Search Results

Search found 6493 results on 260 pages for 'git bash'.

Page 111/260 | < Previous Page | 107 108 109 110 111 112 113 114 115 116 117 118  | Next Page >

  • Script to run chown on all folders and setting the owner as the folder name minus the trailing /

    - by Shikoki
    Some numpty ran chown -R username. in the /home folder on our webserver thinking he was in the desired folder. Needless to say the server is throwing a lot of wobbelys. We have over 200 websites and I don't want to chown them all individually so I'm trying to make a script that will change the owner of all the folders to the folder name, without the trailing /. This is all I have so far, once I can remove the / it will be fine, but I'd also like to check if the file contains a . in it, and if it doesn't then run the command, otherwise go to the next one. #!/bin/bash for f in * do test=$f; #manipluate the test variable chown -R $test $f done Any help would be great! Thanks in advance!

    Read the article

  • "Reset" colors of terminal after ssh exit/logout

    - by dgo.a
    When I ssh into a remote server, I like the colors of the terminal to change. I use setterm on my remote ~/.bashrc file to get this done. However, when I exit, the terminal colors are not reset to the local ones. I solved the problem, but I am not sure if it is the best solution. This is what I could come up with. On the ~/.bash_logout on the remote server, I put: echo -e "\033[0m" /usr/bin/clear Just out of curiousity: Does anyone know of a better way? (I got the echo -e "\033[0m" line from http://edoceo.com/liber/linux-bash-shell)

    Read the article

  • Custom prompt doesn't work on Mac Terminal

    - by mareks
    I like to use a custom prompt (current path in blue) on my unix machine: export PS1='\[\e[0;34m\]\w \$\[\e[m\] ' But when I try to use it on Mac's terminal it doesn't work: it fails to detect the end of the prompt and overwrites the prompt when I type commands. This also happens when I'm inputting a long command where it wraps over the same line instead of starting a new line. I don't understand why this is the case since I use bash on both machines. Any suggestions on how to remedy this?

    Read the article

  • How do I get `set show-all-if-ambiguous on` in my .inputrc to play nice with the Python interpreter?

    - by ysim
    I noticed that after I added the set show-all-if-ambiguous on line to my ~/.inputrc, whenever I pressed tab to indent a block, it would show me the bash Display all ... possibilities? (y or n) prompt, and leave me unable to indent the actual code. Is there any way to keep that line in my .inputrc but still have the tab key work as expected in the Python interpreter? This is in my VirtualBox Ubuntu 12.04 VM, if it matters. EDIT: Curiously, I now have a different issue with the Python shell that comes with Django -- when I press tab, I get Python tab completion, but only with one Tab press. I've opened a separate question here for it.

    Read the article

  • Shell script only executes partially when run with CRON

    - by binaryorganic
    I've written a shell script that does the following: Retrieve mail from a POP3 account (using GetMail) Save a copy of that email to S3 (using AWS CLI) Email me the filesize of the email The script runs fine manually, and technically runs from CRON, but it only seems to be sending the email. The getmail and S3 bits don't seem to run. Everything I've read seems to hammer home the message that I need to be careful about relative paths and the like when using CRON, but I think I'm using absolute paths everywhere I need to be, so I'm stumped as to what the issue could be. My Shell Script is here: #!/bin/bash # Run GetMail getmail -r /PATH/TO/EMAIL/getmail.email # Save to S3 aws s3 cp /PATH/TO/SCRIPT/email-backup.mbox s3://XXXXXXXXXX/email-backup.mbox # Send Confirmation Email SUBJECT="EMAIL SUBJECT" EMAIL="[email protected]" # Get current filesize FILENAME=/PATH/TO/SCRIPT/email-backup.mbox FILESIZE=$(stat -c%s "$FILENAME") # Email Content EMAILMESSAGE="/tmp/emailmessage.txt" echo "EMAIL BODY" >$EMAILMESSAGE echo "" >>$EMAILMESSAGE echo "Current File Size: $FILESIZE bytes" >>$EMAILMESSAGE # Send the Mail /bin/mail -s "$SUBJECT" "$EMAIL" < $EMAILMESSAGE

    Read the article

  • Mac OS X keeps "old" environment variable around

    - by Xymak1y
    So far I had /Applications/play-1.2.5/ added to my $PATH variable. Now I'm working with 2.2.1, which I installed in /Applications/play-2.2.1 and changed in ~/.bash_profile (which is getting sourced at startup). However, when printing $PATH, 1.2.5 is somehow still around: mbp:~ user$ echo $PATH /usr/local/share/npm/bin:/Applications/play-2.2.1:/usr/local/heroku/bin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/Applications/play-1.2.5:/Applications/XAMPP/xamppfiles/bin/:/opt/X11/bin As far as I now, I only entered $PATH variables in .bash_profile, which looks like this: mbp:~ user$ cat .bash_profile source ~/.git-completion.bash ### Added by the Heroku Toolbelt export PATH="/usr/local/heroku/bin:$PATH" ### Play Framework export PATH="/Applications/play-2.2.1:$PATH" export PATH="/usr/local/share/npm/bin:$PATH" I'm also not sure where the XAMPP extension to the variable comes from. Can I see somewhere which other files are being sourced on startup?

