash
Windows user may not set shellslash but wish to use git-bash.
This requires custom shellescape for sh,bash shells
because builtin shellescape() depends on shellslash.
Tested on Vim
It was using s:esc() which escapes spaces with a backslash.
This does not work on Windows.
&shell could be escaped on because of spaces.
See patch-8.0.1455 and related 8.1.x patches that address this
for $SHELL on Unix and git-bash on Windows.
Related #852, #908Close#890
Having a trailing slash with &rtp entries is problematic when removing
them later: if loading on demand is used, s:reorg_rtp might fail to
remove the previous runtime paths.
Test case has been using maktaba (which triggers unsetting s:middle in
s:reorg_rtp), but because of the trailing slashes being used, the
previous rtp entries were not removed.
chcp parsing is fragile because of the system locale. There's no convenient way to parse out the codepage value without regex just by relying on cmd.exe builtins and default binaries in PATH.
Vim can be used to parse chcp output but it requires an additional `system` per `s:system` and `chcp` can change within the same console so caching the value won't work on the terminal.
Powershell supports regex but it has a long startup even with `-NoProfile` so running it when `&shell` is not powershell slows down `:PlugInstall` more.
Neovim provides `stdpath` to abstract the directory paths that it uses based on the OS. It respects `XDG` specification on Linux and uses `%LOCALAPPDATA%` on Windows.
1. Use `stdpath('data') . '/plugged'` for plugin directory.
2. Use `stdpath('config')` for user-config directory. This is `~/.config/nvim/` on Linux and `%LOCALAPPDATA%\nvim\` on Windows.
* output of chcp was not parsed correctly
On Windows, when wrapping a batch command with the function s:wrap_cmds(),
when calling 'chcp' to get the current code page, the code assumed that
the output was in the format "active code page: XXX" (where XXX is the
code page), whereas the actual output is localized (for instance, in
French, the output would be: "page de code active: XXX").
The parsing of the output relied on that, and this failed for a
message different from "active code page" (i.e., English).
This patch changes the parsing to split the output of chcp on the colon
instead of spaces. Assuming that the output is always in the format
"<localized message>: XXX", regardless of the locale, hopefully, this is
a bit more robust.
Travis defaults to it since this year.
Ubuntu trusty reached EOL since April 2019.
Ubuntu bionic provides Vim 8.
Simplify OS install with Travis' apt addon
so that only Vim nightly will be compiled from source.
Ubuntu 14.04.6 provides ruby v.1.9.3, not 1.8.x or older.
Use it to test the ruby provider for backward compatibility.
Signed-off-by: Jan Edmund Lazo <jan.lazo@mail.utoronto.ca>
Excerpt from `:h shell-powershell`:
To use powershell (on Windows):
set shell=powershell shellquote=( shellpipe=\| shellxquote=
set shellcmdflag=-NoLogo\ -NoProfile\ -ExecutionPolicy\ RemoteSigned\ -Command
set shellredir=\|\ Out-File\ -Encoding\ UTF8
Previously, `:PlugDiff` would show every new commit to a plugin's git
repo. This makes sense for the general case, but makes less sense when a
plugin lives in a subdirectory of the repo (and is configured with the
'rtp' option). This makes it difficult to determine which commits relate
to the plugin and which are unrelated.
This changes `:PlugDiff` to filter out any commits outside of the 'rtp'
folder.
Some consequences:
* This does not change the `:PlugUpdate` UI. This means `:PlugUpdate`
may pull down non-plugin commits, display that it has updated the
plugin, and then `:PlugDiff` will show no updates (since such commits
fall out of the 'rtp' path).
* It also means there's no UI to revert non-plugin updates, as they
don't show up in `:PlugDiff`.
Git v2.8 introduced support for parallel fetching of submodules. This
can result in a substantial performance boost when performing a "git
submodule update" in a repository with many submodules, potentially
further magnified when updating submodules recursively.
This commit introduces a check on the Git version to see if it supports
parallel submodule fetching. If it does, we add the '--jobs=N' argument
to our call to "git submodule update", where N is g:plug_threads or the
vim-plug default of 16.
This allows you to override the default behavior of `o' or `<cr>'
binding in PlugDiff window.
e.g.
" Move to preview window
autocmd! FileType vim-plug nmap <buffer> o <plug>(plug-preview)<c-w>P
Close#749Close#768
`writefile()` always output LF without CR each lines.
But batch file on Windows needs CR and LF, at end of lines.
And if the path of home directory contains non-ASCII
characters like Japanese username (e.g. `C:\Users\太郎`),
batch file without CR can't be executed correctly.
Close#635Close#668Close#539
Use a temporary batchfile for :!, system(), and jobs and run it in cmd.exe.
This bypasses Vim/Neovim issues in Windows and reduces the need to set more options.
Also, s:shellesc_cmd works in a batchfile only.
Set shellredir for system() in Windows
$SHELL sets the default value of 'shell' (see :h 'shell').
This affects shellredir but cmd.exe requires '>%s 2>&1'.
When loading 'deoplete.nvim' for the 2nd time during InsertEnter
manually, the `s:dobufread` (or `s:lod` itself) prevents it to work
properly - likely because the plugin gets resourced.
