home of the madduck/ blog/ feeds/
madduck's Planet Git posts

The following blog posts appear on Planet Git. Please visit my main blog page for all my posts.

All recent articles on packaging using a version control system should really appear over at Planet vcs-pkg. Feel free to just ping me with a feed URL that is vcs-pkg-specific.

Posted Sun 19 May 2013 13:48:01 CEST Tags:

The other day, Romain shared his concerns about using Git for team-maintained packaging. His comment system is currently broken, so I wrote an e-mail reply, which I would like to share.

I agree with Romain that the design decision to not support subtree checkouts like SVN is not without problems. As opposed to a single SVN repo with components in subdirectories that you can individually check out, you might end up with a hundred Git repos, and the same change to all then requires one to iterate all 100.

I'd like to make the distinction between trivial changes (e.g. s/© 2008/&-2009/g) and those that might not be (e.g. Standards-Version, or something even more elaborate).

In case of the former, there's no question, it can be painful to operate across a hundred repos. Tools like mr make that a bit easier, but it's still far from optimal.

The latter, however — updating Standards-Version and adding the appropriate changelog entry — is not really comparable. Neither would be e.g. changing a file location in 100 different repos. In those cases, every single package needs manual intervention, and if only for quality reasons and testing. In this sense, I actually think that a single SVN checkout with all the subtrees and the possibility to easily commit the result of a recursive action is counter-productive.

On the other hand, I don't say that I am pleased with the workflow Git (or any other DVCS for that matter) imposes. It's sometimes quite painful, as Romain says. We are missing higher-level tools that allow for easier and more intuitive bulk operations. I think that they should be implemented outside of the VCS-tool though, true to the Unix principles. SVN integrates it all into a monolithic piece of software, and that often isn't ideal either (think size and slowness, or backup weight, or chance of corruption, or granular access control, or the impossibility to properly track files across subtrees).

mr is a step in the right direction, and we need more tools along those lines. First, however, I think that people need to figure out how exactly to use DVCS for packaging, such that there is any chance of consolidating workflows across a larger number of packages; if everyone does it their own, slightly unique way, then that goal is inifinitely far. This is the reason I started vcs-pkg.org, and even though we're still far from anywhere, I am quite pleased with what we've done so far.

If you're at Debcamp or DebConf, maybe you could join the discussion.

Romain also mentioned that distributed VCS don't allow for the same sort of centralisation as SVN does. I disagree: you can use Git in exactly that way, as a centralised repo from which packages are built. The nice advantage over SVN (one which svk tried to close) is the ability for everyone to easily branch/fork, or work offline.

Once you start down that path, it somehow inherently becomes everyone's own responsibility to ensure that one's changes end up in the central repository (where commit hooks might verify the build-ability, ensure that the test suite still passes, or run simple format/consistency checks).

This sort of workflow is very different from the one with a self-appointed benevolent dictator at the top, who (like Linus, or Junio for Git) sometimes forget to include patches due to overflooding. The question is really: Given that you need some sort of centralised release coordination, do you want a human or a repo to be the central entity (and single point of failure)?

I really prefer the repo, since that places the sole responsibility on the leafs, on the contributors, who need to see their code through all the way.

It's a whole lot more rewarding to commit/push, get a rejection, pull, merge, commit/push, and be done, rather than to send a patch to upstream, wait, reping, notice that it's not in in the new release, ask, ping, change, reping, get angry, ping, hope, wait, ping, wonder why the heck you are still doing this, write angry email but don't send it, reping, ask, and finally notice that it's been accepted after all.

NP: Deep Purple: Made in Japan

Posted Fri 19 Jun 2009 08:28:01 CEST Tags:

Instead of commenting on all the recent Git on Planet Debian, I'd like to point you all to the Git Wiki, and specifically the page BlogPosts.

Please link your Git-related blog posts from there.

Also, there is Planet RCS for you to aggregate RCS/VCS-relatd posts.

NP: Anekdoten: From Within

Posted Sat 13 Dec 2008 14:01:06 CET Tags:

I was excited by Pierre's idea to add Git branch information to the Zsh prompt and even more so when I saw Mike implement support for multiple VCSs.

Unfortunately, Mike's a Bash user, and so I took it upon myself to port the idea to Zsh. The file 60_vcsprompt is sourced from my .zshrc, which sets psvar[1] through psvar[3]. Those are then used in 80_prompt (also sourced from .zshrc) when setting $PS1.

My prompt follows the same principle as Mike's and puts the branch name at the repository root location in the repository path. In the following example, ~, ~/code, and ~/code/netconf/netconf are three separate Git repositories, while ~/code/unionfs-fuse and ~/code/unperish are maintained with Mercurial and Bazaar respectively:

lapse:~|master|% cd code
lapse:~/code|master|% cd netconf 
lapse:~/code|master|netconf% cd netconf
lapse:..e/netconf/netconf|master|% cd src
lapse:..etconf/netconf|master|src% git checkout no-threads
Switched to branch "no-threads"
lapse:..nf/netconf|no-threads|src% cd ../../../unionfs-fuse 
lapse:../unionfs-fuse|hg:default|% cd ../unperish
lapse:..unperish|bzr:unperish@159|%

You'll notice that unlike Mike's prompt, mine's limited to a maximum length of 25 characters. However, the repository root path is kept at least 10 characters long, so the prompt might get longer than 25 characters if you descend deep into a repository's subdirectories.

I couldn't easily figure out how to add support for other version control systems, so if you do, please feed back the patches! And the same goes for suggestions and improvements.