    Read the article

  • I do not understand -printf script

    - by jerzdevs
    I have taken over the responsibility of RHLE5 scripting and I've not had any training in this platform or BASH scripting. There's a script that has multiple pieces to it and I will ask only about the second piece but also show you the first, I think it will help with my question below. The first part of the script shows the output of users on a particular server: cut -d : -f 1 /etc/passwd The output will look something like: root bin joe rob other... The second script requires me to fill in each of the accounts listed from the above script and run. From what I can gather, and from my search on the man pages and other web searches, it goes out and finds the group owner of a file or directory and obviously sorts and picks out just unique records but not really sure - so that's my question, what does the below script really do? (The funny thing is, is that if I plug in each name from the output above, I'll sometimes receive a "cannot find username blah, blah, blah" message.) find username -printf %G | sort | uniq

    Read the article

  • Smart Auto-completion in SVN (and other programs!)

    - by Jimmy
    When I type "svn add path/to/somefile..." and tab to autocomplete, the system should ONLY complete files/directories that are NOT under currently under SVN control. Likewise, when I commit, remove or resolve files, the tab completion should only complete files that are relevant to what I'm doing. This is especially important in SVN where I can waste thousands of keystrokes typing long path and file names, but it of applies to other programs. I know bash has a bash_completion file that can be used to programatically alter this behaviour but I've not found a decent example of SVN completion which actually completes file names rather than SVN command names. My question is: Does anyone have such a setup? Does anyone use a different shell or tool that does something similar? Has anyone given this any thought?

    Read the article

  • Is it possible to have environment variables in the path of the working directory : PS1?

    - by mthpvg
    I am on Lubuntu and I am using bash. My PS1 (in .bashrc) is : PS1="\w> " I like it because I need to paste the working directory all the time. The problem is that the path is always very long and since I use terminator I only have half of my screen's width available to display it... it is ugly and annoying. My command prompt looks like that : /this/is/a/very/long/path/that/i/want/to/make/shorter > I'd like to set in my environment variables : $tiavl=/this/is/a/very/long And then I'll get : $tiavl/path/that/i/want/to/make/shorter > The goal is to have something shorter in the command prompt but I still want to be able to copy paste it and do : cd $tiavl/path/that/i/want/to/make/shorter It is a bit like with $HOME : ~/path/that/i/want/to/make/shorter > I know where I am and I can copy paste the ~. Thanks.

    Read the article

  • Run a specific command from a directory

    - by Cameron Kilgore
    I have a bash script where I need to run an init utility within a directory with a configuration file defined. I don't think it's possible to explicitly tell the utility to run the file as an argument, so what I need to do is go to the directory with the config file, and then run the command. I have some logic in place, but its not working -- the utility never runs. Is there any way I can tell the script to go to this directory, and then run the script? cd /var/www/testing-dev.example.co eval "standardprofile"

    Read the article

  • Install correct libraries depending on 64/32 bit

    - by Rich
    I am using Bash to install a customised version of JBoss, and one of the things I would like to do is install the correct version of the Apache Portable Runtime, which is a native binary. This script could be run on both 32 and 64 bit versions of RHEL. What are my options for identifying which version of the APR to install? I think we only have 32bit and x64-based systems here. I would still like to identify i64 systems so that the script can refuse to install on that type of machine. I am aware of using uname -m and grepping /proc/cpuinfo to find out, but was wondering which approach others would recommend?

    Read the article

  • How to write more than one line in a launcher

    - by seraex
    How can I run three commands in a launcher? My commands are cd /home/seraex/MyDoc rm MyDoc.tgz tar cfz MyDoc.tgz * which will go to my documents folder and delete old backup and make a new backup. At the moment I make a text file and then make a launcher and point it to the file, but I want to delete the file and make the launcher run the commands directly. I'm using ubuntu 10.10 ' ubuntu site says 'Unfortunately launchers do not have access to the Bash environment so you cannot just include the multi commands' when i ggole chaining in launchers. thanks, admin may delete the question '

    Read the article

  • gitosis did not generate projects.list automatically, gitweb can't work.

