[Tika Wiki] Update of "UsingGit" by ChrisMattmann

[Apache Wiki <wi...@apache.org>]

Dear Wiki user,

You have subscribed to a wiki page or wiki category on "Tika Wiki" for change notification.

The "UsingGit" page has been changed by ChrisMattmann:
https://wiki.apache.org/tika/UsingGit

Comment:
- add Git instructions page.

New page:
{{{#!wiki caution
'''Pending a VOTE'''

This page will become valid if the VOTE for Apache Tika to move to using Git passes.
}}}

Apache Tika uses the [[http://git-scm.com/|Git]] version control system. Apache provides writeable Git repositories hosted at [[https://git-wip-us.apache.org|https://git-wip-us.apache.org/]]. This guide assumes you have read the guides and information provided on the Apache Git-WP page.

= Checking out a copy of Tika and modifying it =

To check out a copy of Tika, perform the following command:

{{{
git clone https://git-wip-us.apache.org/repos/asf/tika.git trunk
}}}

This will check out the trunk (master) copy of the code. The normal convention,  `master`, used in Git, is replaced with `trunk` in the ASF's repos.

Once you have the code, you can modify a file, or two, then add the files for staging/commit, and then commit them. Once done, you can also decide to push the files up to the master repository if you have write access and are a member of the PMC and/or a committer. If you don't have write access to the repository, you can follow [[http://github.com/apache/tika/#contributing|this guide]] for issuing pull requests that committers/PMC members can merge.

Here are the steps to achieve the above if you have commit access, assuming you have 2 files, `file1` and `file2` in the repository local copy that you modified.

 1. `git add file1 file2`
 2. `git commit -m "message describing change."`
 3. `git push -u origin trunk`

= Suggested Developer Workflow =

As a longtime SVN user, it took me a while to figure out Git and realize its great support for use cases. Here is the one I most commonly use in Git that I think fits the Tika workflow as a PMC member/committer quite well. Imagine that you are working on an issue, call it TIKA-xxx. Either you have filed it yourself, or someone has filed it. My suggestion is to create a feature/working branch for that issue, named exactly after the JIRA issue, e.g., create a branch TIKA-xxx. Commit your changes and modifications there. Then, after the changes are applied, tested, and verified, then you can switch back to the master branch, and push your local copy to the remote. If you'd like a review before pushing (e.g., attaching a patch you have two options). You can create a patch file by diffing the branches, and then attach that to the JIRA issue. This is common and already in use using Subversion in Tika in years past. Let's take this use case to start. Inside of your tika clone, do the following:

 1. `git branch TIKA-xxx`
 2. modify files, change them, test, etc.
 3. `git add <changed files>` (to see them, try `git status` or simply to stage '''all''' changes, `git add *`)
 4. `git commit -m "Fixes for TIKA-xxx contributed by <your first name> <your last name> <your email>"
 5. `git checkout trunk`
 6. `git diff trunk..TIKA-xxx > TIKA-xxx.<your last name>.<yyMMdd>.patch.txt`
 7. Attach the patch created in 6 to JIRA.

If you are '''not''' looking for a review or you have already had a review and are ready to commit the changes, push them:

 0. (if you haven't already merge, the branch into master) `git checkout trunk && git merge TIKA-xxx`
 1. `git push -u origin trunk`

= Suggested User Contribution Workflow =

Users contribute patches to Tika using JIRA and/or Github. Let's take the use case for contributing via JIRA, then we'll tackle the Github one.

== JIRA contribution ==

Grab the user's patch file, and then apply it locally. Before doing so, create the feature branch by following step 1 from the Developer workflow section. Then, assuming the patch file is named according to the conventions from step 6 in the Developer workflow section, perform the following steps (make sure you are on the TIKA-xxx feature branch):

 1. `git apply < TIKA-xxx.<your last name>.<yyMMdd>.patch.txt`
 2. Steps 3-4 from Developer workflow guide.
 3. Final 2 steps from Developer workflow guide (aka the merge of feature branch step and then the push to origin trunk)

== Github contribution ==

If they are contributing using Github they are submitting a Pull Request for review you can easily merge their pull request into your local feature branch using Git. Assuming the Github user is "user01" and the branch they have created is "fix-tika-stuff" (you can find this information in the Tika Pull request, for example in [[https://github.com/apache/tika/pull/65|this pull request #65]], the username is `smahda` and branch name is `TIKA-1803`), you can merge it into your local feature branch like so (again, make sure you are on the local feature branch for your issue, TIKA-xxx first by typing `git checkout TIKA-xxx` if you aren't already):

 1. `git pull https://github.com/user01/tika.git fix-tika-stuff` (will find the remote github branch and merge locally)
 2. test and make sure PR works, etc. Fix conflicts, etc.
 3. Final 2 steps from Developer workflow guide (aka the merge of feature branch step and then the push to origin trunk)