One of the next things I am planning to implement is an indicator for when your working tree contains uncommitted changes, e.g.:

lapse:..etconf/netconf|master|src% touch foo
lapse:..tconf/netconf|master*|src%

So watch those files.

NP: Gazpacho: Bravo

Posted Fri 11 Jul 2008 11:21:07 CEST Tags:

Robin Rosenberg introduced me to the simplest method of creating a new git branch without any ancestors:

$ echo ref: refs/heads/newbranch > .git/HEAD
$ git branch
  master
[...]
$ git commit -m 'creating newbranch'
$ git branch
  master
* newbranch

This comes in handy if you want to maintain two separate components without any common files in the same repository. I am using it currently while experimenting with a new method of storing my home directory in git, which is still far from anywhere. I shall report once I reach a point of success or failure.

NP: Rush: Moving Pictures

Update: Johannes Schindelin taught me how to do the same without touching files in .git/:

$ git symbolic-ref HEAD refs/heads/newbranch
[...]

and also addressed the issue which would have all files already committed to the "master" branch now appear in the git status output as staged.

This is because the index contains the full copy of a revision of a file, as it would be if committed at any point. git status shows the differences between what has been committed, what would be committed, and what is available in the working tree. Since we pointed HEAD to nowhere ("newbranch" does not yet exist), the index and what has been committed (nothing in this case) diverge, the files are still staged, and thus are scheduled to be part of the impending commit.

The way to fix this is to remove the index:

$ rm .git/index

This may seem weird, but it works, because git recreates the index whenever you switch branches:

piper:~> git init-db
Initialized empty Git repository in .git/
piper:~> echo 1 > a; git add a; git commit -m.
Created initial commit e774324: .
1 files changed, 1 insertions(+), 0 deletions(-)
create mode 100644 a
piper:~> git symbolic-ref HEAD refs/heads/newbranch
piper:~> rm .git/index
piper:~> git status
# On branch newbranch
#
# Initial commit
#
# Untracked files:
#   (use "git add <file>..." to include in what will be committed)
#
#       a
nothing added to commit but untracked files present (use "git add" to track)
piper:~> echo 2 > b; git add b; git commit -m.
Created initial commit 54ff342: .
1 files changed, 1 insertions(+), 0 deletions(-)
create mode 100644 b
piper:~> git branch
  master
* newbranch
piper:~> git checkout master
fatal: Untracked working tree file 'a' would be overwritten by merge.
piper:~> git checkout -f master

Switched to branch "master"
piper:~> git status
# On branch master
nothing to commit (working directory clean)
piper:~> ls
a
piper:~> git checkout newbranch
Switched to branch "newbranch"
piper:~> git status
# On branch newbranch
nothing to commit (working directory clean)
piper:~> ls
b

As you can see, the creation of the branch is a bit complex, but once you (forcefully) switched back to master, you can then freely switch between and commit to them.

Posted Fri 11 Jul 2008 11:21:05 CEST Tags:

Introduction

I gave a joint presentation with Manoj at Debconf7 about using distributed version control for Debian packaging, and I volunteered to do an on-line workshop about using Git for the task, so it's about time that I should know how to use Git for Debian packaging, but it turns out that I don't. Or well, didn't.

After I made a pretty good mess out of the mdadm packaging repository (which is not a big problem as it's just ugly history up to the point when I start to get it right), I decided to get down with the topic and figure it out once and for all. I am writing this post as I put the pieces together. It's been cooking for a week, simply so I could gather enough feedback. I am aware that Git is not exactly a showcase of usability, so I took some extra care to not add to the confusion.

It may be the first post in a series, because this time, I am just covering the case of mdadm, for which upstream also uses Git and where I am the only maintainer, and I shall pretend that I am importing mdadm to version control for the first time, so there won't be any history juggling. Future posts could well include tracking Subversion repositories with git-svn, and importing packages previously tracked therewith.

I realise that git-buildpackage exists, but imposes a rather strict branch layout and tagging scheme, which I don't want to adhere to. And gitpkg (Romain blogged about it recently), deserves another look since, according to its author, it does not impose anything on its user. But in any case, before using such tools (and possibly extending them to allow for other layouts), I'd really rather have done it by hand a couple of times to get the hang of it and find out where the culprits lie.

Now, enough of the talking, just one last thing: I expect this blog post to change quite a bit as I get feedback. Changes shall be highlighted in bold typeface.

Setting up the infrastructure

First, we prepare a shared repository on git.debian.org for later use (using collab-maint for illustration purposes), download the Debian source package we want to import (version 2.6.3+200709292116+4450e59-3 at time of writing, but I pretend it's -2 because we shall create -3 further down…), set up a local repository, and link it to the remote repository. Note that there are other ways to set up the infrastructure, but this happens to be the one I prefer, even though it's slightly more complicated:

$ ssh alioth
$ cd /git/collab-maint
$ ./setup-repository pkg-mdadm mdadm Debian packaging
$ exit
$ apt-get source --download-only mdadm
$ mkdir mdadm && cd mdadm
$ git init
$ git remote add origin ssh://git.debian.org/git/collab-maint/pkg-mdadm
$ git config branch.master.remote origin
$ git config branch.master.merge refs/heads/master

Now we can use git-pull and git-push, except the remote repository is empty and we can't pull from there yet. We'll save that for later.

Instead, we tell the repository about upstream's Git repository. I am giving you the git.debian.org URL though, simply because I don't want upstream repository (which lives on an ADSL line) hammered in response to this blog post:

$ git remote add upstream-repo git://git.debian.org/git/pkg-mdadm/mdadm

Since we're using the upstream branch of the pkg-mdadm repository as source (and don't want all the other mess I created in that repository), we'll first limit the set of branches to be fetched (I could have used the -t option in the above git-remote command, but I prefer to make it explicit that we're doing things slightly differently to protect upstream's ADSL line).

