Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 6 years ago.
Improve this question
Why should I do Nightly Builds?
You should do nightly builds to ensure that your codebase stays healthy.
A side effect of doing nightly builds is that it forces the team to create and maintain a fully automated build script. This helps to ensure that your build process is documented and repeatable.
Automated builds are good at finding the following problems:
Somebody checked in something that breaks stuff.
Somebody forgot to check in a necessary file or change.
Your build scripts no longer work.
Your build machine is broken.
Doing this nightly ensures that you catch such problems within 24 hours of when they occur. That is preferable to finding all the problems 24 hours before you are supposed to deliver the software.
You should also, of course, have automated unit tests that are run for each nightly build.
I've personally found continuous integration to be better than nightly builds:
http://en.wikipedia.org/wiki/Continuous_integration
I even use it on one man projects, it's amazing how fast you can expose issues and take care of them right there.
I've been doing build engineering (among other things) for 16 years. I am a strong believer in build-early, build-often, continuous integration. So the first thing I do with a project is establish how it will be built (Java: Ant or Maven; .NET: NAnt or MSBuild) and how it will be managed (Subversion or some other version control). Then I'll add Continuous Integration (CruiseControl or CruiseControl.NET) depending upon the platform, then let the other developers loose.
As the project grows, and the need for reports and documentation grows, eventually the builds will take longer to run. At that point I'll split the builds into continuous builds (run on checkin) that only compile and run unit tests and daily builds that build everything, run all the reports, and build any generated documentation. I may also add a delivery build that tags the repository and does any additional packaging for a customer delivery. I'll use fine-grained build targets to manage the details, so that any developer can build any part of the system -- the Continuous Integration server use the exact same build steps as any developer. Most importantly, we never deliver a build for testing or a customer that wasn't built using the build server.
That's what I do -- here's why I do it (and why you should too):
Suppose you have a typical application, with multiple projects and several developers. While the developers may start with a common, consistent development environment (same OS, same patches, same tools, same compilers), over the course of time their environments will diverge. Some developers will religiously apply all security patches and upgrades, others won't. Some developers will add new (maybe better) tools, others won't. Some will remember to update their complete workspace before building; others will only update the part of the project they're developing. Some developers will add source code and data files to the project, but forget to add them to source control. Others will write unit tests that depend upon specific quirks of their environment. As a consequence, you'll quickly see the ever-popular "Well, it builds/works on my machine" excuses.
By having a separate, stable, consistent, known-good server for building your application, you'll easily discover these sorts of problems, and by running builds from every commit, you'll be able to pinpoint when a problem crept into the system. Even more importantly, because you use a separate server for building and packaging your application, it will always package everything the same way, every time. There is nothing worse than having a developer ship a custom build to a customer, have it work, and then have no idea how to reproduce the customizations.
When I saw this question, first I searched for Joel Spolsky's answer. Bit disappointed, so I planned to add it here.
Hope everyone is aware of Joel Test on Careers.
From his blog on The Joel Test: 12 Steps to Better Code
3. Do you make daily builds?
When you're using source control, sometimes one programmer
accidentally checks in something that breaks the build. For example,
they've added a new source file, and everything compiles fine on their
machine, but they forgot to add the source file to the code
repository. So they lock their machine and go home, oblivious and
happy. But nobody else can work, so they have to go home too, unhappy.
Breaking the build is so bad (and so common) that it helps to make
daily builds, to insure that no breakage goes unnoticed. On large
teams, one good way to insure that breakages are fixed right away is
to do the daily build every afternoon at, say, lunchtime. Everyone
does as many checkins as possible before lunch. When they come back,
the build is done. If it worked, great! Everybody checks out the
latest version of the source and goes on working. If the build failed,
you fix it, but everybody can keep on working with the pre-build,
unbroken version of the source.
On the Excel team we had a rule that whoever broke the build, as their
"punishment", was responsible for babysitting the builds until someone
else broke it. This was a good incentive not to break the build, and a
good way to rotate everyone through the build process so that everyone
learned how it worked.
Though I haven't got an opportunity to make daily builds, I'm a great fan of it.
Still not convinced? Check out the brief here in Daily Builds Are Your Friend!!
