I started using AutoComplPop for automatic code completions. It works great on the single file I am editing, but if file1 is making a reference to a method defined in file2, it doesn't find it.
The docs don't specify if there is a way to make it search a whole project directory, or even just all open buffers, so I can't tell if this is simply not something the plugin does, or if I need to enable something.
I was testing it out on two Ruby files, if that's relevant. Thanks!
Looks like that the cause of the problem is that ACP set the complete option for its purposes to .,w,b,k (see line #125 in autocomplpop/plugin/acp.vim),
call l9#defineVariableDefault('g:acp_completeOption', '.,w,b,k')
while the default value that is used when pressing \<C-n> is .,w,b,u,t,i. And it appears that the very last letter i actually makes the difference: for some reason vim would not use word from an include file opened in a buffer to complete words in another buffer. So, b option is not enough, i must also be included. Adding the following line into my .vimrc helped
let g:acp_completeOption = '.,w,b,u,t,i'
At least it worked for C++ files, but I'm not sure it fixes the problem for the case of Ruby scripts.
Depending on what is on the left of the cursor, ACP (like all the alternatives) decides what completion mechanism to use.
But ACP only uses Vim's default completion mechanisms: if <C-x><C-o> and <C-n>/<C-p> don't provide what you are looking for, ACP won't help. Try them out first.
Oh cool, this plugin looks a lot like neocomplcache but maybe cleaner...looks a little old. Little concerning that there are so many open tickets on that project and no updates in two years.
Anyway, according to the documentation it doesn't...really...say. Very likely its one of the following things:
Your pwd. If the root directory for your source is some/path then that should also be your current working directory. Try typing :cd some/path to see if that makes a difference.
The runtime path rtp. See if adding the directory with your source files to &rtp does the trick.
The path. Same deal as the &rtp setting.
Very likely this plugin is just falling back on the built in ruby omni completion functions bundled with vim. Try help ft-ruby-omni.
I just had the same problem, and I actually found a solution for this.
Apparently you have to set in your .vimrc file the following:
let g:acp_behaviorKeywordCommand = "\<C-x>\<C-i>"
This will make acp look in every file included by your source for completions, as if you were actually typing <C-p>. However, it is slow, after trying it I decided to revert using <C-p> when there are no matches and default behaviour in the other cases.
Related
I'm trying to configure gf to open files under the cursor.
It works always... almost.
I found an issue when the text under the cursor has unfortunately a corresponding directory. In that case, netrw plugin is opened.
Let me give you an example. I am in this code:
[...], MyObject myobject, [...]
I am over MyObject and press gf.
Unfortunately I have in a folder:
myobject <-- a directory
MyObject.java <-- the file to open
netrw is activated.
I tried to check doc to tinker a little bit (suffixesadd, ...), but probably I am missing how to do it properly.
I found this answer, but it is a little bit different in my opinion because in that case the match of the text and the directory were the only 1st one and it was perfect.
Any help?
P.S. what I am trying to do is creating a small vim plugin that could be used to navigate Java projects based on Maven (it's called vim-java-maven).
Just for learning VIM.
As silly as it may sound, Vim considers directories as valid targets so…
:help 'suffixesadd' doesn't help because the directory name is an exact match,
:help 'includeexpr', which is only invoked if there is no match, is not invoked since there is a match.
That behaviour is hardcoded and there is no way to affect it at runtime. The only solution is to write your own Gf() that handles directories more sensibly and map it to gf.
I've got two c++ files that I want to diff with vimdiff. One of them has a lot more function definitions at the start, before both have a common function that I'm actually interested in. However, vimdiff seems incapable to ignore all the function defs before the common one (perhaps because of different arguments).
Is there any way I can give a hint to vimdiff that, say, line xxx in file1.cxx is equals to line yyy in file2.cxx?
I'm open for alternative solutions without vimdiff, but they must be on linux and very preferably command line, since I'm ssh-ing and any graphical interface is a bit uncomfortable.
Vim just delegates the actual work of comparing the files to the external diff utility, cp. :help diff-diffexpr. The help page also shows how a different utility can be used. Unfortunately, I'm not aware of any more "intelligent" or configurable diff tool that would help in your situation.
A workaround might be (temporarily) removing the excess functions that you're not interested in, anyway. With the BlockDiff plugin, you don't actually need to modify the files. Just select the interesting lines in both windows and execute :[range]BlockDiff on them. Only those sections will then be diffed in a separate tab page. (The plugin mentions this requires a GUI, but Vim in a terminal supports tab pages just as well.)