    - by Readon Shaw
    I setup a gitosis managed git server. git clone is ok. but when I set gitweb via gitweb.conf as below: $projectroot = "/srv/gitosis/repositories"; $git_temp = "/tmp"; $home_text = "indextext.html"; $projects_list = "/srv/gitosis/gitosis/projects.list"; $stylesheet = "/gitweb/gitweb.css"; $logo = "/gitweb/git-logo.png"; $favicon = "/gitweb/git-favicon.png"; Btw, the commet was deleted because of the special symbol # is using as bold prefix. "403 Forbidden - No projects found" is reported when I access gitweb through "http://localhost/cgi-bin/gitweb.cgi" I checked the projects.list file it is empty, is that the reason why gitweb access failed? what would be the correct content? can i add it manually?

    Read the article

  • How to permanently remove xcuserdata under the project.xcworkspace and resolve uncommitted changes

    - by JeffB6688
    I am struggling with a problem with a merge conflict (see Cannot Merge due to conflict with UserInterfaceState.xcuserstate). Based on feedback, I needed to remove the UserInterfaceState.xcuserstate using git rm. After considerable experimentation, I was able to remove the file with "git rm -rf project.xcworkspace/xcuserdata". So while I was on the branch I was working on, it almost immediately came back as a file that needed to be committed. So I did the git rm on the file again and just switched back to the master. Then I performed a git rm on the file again. The operation again removed the file. But I am still stuck. If I try to merge the branch into the master branch, it again says that I have uncommitted changes. So I go to commit the change. But this time, it shows UserInterfaceState.xcuserstate as the file to commit, but the box is unchecked and it can't be checked. So I can't move forward. Is there a way to use 'git rm' to permanently remove xcuserdata under the project.xcworkspace? Help!! Any ideas?

    Read the article

  • Private VCS hosts for school projects?

    - by Ibrahim
    I want to use version control for a partner project for school, but these days it seems like there are no private, free VCS hosts that I could use. I would prefer to use git if possible, but I wouldn't mind SVN. Alternatively, if there aren't any, shouldn't there be some way for me to use git without a central repository? I don't know enough about git, but I assume that is the point of a DVCS, no? I've considered scp'ing a clone of the repository to my school unix account and then giving my partner access to that, but it seems like it would be a bit of a pain. What are your thoughts/suggestions? Edit: I do know of one site called xp-dev, but I'm not sure how much I trust it. But I could use that and use git-svn on my side, since my partner has actually only ever used svn. But still wondering if there are any alternatives.

    Read the article

  • Android Eclipse test projects cannot be used with a project being built in an Android build tree

    - by orospakr
    An Android Java project placed in a git repository and built in an Android tree in /packages/apps needs to have the project files located at the root of the git repository. This is problematic for creating a complementary Test project, which should ideally be included in the same git repository so commits are atomic for both code and tests. Eclipse gets very unhappy if you include the Test project as a subdirectory. Is there an appropriate approach for dealing with this other than creating a second repository?

    Read the article

  • How do I get Eclipse CDT to ignore files

    - by brianegge
    I have a C++ project in Eclipse. The project uses Perforce and Eclipse has the Perforce plugin installed. Everything was fine, until I decided to create a git repo in my project. I created the git repo to snapshot some changes which I wasn't ready to commit. Everything was fine until I refreshed my files in Eclipse. Two problems have occurred: Eclipse found my .git folder, and indexed all of the files inside of it. Eclipse also decided to add all the git file to my pending change list. If I create a new file within Eclipse, I'd like it to add it to Perforce, but if it happens to find a file, I don't want it to do anything with it. I'd also like to give Eclipse a list of file types to always ignore, just like I do with my .hgignore file.

    Read the article

  • Shell Script Sequencing with Rake

    - by Haseeb Khan
    Hi All, I am working on a rake utility and want to implement something mentioned below: There are some shell commands in a sequence in my Rake file. What I want is that the sequence should wait for the previous command to finish processing before it moves to the next one. sh "git commit -m \"#{args.commit_message}\"" do |ok, res| # Do some processing end sh "git push heroku master" So, in the above example what I want is that sh "git push heroku master" shouldn't be executed until the processing in the sh "git commit -m \"#{args.commit_message}\"" do |ok, res| # Do some processing end is completed. Also another nice to have would be that if I can store the output of the shell command in a Ruby variable so it can be used in further manipulation if required. Looking forward to a reply from the fellow community member shortly. Thanks in advance.