$ git config remote.upstream-repo.fetch \
    +refs/heads/upstream:refs/remotes/upstream-repo/upstream

And now we can pull down upstream's history and create a local branch off it. The "no common commits" warning can be safely ignored since we don't have any commits at all at that point (so there can't be any in common between the local and remote repository), but we know what we're doing, even to the point that we can forcefully give birth to a branch, which is because we do not have a HEAD commit yet (our repository is still empty):

$ git fetch upstream-repo
warning: no common commits
[…]
  # in the real world, we'd be branching off upstream-repo/master
$ git checkout -b upstream upstream-repo/upstream
warning: You appear to be on a branch yet to be born.
warning: Forcing checkout of upstream-repo/upstream.
Branch upstream set up to track remote branch
  refs/remotes/upstream-repo/upstream.
$ git branch
* upstream
$ ls | wc -l
77

Importing the Debian package

Now it's time to import Debian's diff.gz — remember how I pretend to use version control for package maintenance for the first time. Oh, and sorry about the messy file names, but I decided it's best to stick with real data in case you are playing along:

Since we're applying the diff against version 2.6.3+200709292116+4450e59, we ought to make sure to have the repository at the same state. Upstream never "released" that version, but I encoded the commit ID of the tip when I snapshotted it: 4450e59, so we branch off there. Since we are actually tracking the git.debian.org pkg-mdadm repository instead of upstream, you can use the tag I made. Otherwise you could consider tagging yourself:

$ #git tag -s mdadm-2.6.3+200709292116+4450e59 4450e59
$ git checkout -b master mdadm-2.6.3+200709292116+4450e59
$ zcat ../mdadm_2.6.3+200709292116+4450e59-2.diff.gz | git apply

The local tree is now "debianised", but Git does not know about the new and changed files, which you can verify with git-status. We will split the changes made by Debian's diff.gz across several branches.

The idea of feature branches

We could just create a debian branch, commit all changes made by the diff.gz there, and be done with it. However, we might want to keep certain aspects of Debianisation separate, and the way to do that is with feature branches (also known as "topic" branches). For the sake of this demonstration, let's create the following four branches in addition to the master branch, which holds the standard Debian files, such as debian/changelog, debian/control, and debian/rules:

  • upstream-patches will includes patches against the upstream code, which I submit for upstream inclusion.
  • deb/conffile-location makes /etc/mdadm/mdadm.conf the default over /etc/mdadm.conf and is Debian-specific (thus the deb/ prefix).
  • deb/initramfs includes the initramfs hook and script, which I want to treat separately but not submit upstream.
  • deb/docs similarly includes Debian-only documentation I add to the package as a service to Debian users.

If you're importing a Debian package using dpatch, you might want to convert every dpatch into a single branch, or at least collect logical units into separate branches. Up to you. For now, our simple example suffices. Keep in mind that it's easy to merge two branch and less trivial to split one into two.

Why? Well, good question. As you will see further down, the separation between master and deb/initramfs actually makes things more complicated when you are working on an issue spanning across both. However, feature branches also bring a whole lot of flexibility. For instance, with the above separation, I could easily create mdadm packages without initramfs integration (see #434934), a disk-space-conscious distribution like grml might prefer to leave out the extra documentation, and maybe another derivative doesn't like the fact that the configuration file is in a different place from upstream. With feature branches, all these issues could be easily addressed by leaving out unwanted branches from the merge into the integration/build branch (see further down).

Whether you use feature branches, and how many, or whether you'd like to only separate upstream and Debian stuff is entirely up to you. For the purpose of demonstration, I'll go the more complicated way.

Setting up feature branches

So let's commit the individual files to the branches. The output of the git-checkout command shows modified files that have not been committed yet (which I trim after the first example); Git keeps these across checkouts/branch changes. Note that the ./debian/ directory does not show up as Git does not know about it yet (git-status will tell you that it's untracked, or rather: contains untracked files since Git does not track directories at all):

$ git checkout -b upstream-patches mdadm-2.6.3+200709292116+4450e59
M Makefile
M ReadMe.c
M mdadm.8
M mdadm.conf.5
M mdassemble.8
M super1.c
$ git add super1.c     #444682
$ git commit -s

  # i now branch off master, but that's the same as 4450e59 actually
  # i just do it so i can make this point…
$ git checkout -b deb/conffile-location master
$ git add Makefile ReadMe.c mdadm.8 mdadm.conf.5 mdassemble.8
$ git commit -s

$ git checkout -b deb/initramfs master
$ git add debian/initramfs/*
$ git commit -s

$ git checkout -b deb/docs master
$ git add RAID5_versus_RAID10.txt md.txt rootraiddoc.97.html
$ git commit -s

  # and finally, the ./debian/ directory:
$ git checkout master
$ chmod +x debian/rules
$ git add debian
$ git commit -s

$ git branch
  deb/conffile-location
  deb/docs
* master
  upstream
  upstream-patches

At this time, we push our work so it won't get lost if, at this moment, aliens land on the house, or any other completely plausible event of apocalypse descends upon you. We'll push our work to git.debian.org (the origin, which is the default destination and thus needs not be specified) by using git-push --all, which conveniently pushes all local branches, thus including the upstream code; you may not want to push the upstream code, but I prefer it since it makes it easier to work with the repository, and since most of the objects are needed for the other branches anyway — after all, we branched off the upstream branch.

