Let say there is a team working on main Git repository using branching model. Now a second team joins and is starting to work a subset of a project. As the starting point they need to collaborate on one folder from the repository. They are not allowed to see rest of the code base. What is the best way to achieve that?
Going forward they would need to be able to merge their changes into the main code base and get any updates from that one folder along the way too.
This is all based on Windows OS with Atlassian Stash and Git on internal network.
That would mean that one folder needs to be its own repo:
added as a submodule (tracking a branch) in the main Git repo
forked by the second team, for them to push to the fork and make PR or synchronize from the original folder git repo.
I would suggest separating the subset into a sub-project, and use language-specific ways to deliver it during building of you main project. For example, if you use MS Visual Studio, you could turn it into library or module and use nuget to deliver it during build of your main project.
In my experience it appears to be much more convenient than using submodules, when it comes to merging.
Another reason to do that - and maybe even more important one - that the other team would be able to handle the project as a compilable and testable unit, instead of a pile of source files.
Related
My team is working on translating several legacy mobile applications to Xamarin Forms apps. Currently each application is in its own solution, which is not ideal when it comes to the fact that they all use a common set of backend software libraries. We were planning to consolidate all the smaller solutions into a single solution, containing the apps as well as the common libraries.
However, one of my teammates brought up a valid concern about how with a single Xamarin Forms app, several projects could get generated (core, Android, iOS, etc.), with the eventual result of a generally unwieldy solution. I agree with him that the current setup probably would not scale too well as we add more apps -- even if we group projects in solution folders, Visual Studio will eventually slow to a crawl after a certain amount of projects exist in the solution.
So we are considering just going back to having each app in its own solution, each solution containing the few Xamarin Forms projects for that app, as mentioned above. But this brings us back to the question of how to reasonably manage the shared library code. My current thought would be to just use shared project(s) for the libraries, or maybe assemble them into NuGet package(s) the app solutions would consume. Am I on the right track here, or does anyone know of a better way to do this?
There are several different ways to manage a shared code project using subtrees, submodules, NuGet packages, etc. There are pros and cons to each so it's best to decide based on the expected use case for that project.
Subtrees essentially take a copy of the remote repo and pull it into the parent repo. This makes it easy to pull in changes from the remote repo but if changes are expected to be pushed back it can be significantly more difficult since it has no knowledge of the remote repo. While it is possible to push changes back it can take a significant among of time to do depending on the amount of history of the repos.
Submodules are similar to subtrees except that instead of taking a copy it tracks the remote repo based on a specific commit it's pointed to. This essentially can be thought of as another repo inside of the parent that makes pushing changes back to the remote repo much easier but at the cost of making pulling/updating from it a little bit more difficult.
NuGet packages are extremely convenient to install, update, and release to others without having to make the source code public, but that comes with a bit more initial setup to generate each package version and comes at the cost of making it more difficult to debug than with the actual source code. This is particularly a great option if the shared code library will be distributed to others.
For most projects, if changes are expected to be potentially made to that shared project from a consuming one I'd recommend a repo for each project and set up the shared one as a submodule in each. It does take a bit of learning to get used to the different processes of checking out and updating a submodule but actually isn't all that difficult and worth learning the few git commands required. The docs provide a great example of how to get started using submodules.
I have several projects which use code from a large set of component libraries. These libraries are under source control.
The libraries repository contains all the libraries used by all my projects and contains multiple versions of multiple libraries. Each library/version pair lives in its own folder. Each of my projects identifies the specific library/version pairs it needs through the folder paths of the references in its project file.
For example $(LibraryPath)\SomeLibrary\v1.1.5
Please note that the libraries repository is only ever added to. No changes are made to stuff already in the repository. Ever.
I have been of course been able to configure my build plan to pull the libraries repository to a libraries subfolder of the working directory. So far so good. However, using the auto branch management feature of Bamboo, this setup means that the libraries repository is cloned for each and every branch in all projects.
Not funny. No, really, not funny...