I am trying to use the icicle-search-file/icicle-search-buffer commands, but am having the same problem with both where C-RET seems to perform no operation on completion options once I have already entered the search context pattern.
The operations described are using Emacs 24.3.1 on Ubuntu 13.10 with icicles, workgroups2, and some custom key bindings all done with H-_ or C-c _ to avoid conflicts. I have icicles and workgroups2 set to be on at start in my .emacs file. I don’t have anything custom bound to C-c `.
I have learned to use icicle-search (C-c `) with progressive completion (S-SPC) and it works well, which is why I wanted to figure out these commands.
I have called the icicle-search-buffer command with each of these options just to be sure there is no difference, and the result are the same:
C-9 C-c `
M-s M-s b
M-x icicle-search-buffer RET
I enter the search regex (.* RET - to make things simple) just as with icicle-search, and am then prompted for the buffers to search. I move through the completion options with arrow keys and use C-RET on the buffers I want to search. The result of this is the buffer I chose disappearing from the completion list and no obvious result appearing. Hitting RET at any point just searches the buffer I had active before beginning the search, just as if I had used C-c `. It does not search the currently selected buffer in the completions list.
Does anyone see where I am failing to understand the usage of these commands? I have read through the information on the Icicles - Search Commands, Overview page.
The "Search Multiple Buffers, Files, and Bookmarks" section seems to describe this simply, but beyond using C-RET not much is said that suggests to me where my problem is.
I also followed the link in that section to the Icicles - Multi-Commands page
which was helpful in seeing how icicles modifies vanilla Emacs commands and how multi-commands work in general, but I could not determine from that where my mistake is.
You were doing nothing wrong. (And your description of using Icicles search is a good one.)
I introduced a bug recently.
Should be OK now. Please download the latest version of icicles-mac.el. Or wait (up to a day) for it to be mirrored on MELPA and download it from there. Sorry for your trouble, and thanks for bringing this to my attention.
IMPORTANT:
If you byte-compile Icicles (recommended), then whenever (as now) icicles-mac.el is updated, you
must load icicles-mac.el (not just icicles-mac.elc), then compile it, then *recompile ALL***of the other **Icicles source files as well. You want to make sure that all byte-compiled files are up-to-date.
This is normal for Lisp. Code that depends on macros needs to be byte-compiled anew after loading the updated macros.
Two part question:
First, when using VIM what process do I take and what keys do I type to "go to definition" or "go to declaration" etc.? This document might be the answer to my question, but I can't get it to work, so I'm unsure. It looks like its merely text matching the string rather than finding the true definition. If I can get this to work, then will I be able to jump outside of the current document to a definition/declaration? Or does this only work within a single document?
Second, how do I make this work specifically with the Go programming language? It sure would be nice to "click" the Client in
clnt := &http.Client{Transport: tr}
And be taken to the actual code that defines an http.Client.
Possible? How?
As you guess, gd (and other commands) is merely text matching, vim doesn't understand the syntax as it is just a text editor, :h gd will explain how gd works.
Usually, 'go to definition' is brought by using CTRL-] and tag files. A user manual about this topic can be read by :h 29.1.
First you need to generate a tags file for your project, as latest Exuberant Ctags has supported golang (from here), command
cd /path/to/your/project
ctags -f tags -R --fields=+K+a
will do the job.
Second, open vim, by default vim will find tag files under working directory (according to 'tags' option), if the tag file is found successfully, then CTRL-]` should works well.
Also check two useful plugins Tagbar and Easytags.
For golang, you can use the application godef to do it. The pluging vim-go helps you on setting everything, so, you just type 'gd' in a definition and it goes to the exact definition.
https://github.com/fatih/vim-go/blob/master/doc/vim-go.txt
I am rather new to VIM. I got some source code and this is a mess. At a first sight I would like at least to get a clear and organised view of the code, so I like to get it rightly formatted, I mean indented depending on the depth of the functions and so.
I wonder if it can be done with VIM, and otherwise which other commandline tools for that can you recommend.
Thanks
While vim is a true Swiss-knife I still prefer external tools for some jobs. This approach is some times much more intuitive and easy to remember than using the built-in equivalent.
In the case of indenting, I filter the whole file buffer through astyle. The astyle parameters are much easier to grasp in a couple of minutes, especially if you are not a vim guru. Also astyle provides much more flexibility in fine-tuning the output.