Specifying --tags instead of --all pushes tags instead of heads (branches); you couldn't have guessed that! See this thread if you (rightfully) think that one should be able to do this in a single command (which is not git push refs/heads/* refs/tags/*)…

$ git push --all
$ git push --tags

Done. Well, almost…

Building the package (theory)

Let's build the package. There seem to be two (sensible) ways we could do this, considering that we have to integrate (merge) the branches we just created, before we fire off the building scripts:

  1. by using a temporary (or "throw-away") branch off upstream, where we integrate all the branches we have just created, build the package, tag our master branch (it contains debian/changelog), and remove the temporary branch. When a new package needs to be built, we repeat the process.

  2. by using a long-living integration branch off upstream, into which we merge all our branches, tag the branch, and build the package off the tag. When a new package comes around, we re-merge our branches, tag, and build.

Both approaches have a certain appeal to me, but I settled for the second, for two reasons, the first of which leads to the second:

  1. When I upload a package to the Debian archive, I want to create a tag which captures the exact state of the tree from which the package was built, for posterity (I will return to this point later). Since the throw-away branches are not designed to persist and are not uploaded to the archive, tagging the merging commit makes no sense. Thus, the only way to properly identify a source tree across all involved branches would be to run git-tag $branch/$tagname $branch for each branch, which is purely semantic and will get messy sooner or later.

  2. As a result of the above: when Debian makes a new stable release, I would like to create a branch corresponding to the package in the stable archive at the time, for security and other proposed updates. I could rename my throw-away branch, if it still existed, or I could create a new branch and merge all other branches, using the (semantic) tags, but that seems rather unfavourable.

So instead, I use a long-living integration branch, notoriously tag the merge commits which produced the tree from which I built the package I uploaded, and when a certain version ends up in a stable Debian release, I create a maintenance branch off the one, single tag which corresponds to the very version of the package distributed as part of the Debian release.

So much for the theory. Let's build, already!

Building the package (practise)

So we need a long-living integration branch, and that's easier done than said:

$ git checkout -b build mdadm-2.6.3+200709292116+4450e59

Now we're ready to build, and the following procedure should really be automated. I thus write it like a script, called poor-mans-gitbuild, which takes as optional argument the name of the (upstream) tag to use, defaulting to upstream (the tip):

#!/bin/sh
set -eu
git checkout master
debver=$(dpkg-parsechangelog | sed -ne 's,Version: ,,p')
git checkout build
git merge ${1:-upstream}
git merge upstream-patches
git merge master
for b in $(git for-each-ref --format='%(refname)' refs/heads/deb/*); do
  git merge -- $b
done
git tag -s debian/$debver
debuild -i.git
git checkout master

Kumar Appaiah spotted that -i.git is actually needed in the debuild call to make it exclude the .git directory from the generated diff.gz.

Note how we are merging each branch in turn, instead of using the octopus merge strategy (which would create a commit with more than two parents) for reasons outlined in this post. An octopus-merge would actually work in our situation, but it will not always work, so better safe than sorry (although you could still achieve the same result).

If you discover during the build that you forgot something, or the build script failed to run, just remove the tag, undo the merges, checkout the branch to which you need to commit to fix the issue, and then repeat the above build process:

$ git tag -d debian/$debver
$ git checkout build
$ git reset --hard upstream
$ git checkout master
$ editor debian/rules    # or whatever
$ git add debian/rules
$ git commit -s

$ poor-mans-gitbuild

Before you upload, it's a good idea to invoke gitk --all and verify that all goes according to plan:

screenshot of gitk after the above steps

When you're done and the package has been uploaded, push your work to git.debian.org, as before. Instead of using --all and --tags, I now specify exactly which refs to push. This is probably a good habit to get into to prevent publishing unwanted refs:

$ git push origin build tag debian/2.6.3+200709292116+4450e59-3

Now take your dog for a walk, or play outside, or do something else not involving a computer or entertainment device.

Uploading a new Debian version

If you are as lucky as I am, the package you uploaded still has a bug in the upstream code and someone else fixes it before upstream releases a new version, then you might be in the position to release a new Debian version. Or maybe you just need to make some Debian-specific changes against the same upstream version. I'll let the commands speak for themselves:

$ git checkout upstream-patches
$ git-apply < patch-from-lunar.diff   #444682 again
$ git commit --author 'Jérémy Bobbio <lunar@debian.org>' -s

  # this should also be automated, see below
$ git checkout master
$ dch -i
$ dpkg-parsechangelog | sed -ne 's,Version: ,,p'
2.6.3+200709292116+4450e59-3
$ git commit -s debian/changelog

$ poor-mans-gitbuild

$ git push
$ git push origin tag debian/2.6.3+200709292116+4450e59-3

That first git-push may require a short explanation: without any arguments, git-push updates only the intersection of local and remote branches, so it would never push a new local branch (such as build above), but it updates all existing ones; thus, you cannot inadvertedly publish a local branch. Tags still need to be published explicitly.

Hacking on the software

Imagine: on a rainy Saturday afternoon you get bored and decide to implement a better way to tell mdadm when to start which array. Since you're a genius, it'll take you only a day, but you do make mistakes here and there, so what could be better than to use version control? However, rather than having a branch that will live forever, you are just creating a local branch, which you will not publish. When you are done, you'll feed your work back into the existing branches.

Git makes branching really easy and as you may have spotted, the poor-mans-gitbuild script reserves an entire branch namespace for people like you:

$ git checkout -b tmp/start-arrays-rework master

Unfortunately (or fortunately), fixing this issue will require work on two branches, since the initramfs script and hook are maintained in a separate branch. There are (again) two ways in which we can (sensibly) approach this:

  • create two separate, temporary branches, and switch between them as you work.

