I have a Rails 3.2.x app which I recently removed database.yml out of version control in git bit adding /config/database.yml to .gitignore in my app directory and did a git rm --cached config/database.yml in my app directory. I have the database.yml file on my production and staging server under app/shared/config/database.yml and I symlink to it via my Capistrano deploy.rb file.
In my dev environment on my master branch I recreated the database.yml file and I can continue development, but every so often the file disappears from the directory entirely and I have to recreate it. Note, I have Sublime Text 2 open usually with database.yml being one of the files that I have open most of the time.
Things will run fine for a while but all of a sudden the file will disappear from the master branch config directory on my local disk and I have to recreate it.
What could be causing this file to disappear randomly? git ignores it and I just use a local copy in my master branch. Could this be from switching branches? Or because my text editor has it open?
It's not a huge deal because I have a backup copy and I can just drop in, but this only started since I pulled it out of version control/git.
Any thoughts on why this is happening? There are times when I need to demo the app and don't want to have any "hiccups".
Any help is greatly appreciated. I don't want to have to commit database.yml back to version control, that's for sure but this problem is perplexing as to why it disappears randomly.
Related
I was working on a project in IntelliJ, where I put in some code to delete some files from my project. This accidentally deleted the whole project from the filesystem (Windows 10). The only file left is the .idea directory that probably stores some IntelliJ configuration. The Local History option in IntelliJ has been disabled ever since (which means it wiped the Local History of this project), so there is no way of recovering the project via Local History either.
I used recovery programs like Recuva and EaseUS which for some reason didn't recover the lost files (EaseUS in particular shows files from 2 or 3 days ago instead of today).
I also doubt I can recover these files via Git since Git does not use a single file to store its history.
Any suggestions?
I'm new to JavaFX 8 and the IntelliJ IDE. I have a JavaFX8 project that works but not as I would like. I'd like to try another approach but the substantial changes may not work. I don't want to loose code I have working.
To save code I have working, I've been creating a new project and then locally copying all the folders(.idea, out, src) and files except .iml, of the working project into the appropriate folders in the new project with the newly generated .iml.
This always seems to work but is it proper procedure?
I'm not on a team of developers and have yet to learn Git/GitHub.
Please advise. Thanks.
Maybe you should learn how to use a Version Control System like Git, then you can create a project repository and have different branches for things you want to try out. Keeping the working code in your master branch will prevent you loosing your working code. Also, when using a vcs you can always revert to versions of your code that have been working. The IntelliJ Idea IDE has perfect support for working with all different types of version control systems. If you don't want to learn any forms of vcs then there is no other way to "backup" your working code.
Is it proper procedure? It's probably not how most people would go about achieving what you want to achieve but it's certainly workable. If you wanted to stick with that for simplicity now, I'd copy the whole directory structure, delete the .idea and .iml files, and then create a new project in IntelliJ on that clean copy: IntelliJ will automatically set up folder structure based on the existing source without you having to go through any additional manual setup.
If you're willing to experiment with the git route, to achieve the basics of what you want to achieve is not very complicated and I've written a small quick-start below. IntelliJ offers very good support for Git, and once your repository is created you can do everything you need from the IDE. I'm going to assume you're working on Windows, although the steps shouldn't be too far removed on other platforms.
Install Git
You can download and install Git from https://git-scm.com/download/win, which will install a command shell called Git Bash.
One-off setup for your project
Open up git bash and go into the directory containing your source. Rather than seeing separate drives as Windows does, Git Bash assumes there is a logical 'root' directory under which all your files are accessible. Your C: drive will be /c. To move around you can use cd to change directory (using / instead of ) and ls to list files instead of using dir.
Assuming your source code is in C:\projects\myproject:
cd /c/projects/myproject
git init
The second line above creates a git repository in that directory. This doesn't affect your code, it just creates a folder called .git that contains all of the book-keeping information.
You don't want to have every file under version control - in particular you don't want your build outputs. You need to set up a file in your project directory called .gitignore which tells git which files and directories should be ignored. As a starting point you can copy https://github.com/github/gitignore/blob/master/Java.gitignore and rename the file to .gitignore
Basic Commands and committing your initial version
There are a small number of basic commands:
git status
Running git status will tell you which files have been modified, which are not under version control, and which files have been added to the staging area to be committed next time.
git add path/to/file
This adds a file to the staging area waiting to be committed. You can add multiple files to the staging area before committing them in one go.
git commit -m "description of your change"
This commits all of the staged files as a new version, which the specified commit message.
If you go into your project directory, do a git status and check through the list to make sure there's nothing you don't want to have under version control, then you can do git add . to add everything to the staging area and git commit -m "Check in initial version of the source code" to commit it to the repository.
After you've committed, you can run
git log
To see a history of all of the changes. IntelliJ has a view that will show you the same thing.
Creating an experimental branch
This is where git shines; if you want to try something experimental you can create a branch of your project while allowing git to preserve the original version.
git checkout -b experiment1
Will create and switch to a branch called experiment1. You can delete, rename, move, rewrite and develop whatever you like on this branch. The changes you commit will be independent of your original working version.