What I would like to do is:
pull the libraries repository in each build plan
but pull it to a fixed location that is the same for all build plans
it doesn't have to be an absolute path
but it does need to be outside the working directory of the current build plan to avoid unnecessary duplication
Unfortunately the Checkout Directory of the Source Code Checkout configuration task in a Bamboo build plan doesn't allow me to specify either an absolute path or a relative one that goes "up" for one or more levels from the working dir. The hint text explicitly states "(Optional) Specify an alternative sub-directory to which the code will be checked out." And indeed, specifying something like ..\Library gets punished with the message "Checkout to parent directory is forbidden".
I have seen information on the "artifact sharing" feature of Bamboo. This will probably work, but it seems like overkill for what I want to achieve.
What would be the easiest and least complicated way to achieve my goal using Atlassian's Bamboo Continuous Integration?
Out-of-the-box alternatives are welcome, but please don't direct me to any products that require intimate CLI use and/or whose documentation assumes (extensive) knowledge of 'nix and/or Java setup. I am on Windows and spoiled rotten by powerful (G)UI's.
I have the same problem - with a repository weighing in at around 2GB.
I'd like to simply "git checkout myBranch" and "git clean -fxd" instead of cloning every time (which should save a lot of time and disk space). However I also like Bamboo's automatic trigger with new branches showing up.
Like the OP, I'd love to be able to put "..\SharedDirectory" in the "CheckoutDirectory" for the
"Source Code Checkout" task but it won't let me go out above the \JOB_KEY\ folder
One possible solution is: replacing the "Source Code Checkout" task with the two git commands above. That way I can specify exact when/where/how to do the checkout. I think there may be problems with the initial checkout in this case - but once that is solved, all subsequent branches would use the same shared folder, and no more pulling down 2GB every time.
We develop with Visual Studio 2010 (in C#) and migrated a while ago from SVN to GIT. Now we try to split up our repository (which is quite big - ~30.000 files) to many git repositories - one for each solution.
The solutions share some projects, mostly libraries we develop in-house and like to add to from all the solutions.
The new repositories have a flat layout. One subdirectory for each project (shared projects are submodules).
In the big old repo, the projects are in a tree structure.
The Problem occurs with external references in the submodules. In the new repos, the path to a referenced project may be "......libs\someproject", while in the new layout the correct path would be "..\someproject".
We already had some edit wars concerning this and are not keen on more.
Half-baked Solutions I could think of:
use "Reference Paths" in ...csproj.user and exclude this file from version control (has to be redone for each developer and after each reopsitory cleanup)
use branches for each situation and try to teach everyone where "real" commits should go and where "environment-change" commits should go (submodules are already not the simplest concept...)
embed binaries instead of the submodules (but what about developing changes to the submodules? what about different log4net versions?)
Does anyone know of a sane solution?
Since you are asking for a sane solution, I can only advise you to look into setting up your own NuGet service (look at http://www.MyGet.org for inspiration)
http://nuget.codeplex.com/
IF you go down the route of package management, consider OpenWrap. However, embedding the package management artefacts in source code is a bad idea. You can use such tools to update what is actually stored in submodules, but don't rely on them at build time. Expect the binaries to be there from the point of view of your build scripts.
So if I understand you correctly, the problem is with Visual Studio and not with Git? If that's the case, use the old tree structure that worked with Visual Studio. Make your submodules structure a tree structure too. So the top of the tree would be one super repo whose sub modules (the branches) would have submodules of their own, until you get down to the leaves of your tree. It would be a pain to setup at first, but it should just work.
Use one submodule to house all "common libraries". Just one level. But you should move the common libraries as services with well defined contracts. This way you can incrementally rollout new versions with no down time. This way you only have a submodule in each that holds the contracts. These could be interfaces or messages.
I have a similar problem using VS 2013.
I want to use git-svn instead of SVN directly. SVN has a gigantic set of directories. I could not create a single git-repository that would contain all of our trunk folder. Git-always exited with an error and the repository was corrupted. I worked around the problem by doing as follows:
Using git-svn, I cloned the subset of folders off SVN/trunk that I needed by creating one git-repository per folder.
Created a local parent git repository that contains all my git-svn-cloned folders.
Each git-repository was added as a sub-module to the parent git-repository.
The problem with Visual Studio is that it does not recognize the multiple projects outside the main project where I opened the solution. This solution is in a folder that contains the only files recognized by Visual Studio as being under git-source control.