    Read the article

  • Confusing .gitignore syntax

    - by tmslnz
    I was reading http://www.kernel.org/pub/software/scm/git/docs/gitignore.html and the 6 points used to explain the ignore patterns seem to be describing a custom variant of a glob search syntax. I am more familiar with Mercurial, which allows to explicitly ignore via glob or regex patterns, no questions asked. Is there anything similar functionality in Git? Can anyone point me to some more exhaustive reference than the Git man page? Best, t

    Read the article

  • GitHub solution for personal repo

    - by Luke Maurer
    So I've got my private SVN repo on my home server, and it has maybe 30 different modules thrown together in it, ranging from abortive throw-away larks to a few endeavors that might actually go somewhere someday. But a recent filesystem failure (BTW, never ever EVER use XFS without a battery-backed hardware RAID) has me spooked and thinking of using a DVCS for all that. I've also just had quite the swig of the Git koolaid, and I've been working with GitHub of late, so that's where I'm looking right now. Of course, it would be silly to shell out major cash for a separate private Git repo for every little project, and I don't want to have to be selective about what I throw up there (I love all my children :-D ), so I'll have to be somewhat creative about this. I can happily use SSH to my home box to use Git the way I've been using SVN, and I'm thinking from there I could amalgamate everything into, say, a big project with 30 submodules, which I then push to GitHub. What'd be a sane way to set this up? Does using submodules sound feasible? How do I sync it all to my private GitHub repo? Cron job? Git hook? I'd love to hear it if anyone's done something similar. I'm not really married to Git or GitHub, so a sufficiently compelling feature of another solution might sway me. But if your answer does involve a different system (especially a different VCS), be advised it'll be a tougher sell :-)

    Read the article

  • Stage untracked files for commit without staging tracked file changes

    - by Blair Holloway
    Oftentimes, when using git, I find myself in this situation: I have changes to several files, but I only want to commit parts of them. I have added several untracked files, which I want to track and commit. Solving the first part is easy; I run: git add -p Then, I choose which hunks to stage, and which hunks remain in my working tree, but unstaged. However, git's patch mode skips over untracked files. What I would like to do is something like: git add --untracked But no such option appears to exist. If I have, say, six untracked files, I could stage them using add in interactive mode and the add untracked option, like so: git add -i a<CR> 1<CR> 2<CR> 3<CR> 4<CR> 5<CR> 6<CR> <CR> q<CR> I feel like there is, or should be, a quicker way of doing this, though. What am I missing?

    Read the article

  • GitHub solution for personal repo

    - by Luke Maurer
    So I've got my private SVN repo on my home server, and it has maybe 30 different modules thrown together in it, ranging from abortive throw-away larks to a few endeavors that might actually go somewhere someday. But a recent filesystem failure (BTW, never ever EVER use XFS without a battery-backed hardware RAID) has me spooked and thinking of using a DVCS for all that. I've also just had quite the swig of the Git koolaid, and I've been working with GitHub of late, so that's where I'm looking right now. Of course, it would be silly to shell out major cash for a separate private Git repo for every little project, and I don't want to have to be selective about what I throw up there (I love all my children :-D ), so I'll have to be somewhat creative about this. I can happily use SSH to my home box to use Git the way I've been using SVN, and I'm thinking from there I could amalgamate everything into, say, a big project with 30 submodules, which I then push to GitHub. What'd be a sane way to set this up? Does using submodules sound feasible? How do I sync it all to my private GitHub repo? Cron job? Git hook? I'd love to hear it if anyone's done something similar. I'm not really married to Git or GitHub, so a sufficiently compelling feature of another solution might sway me. But if your answer does involve a different system (especially a different VCS), be advised it'll be a tougher sell :-)

    Read the article

  • How can I move all my modification to a branch

    - by michael
    Hi, I create a working repository in HG. And I have modified some files. How can i move my all my modification to a branch (a branch that I have not created)? (kind of 'git stash' and the move the stash away change to a branch. Actually, I am not sure how I can do that in git either. If you know, I appreciate if you can tell me in git as well) Thank you.

    Read the article

  • SSH is looking in the wrong place for the public/private key pair on Windows

    - by Michael Morisy
    I'm trying to configure GIT on my Windows XP machine, but SSH keeps creating and looking for the public/private key pair in non-sensical places, e.g. /.ssh/id_rsa Is there a configuration file in the GIT Installation for Windows where I can switch this to my home directory, or another user defined place? By default, it offers to create new key pairs at //.ssh/id_rsa but that gives me the error "Could not create directory '//.ssh'." And when I'm able to find a directory I can create it in, GIT won't look there.

    Read the article

< Previous Page | 107 108 109 110 111 112 113 114 115 116 117 118  | Next Page >