Maybe there could be a way to force this (and reload plugins always),
but by default it seems to make sense to skip already loaded plugins.
Checking out a branch or tag may not succeed if there is a path entry
with the same name. This commit appends -- to git checkout commands to
clarify that we're not referring to a file or a directory.
Close#602
Installer is not synchronous when started on VimEnter since
has('vim_starting') returns 0 in that case. We needed a way to make the
installer synchronous.
Related: #574
- Same output format across different installers
- Avoid unnecessary construction of result string
- Fix multi-line error logging when exit_cb is called before out_cb
Having VisualHostKey enabled for ssh can cause lines containing a single
"^M" character to appear in the git output, which causes E684 to be
thrown if not filtered out.
YouCompleteMe's post-installation script only seems to work with python2
(this is actually mentioned in their docs).
It seems that the sample documentation uses the system default python
(which is no longer always python2), so let's better be explicit about
it.
Now `:PlugInstall!` and `:PlugUpdate!` forces to update submodules.
`:PlugInstall!` now also checks out to new specified branch, tag, or
commit.
Also added tests for changing branch, tag, or commit.
Prevent escape sequences in the preview window from PlugDiff
Steps to reproduce:
1. `git config color.ui always`
2. `PlugDiff`
3. See the content of a commit with `o`
This commit extracts duplicate logic out of three installers.
Pros.
- Better maintainability
- Easier to add/extend the features
- Fixes a bug when 'commit' option is used, submodules are updated
before the designated commit is checked out
Cons.
- The whole process takes slightly longer due to lost parallelism after pull
- Especially, submodule updates are not parallelized
- However, this shouldn't matter much in practice as there are few
plugins that heavily rely on submodules
* Do not map during prepare when function can be async.
Delay to s:finish()
* Always create new buffer, due to mappings change.
* Don't show Pending updates when no appropriate plugs.
* No need for print brackets on python 2.
Previously, syntax highlighting was turned on unconditionally in the
buffers opened by commands such as `:PlugUpdate`. With this commit,
highlighting is not applied if the user has disabled syntax highlighting
by calling `syntax off` after `call plug#end()` in their `vimrc`.
- The output file is no longer executable but a source-able vim script
- PlugSnapshot FILENAME to prompt the user if the file already exists
- Add PlugSnapshot! FILENAME variant to force overwrite existing file
- Apply -complete=file option to PlugSnapshot command
Specify PIPE as stdin for subprocess.Pipe for gvim.exe.
Also fixes some ruby implementation.
* windows doesn't have pgrep.
* windows can't handle SIGTERM.
* windows can't handle /dev/null
* redraw always
* Instead of having a fixed sleep duration, start a thread
that monitors the subprocess.
* Started thread is joined, if process finishes thread returns
before timeout.
* Should increase speed and reduce overhead on CPU.
* Refactor Command class for easier reading.
* Some other minor clean ups & method renames.
* Change travis to use `env` and `matrix` to select builds.
* Use case instead of ifs to select behaviour.
* Buffer updates now managed by __main__ loop with buf_q.
* Synchronous neovim install temporarily provided by python installer.
* Known issues:
* No ctrl-c/interrupt support on nvim.
* Graphical bug: https://github.com/neovim/python-client/issues/103
Checking for the existence of `jobwait()` function seems like a more
robust way to check if the current Neovim process meets the requirement
of the parallel installer. This condition precludes a very old version
of Neovim built even before the initial Job API was implemented.
The older documentation of vim-plug suggested that the second string
parameter to Plug command was for both branches and tags. But we've
realized that tags and branches are not interchangeable (#174), and it
is now recommended that the user explicitly specify whether it's a
branch or a tag. The now-undocumented second parameter is currently for
branches, but I'd like to change it for tags, as the use of tags are
much more common.
* Main differences from ruby are inside Command class, notably it
poll/sleeps on the subprocess to check output & timeout.
* Another difference, interrupt is not instaneous due to checking var.
* Otherwise, I mainly just mirrored code into objects.
* Note that due to GVim freeze, disabling use on Windows
This makes it easier and faster to kill spawned git processes.
Repetitive `pgrep` command for collecting PIDs is no longer required.
Unfortunately, :pgroup option is not supported on Ruby 1.8 so we can't
remove the old implementation of `killall`.
In order to install Gist as a plugin, you have to modify its `dir` to be
the `plugin` subdirectory of the normal path of the plugin if it were an
ordinary pathogen-compatible Vim plugin.
Plug 'https://gist.github.com/952560a43601cd9898f1.git',
\ { 'dir': g:plug_home.'/xxx/plugin', 'rtp': '..' }
PlugClean command was incompatible with the above trick, so it was
updated not to remove the parent directory of `dir` of each plugin.
There's a fairly widespread convention which recommends indicating the default
confirmation option in uppercase. Granted, the number of applicable instances
in vim-plug is currently only two, but perhaps this'll save users with such an
inclination having to hit Enter just to sate their curiosities.
In the recent versions of NeoVim, jobstart() does not return
monotonically increasing numbers, this caused vim-plug to miss updating
the progress of a task when the job ID for the task is already
reassigned to a new task.
File diff suppressed because it is too large.
Load diff
Reference in new issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Junegunn Choi
