Ⅰ. 插件描述
Grep 2G source code in 0.23 second.Ⅱ. 基本信息
|
Ⅲ. 安装方法
使用Vundle管理器安装
在你的.vimrc下添加:Plugin 'baohaojun/beagrep':source %
:PluginInstall对于Vundle版本 < 0.10.2，请用上面的Bundle替换Plugin。
使用NeoBundle管理器安装
在你的.vimrc下添加:NeoBundle 'baohaojun/beagrep':source %
:NeoBundleInstall使用VimPlug管理器安装
在你的.vimrc下添加:Plug 'baohaojun/beagrep':source %
:PlugInstall使用Pathogen管理器安装
在终端中运行以下命令:cd ~/.vim/bundle
git clone https://github.com/baohaojun/beagrepⅣ. 文档说明
- beagrep
Grep 2G source tree in 0.23 seconds. A speed up of more than 10 fold.
It uses a search engine before using grep. The search engine is
beagle, thus the name beagrep.
For more details, visit [http://baohaojun.github.com/beagrep.html] (man page included).
- How does it compare to ag
[https://github.com/ggreer/the_silver_searcher] claims to be very fast, beating [http://beyondgrep.com/] and
GNU grep.
I compared ag and beagrep on my laptop:
| Model | CPU | Memory |
|---|---|---|
| Thinkpad T420 | Intel(R) Core(TM) i5-2520M CPU @ 2.50GHz | 8G |
** ag speed
+BEGIN_EXAMPLE
cd ~/src/android
One nice thing about ag is that it filters
VCS and binary files automatically.
time ag readlink . +END_EXAMPLE
The first time it took 3 minutes, the second time it took 8 seconds,
very impressive! I think if I had learned about it earlier I probably
wouldn't have started working on beagrep in the first place, because I
would have thought it is quick enough already.
** beagrep speed
+BEGIN_EXAMPLE
time beagrep -e readlink+END_EXAMPLE
The first time it took 30 seconds, the second time it took 0.42
second.
Another test is done on Linux kernel source code, where ag takes
1m35s/1.8s for cache hot/cold, while beagrep takes 15s/0.25s
respectively.
(The claim that it took 0.23 second to grep 2G Android source code is
achieved on my Dell Optiplex 960).
- Pros and cons of beagrep
** Pros
- Very fast
- Output format compatibility with grep, so it can be used by, for
e.g., Emacs grep-mode directly. - Match not only file content, file names too, like locate(1).
** Cons
- You need build the search engine database beforehand (the first
time you do this will take a long while, but subsequent updating is
reasonably fast)
- Works on whole words only: can not use partial word (=beagrep -e
readli=) to find =readlink=. - Other grep like tools
Ack's author Andy Lester has written [http://betterthangrep.com/more-tools/] about all kinds of
tools for searching source code.
Of particular interest to beagrep is Google's Code Search tool.
- To-do
- Push beagrep into Debian distribution.
- Acknowledgment
Thank the Beagle project for making the original engine. And of
course, the Apache Lucene project.
Thank LDD's author for saying ``(LDD) is the result of hours of
grepping'' and getting me hooked to grep forever.
Last but not least, I submitted this README on [http://redd.it/14tybj], and has refined
it several times according to the comments there. To those who
commented: thanks, I have learned from you!