I tried setting the git-preferences to use the upper level parent directory as the location of the git-repostitory without noticing any difference.
If I have a C# solution with multiple projects in it, what would be better, to have the Git repo created in the solution folder, or in each individual project folder? Multiple developers will be working on the projects. What are your experiences with this?
I use several (sometimes overlapping) solutions to contain a collection of related independent applications and shared libraries. As others have mentioned, you really don't want to have a single Git repository containing the source for multiple, independent projects as it makes it much too difficult to track isolated changes.
So, if your solution is structured as mine is then you will definitely want individual Git repositories for each project. This has worked well for me for ten to twelve applications and doesn't create as much maintenance overhead as you might think.
If your solution is truly monolithic (and you're sure you want it that way forever and ever), then it probably makes sense to only have a single repository.
It depends. git repositories are most suited to containing a single configuration item with its own independent lifecycle. If your projects have their own release cycle and are shared between multiple solutions then it might make sense to have them in their own repositories. Usually, though, it is the solution that represents a configuration item with all the constituent projects forming part of the same build. In this case a single git repository at the solution level makes more sense.
git submodule is probably worth consideration here. Each project gets its own repo, the solution gets a repo, and the projects are submodules.
I assume that your solution represents some kind of a product while the projects are just a part of the product.
In this situation I would create the repository on the solution level. This way it is a lot easier to build the whole product at once, especially if the projects depend on each other.
Some though and 3 solutions on the subject can be read on that blog:
https://www.atlassian.com/blog/git/git-and-project-dependencies
package management tool, i.e. nuget in VS, so using reference to a package/compiled module
git submodule (only with command line in VS?)
other build and cross-stack dependency tools
Another solution is just to add a project from the other repo and let it out of the current repo, and latter use the Team Explorer to commit its changes.
Just moved to subversion...from visual studio. I love it already! Can someone briefly explain
Repository
Branches
Tags
Trunk
Do I need to create a new repository for every project? Or a new trunk?
Thanks
You don't need a separate repository, but you can if you want. I recommend reading the book at http://svnbook.red-bean.com/. Grab the pdf version or whatever. It doesn't take too long, and it explains some things pretty well. I read it, and found that I'm glad I did.
Remember that subversion is just a fancy filesystem that supports versioning. Think of a repository as a "drive root" like "C:/".
Each project gets a trunk, tags and branches directory. All of your day to day work happens in the trunk. Experimental code is done in a branch and then merged back into the trunk at a later date. Tags are for when you release the software. These are not to be edited. When you release the software, you create a tag with a unique name based on what is currently in the trunk.
I can't say whether or not you need a separate repository for each project, there are pros and cons. This blog posting details them:
Simplified administration. One set of hooks to deploy. One repository
to backup. etc.
Branch/tag flexibility. With the code all in one repository it makes it
easier to create a branch or tag
involving multiple projects.
Move code easily. Perhaps you want to take a section of code from
one project and use it in another, or
turn it into a library for several
projects. It is easy to move the code
within the same repository and retain
the history of the code in the
process.
Here are some of the drawbacks to the
single repository approach, advantages
to the multiple repository approach.
Size. It might be easier to deal with many smaller repositories than
one large one. For example, if you
retire a project you can just archive
the repository to media and remove it
from the disk and free up the storage.
Maybe you need to dump/load a
repository for some reason, such as to
take advantage of a new Subversion
feature. This is easier to do and with
less impact if it is a smaller
repository. Even if you eventually
want to do it to all of your
repositories, it will have less impact
to do them one at a time, assuming
there is not a pressing need to do
them all at once.
Global revision number. Even though this should not be an issue,
some people perceive it to be one and
do not like to see the revision number
advance on the repository and for
inactive projects to have large gaps
in their revision history.
Access control. While Subversion's authz mechanism allows
you to restrict access as needed to
parts of the repository, it is still
easier to do this at the repository
level. If you have a project that only
a select few individuals should
access, this is easier to do with a
single repository for that project.
