Explain web-based development principles through bitbucket



  • Hello! I took Bitbucket's aza a day ago, put SourceTree. The point is, do I understand the algorithm of development through Bitbucket?

    1. Several people write code(php,js,mysql,html). They' it. Push. Code To serve Bitbucket;
    2. Next, the whole Pull-They synchronize the code on their cars.
    3. The code is on FTP. How do you do that on FTP? Or I'm not. Do you understand the match?

    I think I should start the code on the clients. There's every member of the commandy who's running the code through what he wants. Apache+MySQL+PHP.

    Explain the plunge, the sedentary masters!



  • Several people write code(php,js,mysql,html). Commit it. Push. The code is sent to the Bitbucket server;

    Almost there. In general, there are several additional rules in the command-and-work process.

    1. A separate developmental branch shall be under each individual task.
    2. The developer should never put a committ into the master's branch. He puts them in his development branch.
    3. When the task has been brought to a certain end, a merger request (merge request, pull request) is made. Either in some other form (whether orally) the branch is ready for merging. Details: https://ru.stackoverflow.com/q/505731/181472
    4. The proposed branch passes code log, testing and any other checks.
    5. In each case, there must be one responsible person who decides on the merger. It could be a timlid, a tester, and someone else would agree.

    Next, the whole Pull-They synchronize the code on their cars.

    Almost like that.

    Team git pull Actually includes two: git fetch + git merge♪ There are no problems with the first one, the second may be in conflict. I usually update the master and then rebase the current branch on him.

    git fetch
    # посмотрим, что к нам пришло
    git log --oneline --graph --decorate --all
    

    обновим master

    git checkout master
    git merge origin/master # или просто git pull

    переставим текущую ветку на новый master.

    при наличии конфликтов лучше разрешить их сейчас, чем потом при слиянии в master

    git checkout myfeature
    git rebase master

    The code is on FTP. How do you do that on FTP? Or do I not understand the match?

    Don't confuse the code storage and the application. The code inside the guita is stored in his own database. There's no file on a server like Bitbucket that's in your work folder, because it's bare repositories and there's only your own vault.

    How will you deploy the annex is a completely separate question. If it's an interpretable code like php or js, it's convenient to use rsync. There is a more popular, but much less reliable way of deploying through production-server and git-hook repositories. If only FTP is available, we'll have to go through FTP. Details: https://ru.stackoverflow.com/q/428483/181472

    I think I should start the code on the clients. There's every member of the commandy who's running the code through what he wants. Apache+MySQL+PHP.

    See that you have an identical environment at all stages of the process: development, testing, staging (if any), production.

    • Single DS
    • Single versions of all relationships
    • Single mode of deployment, including configuration
    • Single method of launching the annex

    All of the above-mentioned should be described in at least some document and ideally in the code.

    So you fear your team:

    • (parts) Bags which are scattered on the market;
    • from the "works on my machine" phenomenon.
    • From the art servers who were set up a year ago by a man who then resigned, and now no one can do the same, so there's nothing to replace them if they're broken.




Suggested Topics

  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2