First install astyle:# apt-get install astyle
Then inside vim:
:%!astyle (simple case - astyle default mode is C/C++)
or
:%!astyle --mode=c --style=ansi -s2 (ansi C++ style, use two spaces per indent level)
or
:1,40!astyle --mode=c --style=ansi (ansi C++ style, filter only lines 1-40)
you can do the following:
gg=G
I would highly recommend clang-format nowadays. It allows simple integration of clang-format into Vim, once you have clang-format installed:
http://clang.llvm.org/docs/ClangFormat.html#vim-integration
It is the only code beautifier that really understands your C++ code, and it is really intelligent to beautify the code more like a human being than a machine. E.g.:
void TestFunction(int argument1, int argument2,
int argument3);
void TestFunctionVeryLongName(int argument1,
int argument2,
int argument3);
void TestFunctionWithRidiculouslyLongName(
int argument1, int argument2, int argument3);
Vim will definitely do this, although the results may not be perfect:
First, select the entire file in visual mode: ggVG
Then hit = to reindent everything.
You can learn more about the equal command with: :help =
There is also a Vim plugin relying on clang-format: vim-clang-format
Then you can simply map the formatting command to whatever suits you.
There is a vim plugin that enables formatting on your code from within vim. It's called vim-autoformat and you can download it here:
https://github.com/vim-autoformat/vim-autoformat
It integrates external code-formatting programs into vim. For example, if you want to format C, C++, C# or Java code, you need to install the program astyle, and vim sets it as the format program automatically.
I don't write C++ code, but I write some Java code.
Instead, Vim supports the formatting of some common languages.
I have set up a short cut for me to format the whole code in the buffer.
It will return to the line I just edited :)
" format the file
map <leader>fm gg=G'.
A generic solution along the lines of m000's idea is to use UniversalIndentGUI as an external tool.
Just had to solve this exact problem, so I thought I'd contribute to save others some time.
You can use gg=G to indent your code. But things get hard to understand the moment you want to tweak how that auto-indenting happens. Therefore, if you only care that errant whitespace is removed and don't really care about formatting style, gg=G is the quickest way to go about it, because its built-in.
If you do want to control the style (for example, you're trying to make your code conform to a style guide), then you're going to need an external tool to process your file. You can invoke that tool from within vim with: :%!<toolname> <options>. This pipes the file through the tool and re-loads the processed result. (You can obviously use this for anything else you want to do to your file too)
So the next question is, what external tool should you choose? Regardless, the method is the same:
Install the tool of choice
Make sure its in your path
Add a line to your vimrc file that creates a shortcut key to use so you save time
Use it.
Now, which tool you use depends on the style you're trying to replicate. If you're trying to replicate a widely used style, then chances are astyle is all you need.
If you're trying to replicate a custom style, then you will need two things:
UniversalIndentGui - a front end that lets you play around with various options and live-preview their effect on the source file
A set of source code formatting tools installed and in your path
Between uncrustify and greatcode, you should be able to completely replicate the style you want.
Actually, I lied. There is another way and its called clang-format. However, you're going to want to read the documentation on it and its still in early stages so some options don't work very well. It is a beautiful tool though (definitely the smartest of the lot because constructs an AST of your code) and it is even available for Windows.
If you're going to take the time to read the manual, you also want to check out GNU Indent.
Of course, there is the last way, which is actually taking the time to learn vim's indent rules and writing one for your style. It will take time, but it will work with gg=G.
Some notes on astyle vs uncrustify vs greatcode:
Astyle is good for general formatting, but can't do things like align the declaration of variables and re-style comments very well.
Uncrustify can do a LOT of stuff that astyle can't, but be prepared to spend an hour playing around until you've found the correct combination of options you need. (Or if you feel like wasting a lot of time, use genetic algorithms to figure out the best combination of options for your style and when you do share the code and give me a link so I can use it too :) )
Note that you don't have to choose one tool. With vim, you can map one keystroke to execute several commands in succession, so theoretically you could use a combination of these tools to get exactly what you're looking for.
Last but not least, here's an excerpt from my .vimrc file, where I have mapped F12 to invoke astyle with some options:
"A2 = attached brackets
"-s8 indent 8 spaces
"-xc attached braces to class declarations
"-xj remove braces for single statement ifs and elses
"-c convert tabs to spaces in the non-indentation part of the line
map <F12> :%!astyle -A2 -s8 -xc -xj -c<CR>
Don't judge me on the style. Just use the tool to reproduce what you want.