What is the cleverest use of source repository that you have ever seen?

Question

This actually stems from on my earlier question where one of the answers made me wonder how people are using the scm/repository in different ways for development.

Answer 1

Definitely Polarion Track & Wiki...

The entire bug tracking and wiki database is stored in subversion to be able to keep a complete revision history.

http://www.polarion.com/products/trackwiki/features.php

Answer 2

My favorite? An unreleased tool which used Bazaar (a DSCM with very well-thought-out explicit rename handling) to track tree-structured data by representing the datastore as a directory structure.

This allowed an XML document to be branched and merged, with all the goodness (conflict detection and resolution, review workflow, and of course change logging and the like) made easy by modern distributed source control. Splitting components of the document and its metadata into their own files prevented the issues of allowing proximity to create false conflicts, and otherwise allowed all the work that the Bazaar team put into versioning filesystem trees to work with tree-structured data of other kinds.

Answer 3

Pre-tested commits

Before (TeamCity, build manager):

The concept is simple, the build system stands as a roadblock between your commit entering trunk and only after the build system determines that your commit doesn't break things does it allow the commit to be introduced into version control, where other developers will sync and integrate that change into their local working copies

After (using a DVCS like Git, that is a source repository):

My workflow with Hudson for pre-tested commits involves three separate Git repositories:

• my local repo (local),
• the canonical/central repo (origin)
• and my "world-readable" (inside the firewall) repo (public).

For pre-tested commits, I utilize a constantly changing branch called "pu" (potential updates) on the world-readable repo.
Inside of Hudson I created a job that polls the world-readable repo (public) for changes in the "pu" branch and will kick off builds when updates are pushed.

my workflow for taking a change from inception to origin is:

* hack, hack, hack
* commit to local/topic
* git pup public
* Hudson polls public/pu
* Hudson runs potential-updates job
* Tests fail?
      o Yes: Rework commit, try again
      o No: Continue
* Rebase onto local/master
* Push to origin/master

Using this pre-tested commit workflow I can offload the majority of my testing requirements to the build system's cluster of machines instead of running them locally, meaning I can spend the majority of my time writing code instead of waiting for tests to complete on my own machine in between coding iterations.

---

(Variation) Private Build (David Gageot, Algodeal)

Same principle than above, but the build is done on the same workstation than the one used to develop, but on a cloned repo:

How not to use a CI server in the long term and not suffer the increasing time lost staring at the builds locally?

With git, it’s a piece of cake.
First, we ‘git clone’ the working directory to another folder. Git does the copy very quickly.
Next times, we don’t need to clone. Just tell git get the deltas. Net result: instant cloning. Impressive.

What about the consistency?
Doing a simple ‘git pull’ from the working directory will realize, using delta’s digests, that the changes where already pushed on the shared repository.
Nothing to do. Impressive again.

Of course, while the build is running in the second directory, we can keep on working on the code. No need to wait.

We now have a private build with no maintenance, no additional installation, not dependant on the IDE, ran with a single command line. No more broken build in the shared repository. We can recycle our CI server.

Yes. You’ve heard well. We’ve just built a serverless CI. Every additional feature of a real CI server is noise to me.

#!/bin/bash
if [ 0 -eq `git remote -v | grep -c push` ]; then
  REMOTE_REPO=`git remote -v | sed 's/origin//'`
else
  REMOTE_REPO=`git remote -v | grep "(push)" | sed 's/origin//' | sed 's/(push)//'`
fi

if [ ! -z "$1" ]; then
  git add .
  git commit -a -m "$1"
fi

git pull

if [ ! -d ".privatebuild" ]; then
  git clone . .privatebuild
fi

cd .privatebuild
git clean -df
git pull

if [ -e "pom.xml" ]; then
  mvn clean install

  if [ $? -eq 0 ]; then
    echo "Publishing to: $REMOTE_REPO"
    git push $REMOTE_REPO master
  else
    echo "Unable to build"
    exit $?
  fi
fi

---

Dmitry Tashkinov, who has an interesting question on DVCS and CI, asks:

I don't understand how "We’ve just built a serverless CI" cohere with Martin Fowler's state:
"Once I have made my own build of a properly synchronized working copy I can then finally commit my changes into the mainline, which then updates the repository. However my commit doesn't finish my work. At this point we build again, but this time on an integration machine based on the mainline code. Only when this build succeeds can we say that my changes are done. There is always a chance that I missed something on my machine and the repository wasn't properly updated."
Do you ignore or bend it?

@Dmitry: I do not ignore nor bend the process described by Martin Fowler in his ContinuousIntegration entry.
But you have to realize that DVCS adds publication as an orthogonal dimension to branching.
The serverless CI described by David is just an implementation of the general CI process detailed by Martin: instead of having a CI server, you push to a local copy where a local CI runs, then you push "valid" code to a central repo.

@VonC, but the idea was to run CI NOT locally particularly not to miss something in transition between machines.
When you use the so called local CI, then it may pass all the tests just because it is local, but break down later on another machine.
So is it integeration? I'm not criticizing here at all, the question is difficult to me and I'm trying to understand.

@Dmitry: "So is it integeration"?
It is one level of integration, which can help get rid of all the basic checks (like format issue, code style, basic static analysis detection, ...)
Since you have that publication mechanism, you can chain that kind of CI to another CI server if you want. That server, in turn, can automatically push (if this is still fast-forward) to the "central" repo.

David Gageot didn't need that extra level, being already at target in term of deployment architecture (PC->PC) and needed only that basic kind of CI level.
That doesn't prevent him to setup more complete system integration server for more complete testing.