  • merge both into the temporary branch and later cherry-pick the commits into the appropriate branches.

I am undecided on this, but maybe the best would be a combination: merge both into a temporary branch and later cherry-pick the commits into two additional, temporary branches until you got it right, and then fast-forward the official branches to their tips:

$ git merge master deb/initramfs
$ editor debian/mdadm-raid                     # …
$ git commit -s debian/mdadm-raid
$ editor debian/initramfs/script.local-top     # …
$ git commit -s debian/initramfs/script.local-top
[many hours of iteration pass…]

[… until you are done]
$ git checkout -b tmp/start-arrays-rework-init master
  # for each commit $c in tmp/start-arrays-rework
  # applicable to the master branch:
$ git cherry-pick $c
$ git checkout -b tmp/start-arrays-rework-initramfs deb/initramfs
  # for each commit $c in tmp/start-arrays-rework
  # applicable to the deb/initramfs branch:
$ git cherry-pick $c

This is assuming that all your commits are logical units. If you find several commits which would better be bundled together into a single commit, this is the time to do it:

$ git cherry-pick --no-commit <commit7>
$ git cherry-pick --no-commit <commit4>
$ git cherry-pick --no-commit <commit5>
$ git commit -s

Before we now merge this into the official branches, let me briefly intervene and introduce the concept of a fast-forward. Git will "fast-forward" a branch to a new tip if it decides that no merge is needed. In the above example, we branched a temporary branch (T) off the tip of an official branch (O) and then worked on the temporary one. If we now merge the temporary one into the official one, Git determines that it can actually squash the ancestry into a single line and push the official branch tip to the same ref as the temporary branch tip. In cheap (poor man's), ASCII notation:

- - - O             >> merge T >>     - - - = - - OT
       ` - - T      >>  into O >>

This works because no new commits have been made on top of O (if there would be any, we might be able to rebase, but let's not go there quite yet; rebasing is how you shoot yourself in the foot with Git). Thus we can simply do the following:

$ git checkout deb/initramfs
$ git merge tmp/start-arrays-rework-initramfs
$ git checkout master
$ git merge tmp/start-arrays-rework-init

and test/build/push the result. Or well, since you are not an mdadm maintainer (We\^W I have open job positions! Applications welcome!), you'll want to submit your work as patches via email:

$ git format-patch -s -M origin/master

This will create a number of files in the current directory, one corresponding for each commit you made since origin/master. Assuming each commit is a logical unit, you can now submit these to an email address. The --compose option lets you write an introductory message, which is optional:

$ git send-email --compose --to your@email.address <file1> <file2> <…>

Once you've verified that everything is alright, swap your email address for the bug number (or the pkg-mdadm-devel list address).

Thanks (in advance) for your contribution!

Of course, you may also be working on a feature that you want to go upstream, in which case you'd probably branch off upstream-patches (if it depends on a patch not yet in upstream's repository), or upstream (if it does not):

$ git checkout -b tmp/cool-feature upstream
[…]

… when a new upstream version comes around

After a while, upstream may have integrated your patches, in addition to various other changes, to give birth to mdadm-2.6.4. We thus first fetch all the new refs and merge them into our upstream branch:

$ git fetch upstream-repo
$ git checkout upstream
$ git merge upstream-repo/master

we could just as well have executed git-pull, which with the default configuration would have done the same; however, I prefer to separate the process into fetching and merging.

Now comes the point when many Git people think about rebasing. And in fact, rebasing is exactly what you should be doing, iff you're still working on an unpublished branch, such as the previous tmp/cool-feature off upstream. By rebasing your branch onto the updated upstream branch, you are making sure that your patch will apply cleanly when upstream tries it, because potential merge conflicts would be handled by you as part of the rebase, rather than by upstream:

$ git checkout tmp/cool-feature
$ git rebase upstream

What rebasing does is quite simple actually: it takes every commit you made since you branched off the parent branch and records the diff and commit message. Then, for each diff/commit_message pair, it creates a new commit on top of the new parent branch tip, thus rewrites history, and orphans all your original commits. Thus, you should only do this if your branch has never been published or else you would leave people who cloned from your published branch with orphans.

If this still does not make sense, try it out: create a (source) repository, make a commit (with a meaningful commit message), branch B off the tip, make a commit on top of B (with a meaningful message), clone that repository and return to the source repository. There, checkout the master, make a commit (with a …), checkout B, rebase it onto the tip of master, make a commit (with a …), and now git-pull from the clone; use gitk to figure out what's going on.

So you should almost never rebase a published branch, and since all your branches outside of the tmp/* namespace are published on git.debian.org, you should not rebase those.

But then again, Pierre actually rebases a published branch in his workflow, and he does so with reason: his patches branch is just a collection of branches to go upstream, from which upstream cherry-picks or which upstream merges, but which no one tracks (or should be tracking).

But we can't (or at least will not at this point) do this for our feature branches (though we could treat upstream-patches that way), so we have to merge. At first, it suffices to merge the new upstream into the long-living build branch, and to call poor-mans-gitbuild, but if you run into merge conflicts or find that upstream's changes affect the functionality contained in your feature branches, you need to actually fix those.

