What are the benefits of integrating source code for a bug tracking system?

Question

I choose a bug / error tracking system for use in our project. This question and this question are useful for evaluating systems. However, I have a question when considering the various products on offer.

Some systems facilitate the integration of a source code management system as a function. This is not what I used before, and looking at the sites of various tools, I can not find details about what exactly this integration provides. Is this more than viewing my repository through the same web interface that I use to raise errors?

So what are the benefits of such integration? How will it save me or make our product better?

Answer 1

Jira along with FishEye (SCM browser) if, if you commit a set of changes with log messages containing a Jira release key, for example "PROJ-123", insert the link in the corresponding Jira problem, and Jira will display the link in PROJ-123 in the change list which mentions this problem.

Jira can also integrate with Hudson , so when an assembly that includes a fix (i.e. a set of changes) is executed, the Jira issue mentioned in the log message of this commit will receive a comment about the status of the assembly.

Since the error tracker is used to track problems with the source code, there is a boat of functions that can be represented in the integrated problem / source code management system.

Answer 2

The main advantage is that you can create errors in relation to specific versions of the source code. It just helps determine the state of the code base when an error is detected. I think the popular product in this space is Trac , which integrates with SVN.

Answer 3

I did this using Visual Studio Team Foundation Server, which not only integrates with the source control, but also integrates with the data warehouse. This allows, for example, to track which errors were caused by which code fragments show which code fragments need better testing. The upcoming features in the 2010 version are even more convincing.

Answer 4

Well, I would say that there is one big reason, and you can work with a task / problem oriented one .

I mean:

• every code change you make (everything from a small quick fix to a big new feature) will be a problem in Jira or Rally or Bugzilla, Trac, Mantis, the one you like.

• This means that the problem / task tracking system should be easy to use, fast, simple, etc., otherwise developers will hate it.

• Then map each change to the problem, and you're done. You get full traceability (great for debugging diff, you will skip it if you don’t have one), better project tracking, better release management, etc.

You can link each change / commit command (depending on your scm jargon) for the task or, if your scm can do this, create a branch for each fix, which is even better.

Basically, what you get is simple: every change will be documented, or at least related to the right problem / task.

If you use branches, you can promote things like: always working on stable baselines, reducing damage to the main line (or making it clean if you want), deciding which tasks are integrated and which ones until the next version, individually run tests about changes before the merger, etc.

Answer 5

Github is basically the opposite: there aren’t so many bug / problem tracking systems with source code version control turned on, but rather source code version control with some bug / error correction added. I can’t believe that no one mentioned it here, because nowadays it is more common in the opposite direction.

But it is very powerful, you can basically refer to all git sentences also in questions: commits, branch, pull-requests. Github even installed several “automatic” ones, combining PR with content such as “fixes issue 23” automatically closes issue 23.

Github is also very convenient, since most of the open source software currently in use is also hosted there, and you can also reference all of these libraries in your own problems / pull requests, etc. Also, most of the typical modern commercial infrastructure around software development will also be integrated with Github: Travis or Codeship will tell you how your assemblies do and even automatically deploy them, Hound or Rubocop will tell you what your code looks like and Usersnap or Trackduck will report bugs about your Github problems if you don't want to use their software.

Trac, which was mentioned in the answers here, feels very old when compared to Github.