Shell - Remove everything except a single folder - bash

I have a directory containing one single folder named .git along side various other files and folders. I want to delete all files and all folders, except for my .git directory.
Note that if any of the folders contain a subfolder named .git, I want to remove those as well. Only the root git repo should be preserved!

This is that simple as :
shopt -s extglob dotglob
rm -rf !(.git)
See http://mywiki.wooledge.org/glob

Related

Copy whole directory but exclude all folders and subfolders with certain name

I'm not allowed to use rsync on the cluster I'm working on so I need to use cp. I want to copy a large directory including all files and subfolders etc. but without any folders that have the name "outdir".
I tried cp -r -v ./!(outdir) ../target-directory/
but it still copies all folders and contents in deeper directories with the name outdir. It only included the outdir folders in the highest directory.
I also tried cp -r ./*/!(outdir) ../target-directory/ but that one copied all files into the folder without keeping any hirarchy or folders etc.
I also tried certain find commands but it didn't work, but maybe I was just doing something stupid. I'm a beginner with bash so if you could explain your answer and what the flags etc. do that would really be helpfull, I've been trying forever now, on what I think shouldn't be that hard to do.
Instead of cp, you can use tar with option --exclude to control what you want copied or not.
The full command is:
tar --exclude="outdir" -cvpf - . | (cd TARGET_DIRECTORY; tar -xpf -)
So any path that contains the "outdir" pattern will be excluded.
Without the --exclude option, it will copy the entire structure of your current directory under TARGET_DIRECTORY.
You can replace the . in the first tar by your desired source directory.

zip all files and folders recursively in bash

I am working on a project, where compilation of the project involves, zipping up various files and folders and subfolders (html/css/js) selectively. Working on the windows platform, and I could continue to just use the CTRL+A and then SHIFT-click to unselect, but it does get a little tedious. I am working with cygwin, so I was wondering if it is possible to issue a command to zip selected files/folders recursively whilst excluding others, in one command? I already have zip command installed, but I seem to be zipping up the current zip file too and the .svn file too.
I would like this to be incorporated into a shell script if possible, so the simpler the better.
After reading the man pages, I think the solution that I was looking for is as follws:
needs to recurse directories (-r),
needs to exclude certail files/directories (-x)
It works in the current directory, but the . can be replaced with the path of any directory
zip -x directories_to_exclude -r codebase_latest.zip .
I have incorporated this into a short shell script that deletes files, tidy up some code, and then zips up all of the files as needed.
You should read man page of zip command:
-R
--recurse-patterns
Travel the directory structure recursively starting at the current directory; for example:
zip -R foo "*.c"
In this case, all the files matching *.c in the tree starting at the current directory are stored into a zip archive named foo.zip. Note that *.c will match
file.c, a/file.c and a/b/.c. More than one pattern can be listed as separate arguments. Note for PKZIP users: the equivalent command is
pkzip -rP foo *.c
Patterns are relative file paths as they appear in the archive, or will after zipping, and can have optional wildcards in them. For example, given the cur‐
rent directory is foo and under it are directories foo1 and foo2 and in foo1 is the file bar.c,
zip -R foo/*
will zip up foo, foo/foo1, foo/foo1/bar.c, and foo/foo2.
zip -R */bar.c
will zip up foo/foo1/bar.c. See the note for -r on escaping wildcards.
You can also have a look HERE

How can I delete contents in a folder using a bash script?

I would like to clear out my /bin folder in my project directory. How can I do this?
I tried rm -rf ~/bin but no luck
~ is a shorthand to a current user home directory. So unless it's also your project directory you are doing something wrong. Other than that, clearing a directory would be
rm -rf ~/bin/*
And if you also want to clear the hidden files
rm -rf ~/bin/* ~/bin/.[a-zA-Z0-9]*
Make sure you are not doing
rm -rf ~/bin/.*
especially as root as it will try to clear out your entire system.
UPD
Why? Since wildcard (*) is interpreted by shell as zero or more characters of any kind the .* will also match . (current directory) and .. (parent directory), which will result in going all the way up and then down, trying to remove each file in your filesystem tree.
You should say "... my bin folder", not "my /bin folder". /bin is an absolute path, bin is a relative path.
rm -rf ~/bin removes $HOME/bin, so not what you want either.
Now, it depends on where you are: if you are in your project directory when you type the command, just type rm -rf bin.
rm -rf ~/bin/{*,.[^.]*}
would delete all files and directories in ~/bin/, including hidden ones (name starts with .), but not the parent directory (i.e. ..).
The .[^.]* matches all hidden files and directories whose name starts with a dot, the second char is NOT a dot, and with or without more chars.