For instance, let's say that upstream started providing md.txt (which I previously provided in the deb/docs branch), then I need to fix that branch:

$ git checkout deb/docs
$ git rm md.txt
$ git commit -s

That was easy, since I could evade the conflict. But what if upstream made a change to Makefile, which got in the way with my configuration file location change? Then I'd have to merge upstream into deb/conffile-location, resolve the conflicts, and commit the change:

$ git checkout deb/conffile-location
$ git merge upstream
CONFLICT!
$ git-mergetool
$ git commit -s

When all conflicts have been resolved, I can prepare a new release, as before:

$ git checkout master
$ dch -i
$ dpkg-parsechangelog | sed -ne 's,Version: ,,p'
2.6.3+200709292116+4450e59-3
# git commit -s debian/changelog

$ poor-mans-gitbuild

# git push
$ git push origin tag debian/2.6.3+200709292116+4450e59-3

Note that Git often appears smart about commits that percolated upstream: since upstream included the two commits in upstream-patches in his 2.6.4 release, my upstream-patches branch got effectively annihilated, and Git was smart enough to figure that out without a conflict. But before you rejoice, let it be told that this does not always work.

Creating and using a maintenance branch

Let's say Debian "lenny" is released with mdadm 2.7.6-1, then:

$ git checkout -b maint/lenny debian/2.7.6-1

You might do this to celebrate the release, or you may wait until the need arises. We've already left the domain of reality ("lenny" is not yet released), so the following is just theory.

Now, assume that a security bug is found in mdadm 2.7.6 after "lenny" was released. Upstream is already on mdadm 2.7.8 and commits deadbeef and c0ffee fix the security issue, then you'd cherry-pick them into the maint/lenny branch:

$ git checkout upstream
$ git pull
$ git checkout maint/lenny
$ git cherry-pick deadbeef
$ git cherry-pick c0ffee

If there are no merge conflicts (which you'd resolve with git-mergetool), we can just go ahead to prepare the new package:

$ dch -i
$ dpkg-parsechangelog | sed -ne 's,Version: ,,p'
2.7.6-1lenny1
$ git commit -s debian/changelog

$ poor-mans-gitbuild

$ git push origin maint/lenny
$ git push origin tag debian/2.7.6-1lenny1

Future directions

It should be trivial to create the Debian source package directly from the repository, and in fact, in response to a recent blog post of mine on the dispensability of pristine upstream tarballs, two people showed me their scripts to do it.

My post also caused Joey Hess to clarify his position on pristine tarballs, before he went out to implement dpkg-source v3. This looks very promising.

Yet, as Romain argues, there are benefits with simple patch management systems. Exciting times ahead!

In addition to creating source packages from version control, a couple of other ideas have been around for a while:

  • create debian/changelog from commit log summaries when you merge into the build branch. Guido's git-dch might be a lead.

  • integrate version control with the BTS, bidirectionally:

    • given a bug report, create a temporary branch and apply any patches found in the bug report.

    • upon merging the temporary branch back into the feature branch it modifies, generate a patch, send it to the BTS and tag the bug report + pending patch.

And I am sure there are more. If you have any, I'd be interested to hear about them!

Wrapping up

I hope this post was useful. Thank you for reading to the end, this was probably my longest blog post ever.

I want to thank Pierre Habouzit, Johannes Schindelin, and all the others on the #git/freenode IRC channel for their tutelage. Thanks also to Manoj Srivastava, whose pioneering work on packaging with GNU arch got me started on most of the concepts I use in the above. And of course, the members of the the vcs-pkg mailing list for the various discussions on this subject, especially those who participated in the thread leading up to this post. Finally, thanks to Linus and Junio for Git and the continuously outstanding high level of support they give.

If you are interested in the topic of using version control for distro packaging, I invite you to join the vcs-pkg mailing list and/or the #vcs-pkg/irc.oftc.net IRC channel.

NP: Aphex Twin: Selected Ambient Works, Volume 2 (at least when I started writing…)

Posted Fri 11 Jul 2008 11:21:05 CEST Tags:

Despite a distributed version control system by design, git can just as well be used in a centralised fashion. When a user clones a git repository, git sets up the local clone such that it is aware of its origin. Let's have a look at the relevant lines in .git/config:

[remote "origin"]
  url = ssh://server/path/to/repo.git
  fetch = +refs/heads/*:refs/remotes/origin/*
[branch "master"]
  remote = origin
  merge = refs/heads/master

The "remote" stanza defines a remote repository with the nickname "origin". If the user calls git fetch without arguments, it will download all remote branches (refs/heads/*) and store those as remote branches with the origin/ prefix. Thus, the branch "big-fat-feature" in refs/heads/big-fat-feature on the other side becomes the remote branch "origin/big-fat-feature" (refs/remotes/origin/big-fat-feature).

The "branch" stanza informs git about the default source for merges when the "master" branch is current. If the local "master" branch is checked out and the user calls git pull without arguments, it instructs git to fetch all branches (see above), then merge the remote "master" branch (refs/heads/master) into the current branch.

I started a new project in git today and decided to go public after I had already made a few commits and created a number of branches.

So I set out to learn a bit more about git internals and came up with two ways to publish the repository such that my local repository would also know about the remote side as if it had been cloned from the start. The documentation didn't leave me satisfied, so I had to try it all out.

Being new to git, my first thought was to push/publish my repository to a public location, and so I did:

### assuming ssh://server/remote.git resolves to an empty, bare git repo
### and that we are chdir()'d to the local repository:
$ git push --all ssh://server/remote.git
$ git remote add origin ssh://server/remote.git
$ git config branch.master.remote origin
$ git config branch.master.merge refs/heads/master
$ git fetch
$ git merge master
$ git branch
  big-fat-feature
* master
  ponies
$ git branch -r
  origin/big-fat-feature
  origin/master
  origin/ponies

Another method then dawned on me, but it's not necessarily better. Instead of pushing, let's copy a clone over and then clone that once more:

$ git config core.bare true
$ scp -r .git server:/remote.git
$ mv `pwd` `pwd`/../OLD
$ cd ..
$ git clone ssh://server/remote.git
$ cd remote
$ git branch
  big-fat-feature
* master
  ponies
$ git branch -r
  origin/HEAD
  origin/big-fat-feature
  origin/master
  origin/ponies

This does almost the same, except for that origin/HEAD branch, but that's just a local symbolic ref (a "symlink") and can just be removed, really:

$ rm .git/refs/remotes/origin/HEAD
$ git branch -r
  origin/big-fat-feature
  origin/master
  origin/ponies

Now the only thing left is setting core.sharedRepository on the remote side to let git know how to handle the Unix permissions. And that's it.