You can switch back to your original version (preserving all of the changes you've committed on that branch) using:
git checkout master
Where master is just the name of the default branch created when you ran git init. The experimental version will still be there and can be switched to again using git checkout experiment1 or from IntelliJ using the branch selection in the bottom right corner of the status bar.
If you decide that the changes you've made in experiment1 are to become your new "good" version, you can merge them back into the master branch and repeat the cycle from there.
So I had been (stupidly) making changes directly on the live server instead of making them on my local machine and deploying them. This messed up my deployment. So now I was to do "git reset --hard".
On my remote server I have a project.git directory (for the repository... which is bare btw) and a project directory (for my actual application).
But when I try to run "git reset --hard" it tells me I'm not on a working tree. If I go into config and change bare to false... it says the same thing.
Ideas?
Found a better solution. :)
First I did a git reset --hard on the local server (since the remote server is just a bare repository.)
Then I did a git commit -a which told me there were no changes but that there were untracked files.
So I did a git add . to add all the files that weren't being tracked.
Finally I ran git commit -a again and git push.
This updated my repository with all the new files and then cap deploy functioned as expected.
Friends, I tried to deploy my yii production application from cloud9 IDE to OpenShift while do so, I got this error message,
CException
Application runtime path "/var/lib/openshift/51dd48794382ecfd530001e8/app-root/runtime/repo/php/protected/runtime" is not valid. Please make sure it is a directory writable by the Web server process.
Even when I changed folder permissions to 775 (chmod -R 775 directory) on Cloud9 IDE and deployed again, but I get the same error coming.
It's an old question, but I just bumped into the same issue very recently.
When you extracted the "yii" package several folders were empty, "framework/protected/runtime" was one of them.
To deploy to OpenShift you need to commit the yii package to git, and the push the commit to OS. But, git won't commit empty folders, so they are not created in your deployment. You need to create some file inside those folders and add those files to your git repo before committing/pushing. The usual procedure would be to add a ".gitkeep" file to those folders (it's just a empty dummy file, so git would see those folders).
That would fix this particular error.
It may be due the ownership given to the folder.
Check the web server user group, is that directory is writable or not and also What effects a web server when we change the platform.
Hope my suggestion would be useful.
For Yii applications, the assets and protected/runtime folders are special. First, both folders must exist and writable by the server (httpd) process. Second, these two folders contains temporary files, and should be ignored by git. If these temporary files got committed, deployment in plain servers (not Openshift servers) would cause git merge conflicts. So I put these two folders in .gitignore :
php/assets/
php/protected/runtime/
In my deployment, I add a shell script to be called by openshift, creating both folders under $OPENSHIFT_DATA_DIR and creating symbolic link to both of them in the application's folders. This is the content of the shell script (.openshift/action_hooks/deploy) which I adapted from here :
#!/bin/bash
if [ ! -d $OPENSHIFT_DATA_DIR/runtime ]; then
mkdir $OPENSHIFT_DATA_DIR/runtime
fi
# remove symlink if already exists, fix problem when with gears > 1 and nodes > 1
rm $OPENSHIFT_REPO_DIR/php/protected/runtime
ln -sf $OPENSHIFT_DATA_DIR/runtime $OPENSHIFT_REPO_DIR/php/protected/runtime
if [ ! -d $OPENSHIFT_DATA_DIR/assets ]; then
mkdir $OPENSHIFT_DATA_DIR/assets
fi
rm $OPENSHIFT_REPO_DIR/php/assets
ln -sf $OPENSHIFT_DATA_DIR/assets $OPENSHIFT_REPO_DIR/php/assets
The shell script ensures the temporary folders created on each gear after openshift deployment. By default, a new directory's right are u+rwx, and it became writable by the httpd process because the gear runs httpd as the gear user (not apache or something else).
when I'm trying to commit my code project it shows a file called 'UserInterfaceState.xcuserstate' which has to be committed. once i commit it and try to push my project to git ,the Xcode gives me a popup message saying that 'The working copy "app" has uncommitted changes.' and when I try to commit again I get the same file 'UserInterfaceState.xcuserstate' to be committed again. Can anybody help me with this?
UserInterfaceState.xcuserstate is where Xcode saves your GUI states, such as window positions, open tabs, expanded nodes in the project inspector etc.
Simply resizing the Xcode window will cause this file to change and be flagged as modified by your source control system. You can make your SCM system ignore specific files that are not important to the project itself.
You want Git to ignore the file, you can add it to the .gitignore file, but you have to remove the tracking. To stop tracking a file that is currently tracked, use git rm –cached.
git rm --cached ProjectFolder.xcodeproj/project.xcworkspace/xcuserdata/myUserName.xcuserdatad/UserInterfaceState.xcuserstate
git commit -m "Removed file that shouldn't be tracked"
Afterwards the .gitignore will take effect of UserInterfaceState.xcuserstate
Now onwards you wont get a popup message saying that 'The working copy "app" has uncommitted changes.'