Remove file from different directory

How do I remove certain files from a different directory than $PWD using the bash shell script.
Looking at the documentation for rm, it appears that rm only works in $PWD.
Am I forced to use this method:
oDir=$PWD
cd directorytoremovefiles
rm files
cd oDir
rm certainly does work for deleting files in another directory.
Whatever gave you that idea from the man page, I certainly hope it's not this:
rm removes each specified file. By
default, it does not remove
directories.
The documentation you refer to, talks only about having write & execute permission to the directory you are deleting from.
So you only need:
rm directorytoremovefiles/files
As pointed out by YYC,
rm $DIR/files
rm will take any path, relative or absolute. If there is no slash at the beginning of directorytoremovefiles then it is a relative path and you may need to store PWD for later. However, you can do this with pushd and popd or a cd - once you're finished. Or, if you run the cd and rm commands in parens they will run in a subshell, like this: ( cd directory; rm files) then your working shell will not change directory.

Ignoring directories in Git repositories on Windows

How can I ignore directories or folders in Git using msysgit on Windows?
Create a file named .gitignore in your project's directory. Ignore directories by entering the directory name into the file (with a slash appended):
dir_to_ignore/
More information is here.
By default, Windows Explorer will display .gitignore when in fact the file name is .gitignore.txt.
Git will not use .gitignore.txt
And you can't rename the file to .gitignore, because Windows Explorer thinks it's a file of type gitignore without a name.
Non command line solution:
You can rename a file to ".gitignore.", and it will create ".gitignore"
It seems that for ignoring files and directories there are two main ways:
.gitignore
Placing .gitignore file into the root of your repository besides the .git folder (in Windows, make sure you see the true file extension and then make .gitignore. (with the point at the end to make an empty file extension))
Making the global configuration ~/.gitignore_global and running git config --global core.excludesfile ~/.gitignore_global to add this to your Git configuration
Note: files tracked before can be untracked by running git rm --cached filename
Repository exclude - For local files that do not need to be shared, you just add the file pattern or directory to the file .git/info/exclude. Theses rules are not committed, so they are not seen by other users. More information is here.
To make exceptions in the list of ignored files, see this question.
To ignore an entire directory place a .gitignore of “*” there.
For example,
Example System
/root/
.gitignore
/dirA/
someFile1.txt
someFile2.txt
/dirB/
.gitignore
someFile3.txt
someFile4.txt
Goal
ignore the contents of dirB/
Top Level (/root/.gitignore)
You could just “dirB/“ here
Ignored Directory (/root/dirB/.gitignore)
Or you could “*” here
Git watches for gitignore at every step of the file system. So here I choose dirB/.gitignore as “*” to ignore dirB/, including all files and subdirs within.
Done ☺️
To instruct Git to ignore certain files or folders, you have to create .gitignore file.
But in Windows Explorer you have to provide a name for the file. You just cannot create file with just an extension. The trick is that create a empty text file and go to command prompt and change the name of the file to .gitignore:
ren "New Text Document.txt" .gitignore
Now open the file with your favorite text editor and add the file/folder names you wish you ignore. You can also use wildcards like this: *.txt.
I had some issues creating a file in Windows Explorer with a . at the beginning.
A workaround was to go into the commandshell and create a new file using "edit".
If you want to maintain a folder and not the files inside it, just put a ".gitignore" file in the folder with "*" as the content. This file will make Git ignore all content from the repository. But .gitignore will be included in your repository.
$ git add path/to/folder/.gitignore
If you add an empty folder, you receive this message (.gitignore is a hidden file)
The following paths are ignored by one of your .gitignore files:
path/to/folder/.gitignore
Use -f if you really want to add them.
fatal: no files added
So, use "-f" to force add:
$ git add path/to/folder/.gitignore -f
You can create the ".gitignore" file with the contents:
*
!.gitignore
It works for me.
In Windows there's an extra catch with slashes. Excluding a single directory in .gitignore with
dir_to_exclude/
will possibly work, but excluding all directories with
/
causes problems when you have file names with spaces (like my file.txt) in your directory: Git Bash escapes these spaces with a backslash (like my\ file.txt) and Git for Windows doesn't distinguish between / and \.
To exclude all directories, better use:
**/
Two consecutive asterisks signify directory contents.