NP: Antimatter: Lights Out

Posted Fri 11 Jul 2008 11:21:05 CEST Tags:

Previously, I demonstrated a Debian packaging workflow using Git and I mentioned the possibility of a follow-up post; well, here it is: you want to use my workflow (or one that's related) for a package that is currently maintained with Subversion on svn.debian.org and you'd like to keep the history during the conversion.

Make sure to read the previous post before this one.

I am again using the example of mdadm since its Git packaging repository is in a state of shambles and I want to restart to get it right and import the history from the previous Subversion repository. What better way than to write a blog post as I do so? Well, plenty actually. This kind of post isn't really made for a blog, and I have started work on setting up ikiwiki on madduck.net, but it's not yet ready, so I'll stick with the blog for now. I will make sure that links don't break as I move content over, so feel free to bookmark this…

Importing the package into Git

Thanks to git-svn, the initial step of getting your package imported into Git is a breeze:

$ git-svn clone --stdlayout --no-metadata \
    svn+ssh://svn.debian.org/svn/pkg-mdadm/mdadm mdadm

Sit back and enjoy. If that command exits prematurely with an error such as the following:

Malformed network data: Malformed network data at /usr/local/bin/git-svn line 1029

then you should upgrade to a newer Git version, or have a look here. If your Git does not know --stdlayout then upgrade as well (or use -T trunk -t tags -b branches instead).

Sam Vilain notes that it is important to "get the attribution right with the final SVN import - getting the authors map right. I didn't do that. If you look at the repository resulting from the above command, you'll notice strange commit authors, such as madduck@some-unique-uuid-from-svn. git-svn allows you to map these to real names with real email addresses, which ensures that the attributions are good for the whole world to see.

When done, switch to the repository and run git-branch -r. As you'll see, git-svn imported all SVN branches and tags as remote branches. You need those if you want to bidirectionally track the Subversion repository, but we are converting, as you may have guessed by the --no-metadata switch above.

Therefore, we resort to the Dinosaur method of converting branches to tags, which I'll simplify for mdadm. We also just delete all remote branches after tagging, since mdadm never used branches in the SVN repository. Your mileage may vary.

git branch -r | sed -rne 's, *tags/([^@]+)$,\1,p' | while read tag; do
  echo "git tag debian/$tag tags/${tag}^; git branch -r -d tags/$tag"
done

git branch -r | while read tag; do
  echo "git branch -r -d $tag"
done

If that seems to work alright, then you can execute the commands.

Sam Vilain (again) hints me at git-pack-refs and then to edit .git/packed-refs with an editor. This certainly leaves more room for errors but might be significantly faster.

Cleaning up the SVN references

Even though we passed --no-metadata to git-svn, it did leave some traces in .git/, which we can now safely remove:

$ git config --remove-section svn-remote.svn
$ rm -r .git/svn

Setting things straight

You can skip this section unless you want to know a bit about how to fix up stuff with Git.

There was actually some nasty tagging errors leading up to the 2.5.6-9 release for etch and I could never be bothered to fix those in SVN, but now I can (I love Git!):

$ git tag -d debian/2.5.6-10            # never existed
$ git tag -f debian/2.5.6-8 2.5.6-8~2   # mistagged
$ git checkout -b maint/etch 2.5.6-8    # this is when we diverged
$ git apply < /tmp/mdadm-2.5.6-8..2.5.6-9.diff
$ git add debian/po/gl.po debian/po/pt.po debian/changelog
$ git commit -s
$ git tag debian/2.5.6-9

Now that that's fixed, there is one other thing to worry about, namely the very last commit to SVN, which obsoletes the repository and points to the Git repository. But that's not all of it. I was also silly enough to include a fix in the same commit. Let's see what Git can do. Since the process of obsoletion involves all but adding a file, we can simply --amend the last commit and provide a new log message:

$ git checkout master
$ git rm OBSOLETE debian/OBSOLETE
$ git commit --amend

Now the repository is in an acceptable state.

Making ends meet

The pkg-mdadm effort on svn.debian.org only maintained the ./debian/ directory, separate from the upstream code, and boy was that a bad idea. Just to give one example: think about what's involved in preparing a Debian-specific patch against the upstream code… this has to end, and we can make it end right here; let's import upstream's code (again not using his ADSL line, but the upstream branch of the pkg-mdadm Git repository; see the previous post for details):

$ git remote add upstream-repo git://git.debian.org/git/pkg-mdadm/mdadm
$ git config remote.upstream-repo.fetch \
    +refs/heads/upstream:refs/remotes/upstream-repo/upstream
$ git fetch upstream-repo
$ git checkout -b upstream upstream-repo/master

Now we have two unconnected ancestries in our repository, and it's time to join them together. The most logical way seems to be to use the last upstream tag for which we have a Debian tag: 2.6.2.

For this, we branch off the corresponding Debian tag (2.6.2-1) and merge upstream's 2.6.2 tag into the new branch. This will be a temporary branch Then, we rebase (remember, nothing has been published yet) the master branch on top of this temporary branch, before we end that branch's short life. The Debian tag stays where it is since it describes the state of the repository at time of the release of 2.6.2-1.