Administrative flexibility. If you have multiple repositories, then
it is easier to implement different
hook scripts based on the needs of the
repository/projects. If you want
uniform hook scripts, then a single
repository might be better, but if
each project wants its own commit
email style then it is easier to have
those projects in separate
repositories
I agree, read the svnbook. It's a great resource.
Do I need to create a new repository for every project? Or a new trunk?
Kevin covered the single/multiple repository trade-offs pretty well. When we started with svn, we used one repository for all of our development projects. It worked well and had all the advantages mentioned. However, as the repository got bigger it got more difficult to administer because of the size of the dump file and resulting issues during backup. It also became an issue that projects couldn't easily be archived out of the repository - it's certainly possible but it requires dumping and pulling out projects from the repository. They aren't issues you can't get around but it's something to keep in mind.
Repository
Branches
Tags
Trunk
Branches, tags and the trunk are just copies of your files contained in the repository. It allows you to segregate and check-mark your files at whatever time you feel appropriate (usually at a release or a feature branch).
An important thing to keep in mind about branches, tags and trunk is that they just conventions in svn. There is no functional difference between the three locations, they are just an accepted usage model and they can be changed or organized differently if you have a good reason. I'm not recommending that you organize differently but you'll find that svn is very flexible because there isn't really a forced organizational structure other than convention.
Depending on how many projects you decide to have in your repository, you may organize differently.
You can have the subdirectories with projects under it:
\repo
\branches
\...
\tags
\...
\trunk
\..
or you can have projects contain the subdirectories:
\repo
\Project1
\branches
\tags
\trunk
\Project2
\branches
\tags
\trunk
There are trade-offs that are covered in the svnbook. The first method is usually used if you only have one project per repository and the second if there is more than one project in your repository.
The nice thing is that you can just start using svn and then figure out what you prefer. You should have some sort of organization but, with cheap copies, you can always re-arrange the folders as your situation or workflow changes.
An important thing to remember with SVN, compared to other version control systems like CVS or Git, is that SVN doesn't really have a concept or branching or tagging. As far as SVN is concerned it's all just a bunch of folders and files. So while you'll see a lot of people using the branches/tags/trunk setup, this is not required and you are able to deviate from this if you so choose.
Generally speaking 'trunk' is where you keep your active development going. So this is where you do all your commits. Whether or not you checkout trunk or use tags/branches instead is entirely up to you.
Branches, as I've used them, are usually for when you need to do large changes to your application but don't want them in trunk because you want to be able to continue developing against trunk without deploying your other changes. In this case you may have something like
\repo
\trunk
\branches
\version_two
In this case you can develop in both trunk and version_two separately and, assuming your live site is a checkout of trunk, you don't need to worry about 'accidentally' breaking your live site with your other changes. And when those changes are done and ready you just merge them back into trunk whenever you want.
Tags can be used similarly to branches, in that instead of checking out trunk and just using 'svn up' to update your repository you instead of several tags, each representing one release. So your repo may look something like
/repo
/trunk
/branch
/version_one
/version_two
/tags
/1.0.0
/1.0.1
/1.1.0
In this case the general idea is that when you're ready to do a deploy you do an
svn copy
To copy trunk over to a tag (in this case the next one could be 1.1.1, 1.2.0, 2.0.0, etc). How you name your tags it entirely up to you though and, again, depends on your project and requirements. With this route instead of doing a regular 'svn up' you would have to do an svn switch. So you have to deploy with
svn switch https://svn.yourrepo.com/repo/tags/1.1.0
The switch will automatically do updates, adds and deletes on the appropriate files.
When it comes to one repo for many projects or separate repos for each one I am an advocate of one repo per project. It provides the additional benefits of easily managing access to it. But most importantly it means that each project has a separate commit history and separate logs. This m
Reading your tags I see you started using VisualSVN instead of your old VSS system. (Your question says you stopped using Visual Studio.. which makes VisualSVN a strange choice).
One of the major differences between SourceSafe and VSS is that you can choose different tools to access the same repository (and you can switch every time you like as they all share the same workingcopy).
E.g.:
TortoiseSVN for Explorer integration.
The normal subversion client for scripts.
VisualSVN as Visual Studio frontend for TortoiseSVN
AnkhSVN as real SCC (VAPI) package in Visual Studio.