Just in case you need to exclude sub folders you can use the ** wildcard to exclude any level of sub directory.
**/build/output/Debug/
Also in your \.git\info projects directory there is an exclude file that is effectively the same thing as .gitignore (I think). You can add files and directories to ignore in that.
When everything else fails try editing the file
/.git/info/exclude
and adding the directories you want to the end of the file, like this:
# git ls-files --others --exclude-from=.git/info/exclude
# Lines that start with '#' are comments.
# For a project mostly in C, the following would be a good set of
# exclude patterns (uncomment them if you want to use them):
# *.[oa]
# *~
assets/
compiled/
I added the folders "assets" and "compiled" to the list of files and directories to ignore.
I've had some problems getting Git to pick up the .gitignore file on Windows. The $GIT_DIR/info/exclude file always seems to work though.
The downside of this approach, however, is that the files in the $GIT_DIR directory are not included in the check-in, and therefore not shared.
p.s. $GIT_DIR is usually the hidden folder named .git
On Unix:
touch .gitignore
On Windows:
echo > .gitignore
These commands executed in a terminal will create a .gitignore file in the current location.
Then just add information to this .gitignore file (using Notepad++ for example) which files or folders should be ignored. Save your changes. That's it :)
More information: .gitignore
I assume the problem is that your working tree is like:
a-cache/foo
a-cache/index.html
b-cache/bar
b-cache/foo
b-cache/index.html
.gitignore
... with the .gitignore you describe. This will give you git status output like:
$ git status
# On branch master
# Untracked files:
# (use "git add <file>..." to include in what will be committed)
#
# .gitignore
# a-cache/
# b-cache/
... if the index.html files have not yet been added to the repository. (Git sees that there are unignored files in the cache directories, but it only reports the directories.) To fix this, make sure that you have added and committed the index.html files:
git add *cache/index.html
git commit -m "Adding index.html files to the cache directories"
... and your git status will then look like:
$ git status
# On branch master
# Untracked files:
# (use "git add <file>..." to include in what will be committed)
#
# .gitignore
nothing added to commit but untracked files present (use "git add" to track)
(Obviously you do want to commit .gitignore as well. I was just being lazy with this test case.)
On Windows and Mac, if you want to ignore a folder named Flower_Data_Folder in the current directory, you can do:
echo Flower_Data_Folder >> .gitignore
If it's a file named data.txt:
echo data.txt >> .gitignore
If it's a path like "Data/passwords.txt"
echo "Data/passwords.txt" >> .gitignore.
I had similar issues. I work on a Windows tool chain with a shared repository with Linux guys, and they happily create files with the same (except for case) names in a given folder.
The effect is that I can clone the repository and immediately have dozens of 'modified' files that, if I checked in, would create havoc.
I have Windows set to case sensitive and Git to not ignore case, but it still fails (in the Win32 API calls apparently).
If I gitignore the files then I have to remember to not track the .gitignore file.
But I found a good answer here:
http://archive.robwilkerson.org/2010/03/02/git-tip-ignore-changes-to-tracked-files/index.html
Just create .gitignore file in your project folder Then add the name of the folder in it for ex:
frontend/node_modules
This might be extremely obvious for some, but I did understand this from the other answers.
Making a .gitignore file in a directory does nothing by itself. You have to open the .gitignore as a text file and write the files/directories you want it to ignore, each on its own line.
so cd to the Git repository directory
touch .gitignore
nano .gitignore
and then write the names of the files and or directories that you want to be ignored and their extensions if relevant.
Also, .gitignore is a hidden file on some OS (Macs for example) so you need ls -a to see it, not just ls.
Temporarily ignore a directory/file that was already in git:
I have a lot of projects in a multi-project gradle project and they can take a long time to delete them, and they're all pretty much the same but different. From time to time I want to remove those from the gradle build by deleting them altogether. git can get them back after all. However I don't want them showing up in git status either. So I use the following simple procedure;
delete files and folders I don't want.
verify build still works
tell git to ignore the deleted files for a bit (we can get them back)
git ls-files --deleted -z | git update-index --assume-unchanged -z
--stdin
go about life without the dirs until you want them back. Then run the same command as before but switch out assume-unchanged for no-assume-unchanged
git ls-files --deleted -z | git update-index --no-assume-unchanged -z
--stdin

Resources