$ git checkout -b tmp/join debian/2.6.2-1
$ git merge mdadm-2.6.2

$ git rebase tmp/join master
$ git branch -d tmp/join

It just so happens that the head of the SVN repository, which is identical to the tip of our master branch, corresponds to Debian release 2.6.2-2, so we tag it:

$ git tag debian/2.6.2-2

We are now also "born" in the sense that maintenance in Git has started. Let's mark that point in history. There is no real reason I can foresee for this yet, but nonetheless:

$ git tag -s git-birth

Turning dpatch files into feature branches

We want to turn dpatch files into feature branches and we somehow make it "proper". We could branch, apply the patch, delete the patch file, checkout master and delete the patch file there as well, but that appears "improper" to me at least; so instead, we'll cherry-pick:

$ git checkout -b deb/conffile-location
$ debian/patches/01-mdadm.conf-location.dpatch -apply
$ git rm debian/patches/01-mdadm.conf-location.dpatch
$ git commit -s
$ git commit -s $(git ls-files --others --modified)

I should quickly intervene to make sure you are following. I am making use of Git's index here. Applying the patch makes the changes in the working tree, but we did not tell Git that we want those to be part of the commit just yet. Instead, we delete the dpatch with git-rm, which automatically registers the deletion with the index. Thus, the first git-commit creates a commit which deletes the dpatch, while the second git-commit creates a commit with all the changes from the dpatch, using git-ls-files to identify new and modified files.

But for now, let's move on. We have two commits in the deb/conffile-location branch, and one of those is relevant to the master branch, we cherry-pick it:

$ git cherry-pick deb/conffile-location^

If you're confused, let me explain: our goal is to have a number of feature branches, of which master is the one in which most of ./debian/ is maintained. All the branches later come together in the long-living build branch, so deb/conffile-location will never be merged back into master. However, once we applied the dpatch to the feature branch, we can delete it from there and the master branch. By cherry-picking, we "import" the deletion to the master branch.

I repeat the same procedure for deb/docs, merging all the documentation-related dpatches, but I'll spare you the details.

… and then Git let me down

In the next step, I found I had misunderstood Git merging: I thought Git was smart, but Linus had his reasons for calling Git the "stupid content tracker" (more on that later). Read on as I am obsoleting dpatch files that upstream had merged: 99-*-FIX.dpatch.

For consistency, I wanted to cherry-pick each of the appropriate upstream commits into the master branch along with deleting the corresponding dpatch file. Here is one example: 99-monitor-6+10-FIX.dpatch was obsoleted by upstream's commit 66f8bbb; the -x records the original commit ID in the log:

$ git cherry-pick -x 66f8bbb
$ git rm debian/patches/99-monitor-6+10-FIX.dpatch
$ git commit -s -m"remove dpatch obsoleted by $(git rev-parse --short HEAD)"

I repeated the procedure for the other dpatch files, removed the dpatch infrastructure, and then went on to merge it all into build to build the package.

The build branch is a long-living branch off upstream, but which upstream? I'll fast-forward you past a segfault problem with mdadm, which upstream (thought to have) resolved with commit 23dc1ae after 2.6.3, but he had not yet released 2.6.4. Looking at the commits between 23dc1ae and upstream's HEAD at the time, I decided to include them all and snapshot 4450e59:

$ git fetch upstream-repo
$ git checkout upstream
$ git merge upstream-repo
$ git tag mdadm-2.6.3+200709292116+4450e59 4450e59

$ git checkout master
$ git merge --no-commit mdadm-2.6.3+200709292116+4450e59
$ dch -v mdadm-2.6.3+200709292116+4450e59-1
$ git add debian/changelog
$ git commit -s

And then I called poor-mans-gitbuild, which merges master and then deb/* into build. Here is when stuff blew up.

I'll make a long story short (read my description of the problem and Linus' answer if you want to know more): I thought Git was smart to identify merges common to both branches and do the right thing, but it turn out that Git does not care at all about commits, it only worries about content and the end result. In our case, unfortunately (or fortunately), the outcome meant a conflict because the upstream branch introduced a simple change (last hunk) in the lines surrounding the patch we cherry-picked, and Git can't handle it.

The solution is not to cherry-pick, to cherry-pick all commits touching the context of the dpatch, or to simply merge upstream into all out feature branches. In our case, the first is the easiest solution and since importing dpatch files is a one-time thing (thank $DEITY), I'll leave it at that.

Almost.

I have spent two days thinking about this more than I should have. And it was this point Linus made which made me appreciate Git even more:

Conflicts aren't bad - they're good. Trying to aggressively resolve them automatically when two branches have done slightly different things in the same area is stupid and just results in more problems. Instead, git tries to do what I don't think anybody else has done: make the conflicts easy to resolve, by allowing you to work with them in your normal working tree, and still giving you a lot of tools to help you see what's going on.

The end

This concludes today's report. Importing the changes from the old Git repo, tagging and merging the branches is all covered in my previous post, or at least you'll find enough information there to complete the exercise.

I would like to specifically thank Sam Vilain and Linus Torvalds for their help in preparing this post, as well as the #git/freenode inhabitants, as always.

If you are interested in the topic of using version control for distro packaging, I invite you to join the vcs-pkg mailing list and/or the #vcs-pkg/irc.oftc.net IRC channel.

Also, if you are interested in Git in general, you can find a list of blog posts on the Git wiki.

NP: The Police: Zenyatta Mondatta

Posted Fri 11 Jul 2008 11:21:05 CEST Tags: