13 KiB
FRACZ
Fracz is a good excercize to learn basic and advanced git command. Go to the website and follow the instructions to start the game. Make sure you have git installed onto the machine you want to do the excercize on.
1. Exercise: Push a commit you have made
Instructions
The first exercise is to push a commit that is created when you run the git start command. Just try git verify after you have initialized the exercises and be proud of passing the first one :-)
Commands
git push
2. Exercise: Commit one file
Instructions
There are two files created in the root project directory - A.txt and B.txt. The goal is to commit only one of them. NOTE: Remember that you can submit your solutions with git verify command instead of git push.
The easiest solution
TODO
Further info
You prepare changes to be committed with git add command. It adds files from working area to staging area. Only files that are in staging area will be included in the commit when you run the git commit command.
Remember that you can git add -A to add all changed files to staging area. You can also do this in air with -a option for git commit, e.g.
Commands
git commit -am "Some Commit Message"
3. Commit one file of two currently staged
Instructions
There are two files created in the root project directory - A.txt and B.txt. They are both added to the staging area. The goal is to commit only one of them.
The easiest solution
TODO
Further info
When you have added too many changes to staging area, you can undo them with git reset command before the git commit.
The easiest solution
TODO
Further info
When you have added too many changes to staging area, you can undo them with git reset command before the git commit.
Commands
$ git status
On branch commit-one-file-staged
Your branch is up to date with 'origin/commit-one-file-staged'.
Changes to be committed:
(use "git restore --staged <file>..." to unstage)
new file: A.txt
new file: B.txt
FOR209-03+Admin@FOR209-03 MINGW64 ~/Documents/exercises (commit-one-file-staged)
$ git restore --staged B.txt
4. Ignore unwanted files
Instructions
It is often good idea to tell Git which files it should track and which it should not. Developers almost always do not want to include generated files, compiled code or libraries into their project history.
Your task is to create and commit configuration that would ignore:
all files with exe extension all files with o extension all files with jar extension the whole libraries directory Sample files are generated for you.
The easiest solution
Further info
A .gitignore file specifies intentionally untracked files that Git should ignore.
To ignore all files with specific string inside filename, just type it in, i.e. dumb To ignore all files with specific extension use wildcard, i.e. .exe To ignore the whole directories, put a slash in the end of the rule, i.e. libraries/ To specify full path from the .gitignore location start rule with slash, i.e. /libraries
Note that there is a difference between libraries/ and /libraries/ rule. The first one would ignore all directories named libraries in the whole project whereas the second one would ignore only the libraries directory in the same location as .gitignore file.
Also, it's worth to know that there are many predefined .gitignores for specific environments so you don't have to invent your own. There is even a .gitignore generator.
commands
$ touch .gitigniore
$ vim .gitigniore
*.exe
*.o
*.jar
libraries/*
libraries\*
$ git add .gitignore
$ git commit -am "gitignore"
[ignore-them 85217dc] gitignore
1 file changed, 0 insertions(+), 0 deletions(-)
rename .gitigniore => .gitignore (100%)
Status: **PASSED**
You can see the easiest known solution and further info at:
https://gitexercises.fracz.com/e/ignore-them/9l2
Next task: chase-branch
In order to start, execute: git start next
See your progress and instructions at:
https://gitexercises.fracz.com/c/9l2
5. Chase branch that escaped
Instructions
You are currently on chase-branch branch. There is also escaped branch that has two more commits.
HEAD | chase-branch escaped | | A <----- B <----- C You want to make chase-branch to point to the same commit as the escaped branch.
escaped
|
A <----- B <----- C
|
chase-branch
|
HEAD
The easiest solution Further info Because the chase branch was direct ancestor of the escaped branch, the pointer could be simply moved and no merge commit is necessary (also, conflicts are impossible to happen in such situations).
This is what Git calls as Fast-Forward merge because the branch pointer is only fast forwarded to the commit you are merging with.
Note that you could easily fool this task by executing command
git push origin escaped:chase-branch
Remote repository could not tell then if you have done the merge or if you just wanted to set the remote chase-branch to point to the same commit as your local escaped branch (which is what the command above does).
See also: Basic Branching and Merging from Git Book.
commands
$ git merge escaped
6. Resolve a merge conflict
Instructions
Merge conflict appears when you change the same part of the same file differently in the two branches you're merging together. Conflicts require developer to solve them by hand.
Your repository looks like this:
HEAD
|
merge-conflict
|
A <----- B
\
\----- C
|
another-piece-of-work You want to merge the another-piece-of-work into your current branch. This will cause a merge conflict which you have to resolve. Your repository should look like this:
HEAD
|
merge-conflict
|
A <----- B <----- D
\ /
\----- C <----/
|
another-piece-of-work
How to resolve merge conflicts using the command line
The most direct way to resolve a merge conflict is to edit the conflicted file. Open the merge.txt file in your favorite editor. For our example lets simply remove all the conflict dividers. The modified merge.txt content should then look like:
this is some content to mess with
content to append
totally different content to merge later
Once the file has been edited use git add merge.txt to stage the new merged content. To finalize the merge create a new commit by executing:
git commit -m "merged and resolved the conflict in merge.txt"
Git will see that the conflict has been resolved and creates a new merge commit to finalize the merge.
Git commands that can help resolve merge conflicts
General tools
git status
The status command is in frequent use when a working with Git and during a merge it will help identify conflicted files.
git log --merge
Passing the --merge argument to the git log command will produce a log with a list of commits that conflict between the merging branches.
git diff
diff helps find differences between states of a repository/files. This is useful in predicting and preventing merge conflicts.
Tools for when git fails to start a merge
git checkout
checkout can be used for undoing changes to files, or for changing branches
git reset --mixed
reset can be used to undo changes to the working directory and staging area.
Tools for when git conflicts arise during a merge
git merge --abort
Executing git merge with the --abort option will exit from the merge process and return the branch to the state before the merge began.
git reset
Git reset can be used during a merge conflict to reset conflicted files to a know good state
The easiest solution
Further info
Because the branches have diverged, fast-forward merge strategy could not be applied. Therefore, a merge conflict was possible. Because two branches made changes in the same file and near the same line, Git decided not to handle the situation itself but to throw a merge conflict (letting user decide what to do).
After you resolve the conflict, you need to add it to staging area to tell Git that you have handled the situation. git commit then continues the merging process.
However, when Git stops and tells you that there is a conflict to resolve, you are not left on your own. There are some tricks that can make conflict resolution process a lot easier.
- By default, Git shows only your changes and their changes of conflicting lines. This will look like this:
<<<<<<< HEAD
2 + ? = 5
=======
? + 3 = 5
>>>>>>> another-piece-of-work
It is often very helpful to see also how the code looked like before both of these changes. Seeing more context can help figure out good conflict resolution a lot faster. You can checkout each file in diff3 mode that shows all three states of conflicting lines.
git checkout --conflict=diff3 equation.txt
Conflict in equation.txt will be presented now as:
<<<<<< HEAD
2 + ? = 5
||||||| merged common ancestors
? + ? = 5
=======
? + 3 = 5
>>>>>>> another-piece-of-work
If you like the diff3 presentation of conflicts, you can enable them by default with
git config merge.conflictstyle diff3
- Sometimes you want either discard your changes or their changes that introduces the conflict. You can do that easily with
git checkout --ours equation.txt
It's also worth to read the Basic Branching and Merging from the Git Book.
commands
137 vim equation.txt
138 git merge another-piece-of-work
139 cat equation.txt
140 vim equation.txt
141 git merge another-piece-of-work
142 cat equation.txt
143 git add equation.txt
144 git commit -m "merge equation"
145 git verify
146 git start next
7. Saving your work
Instructions
You are working hard on a regular issue while your boss comes in and wants you to fix a bug. State of your current working area is a total mess so you don't feel comfortable with making a commit now. However, you need to fix the found bug ASAP.
Git lets you to save your work on a side and continue it later. Find appropriate Git tool and use it to handle the situation appropriately.
Look for a bug to remove in bug.txt.
After you commit the bugfix, get back to your work. Finish it by adding a new line to bug.txt with
Finally, finished it! Then, commit your work after bugfix.
Hint
- Use git stash to save your current work in background and clean the working area.
- Fix the bug.
- Use git stash pop to reapply your changes to working area. Finish your work (see instructions)
$ git pull
...
file foobar not up to date, cannot merge.
$ git stash
$ git pull
$ git stash pop
The easiest solution
Further info
It's hard to verify if you have done this task correctly.
Its aim was to demonstrate git stash feature. When you run this command on dirty working area, it will save its state in stashed changes. You can do another work then, make any commits, checkout to any branch and then get the stashed changes back. You can think of stash as an intelligent Git clipboard.
An interesting option of stash command is the --keep-index which allows to stash all changes that were not added to staging area yet.
Keep in mind that applying stash might lead to conflicts if your working area introducted changes conflicting with stash. Therefore, its often safer to run git stash apply instead of git stash pop (the first one does not remove stashed changes).
Last thing to remember is that stashes are only local - you can't push them to remote repository.
See Stashing in the Git Book.
Commands
149 **git stash**
150 **git pull**
151 ls
152 git status
153 **vim bug.txt**
154 ls
155 git status
156 **git add bug.txt**
157 **git commit -m "Hotfix"**
158 **git stash pop**
159 **vim bug.txt**
160 **git add bug.txt**
161 git commit -m "after bug fix"
162 git verify
163 git log
164 git reset --soft b62a9b25b4f7da5e94c1fc754fd2803d7bfe8107
165 git status
166 vim bug.txt
167 **git add program.txt**
168 **git commit -m "after bug fix"**
8. Change branch history
Instructions
You were working on a regular issue while your boss came in and told you to fix recent bug in an application. Because your work on the issue hasn't been done yet, you decided to go back where you started and do a bug fix there.
Your repository look like this:
HEAD
|
change-branch-history
|
A <----- B
\
\----- C
|
hot-bugfix
Now you realized that the bug is really annoying and you don't want to continue your work without the fix you have made. You wish your repository looked like you started after fixing a bug.
HEAD
|
change-branch-history
|
A <----- C <----- B
|
hot-bugfix
Achieve that.
Hint
You need to use git rebase command.