You don't actually, what you should be wanting is Continuous Integration and automatic testing (which is a step further than nightly builds).
If you are in any doubt you should read this article by Martin Fowler about Continuous Integration.
To summarize, you want to build and test as early and often as possible to spot errors immediately so they can be fixed while what you were trying to achieve when you caused them is still fresh in your mind.
I'd actually recommend to do builds every time you check in. In other words, I'd recommend setting up a Continuous Integration system.
The advantages of such a system and other details can be found in Fowler's article and on the Wikipedia entry among other places.
In my personal experience, it's a matter of Quality Control: every time code (or tests, which can be seen as a form of requirements) are modified, bugs might be creeping in. To ensure quality you should make a fresh build of the product as it would be shipped and perform all the tests available. The more often this is done, the less likely bugs will be allowed to form a colony. Therefore, daily (nightly) or continuous cycles are preferred.
In addition, whether you restrict access o your project to developers or a larger group of users, a nightly build enables everyone to be on the 'latest version', minimizing the pain of merging their own contributions back into the code.
You want to do builds on a regular schedule in order to catch problems with integration of code between developers. The reason you want to do this nightly, as opposed to weekly or on some longer schedule, is that the longer you wait to discover these kinds of problems, the more difficult it will be to resolve them. The practice of doing a build on every check in (Continuous Integration) is just taking the nightly build process to a logical extreme.
The side benefit of having a repeatable build process is important in the long run as well. If you work on a team where there are multiple projects going on, then at some point you will need to be able to easily recreate an old build, perhaps for creating a patch. :(
The more you can automate the build process, the more time you will save for each subsequent build. It also takes the build process itself off of the critical path of delivering the final product, which should make your manager happy. :)
It also depends on the size and structure of the team(s) working on your project. If there are different teams relying on each others API, it may make a lot of sense to have nightly builds for frequent integration. If you're hacking away with only one or two team mates it may or may not be worth it.
Depending on the complexity of your product continuous integration may or may not be able run a full test suite.
Imagine Cisco testing a router with the literally 1000s of different setups to test. To run a full test suite on some products takes time. Sometimes weeks. So you need builds for different purposes. A nightly build can be the basis for a more thorough test suite.
I think they are very important especially on projects with more than 1 person. The team needs to know ASAP if someone:
checks in a bad file
doesn't check in a file
...
Any build automation is better than no build automation :-)
Personally, I prefer daily builds - that way if the build doesn't work then everyone is around to get it fixed.
In fact, if at all possible then Continuous Integration builds are the way to go (i.e. a build on every check-in) as that minimizes the amount of change between a build and so makes it easy to tell who broke the build and also easy to fix the build.
Well ... I guess it depends a lot on your project, of course. If it's just your hobby project, with no releases, no dependencies, and noone but you submitting code, it might be overkill.
If, on the other hand, there's a team of developers all submitting code, automatic nightly builds will help you ensure the quality of the code in the repository. If someone does something that "breaks the build" for all others, it will quickly be noticed. It is possible to break the build without noticing, for instance by forgetting to add a new file to the repository, and nightly builds in a centralized location will detect these quite quickly.
There are of course other possible benefits, I'm sure others will supply them. :)
Nightly builds are only necessary for significantly large projects (when it takes too long to build it often throughout the day). If you have a small project that does not take long to build you can build it as you get functional pieces of code done so that you know that you did not mess anything up in the procees. However, with larger projects this is not possible so it is important to build the project just so that you know that everything is still in working order
There are several reasons, some will be more applicable than others
If your project is being worked on by two or more people
It's a good way to grab the latest version of code that you aren't working on
A nightly build provides a slice in time of the current state of the code
A nightly build will give you a stable build if you need to send code to people
Nightly builds aren't always necessary - I think they're only really useful on big projects. But if you're on a big project, a nightly build is a good way of checking that everything is working - you can run all your tests (unit tests, integration tests), build all your code - in short, verify that nothing is broken in your project.
If you've got a smaller project your build and test times will be shorter so you can probably afford to do more regular builds.
Nightly builds are ideal for performing static code analysis (see qalab and the projects it collects stats from if you are in java world). Unfortunately, this is something that's rarely done.
Related
I have always worked in environments where developers had to go through a process of working with Network Operations (server guys) to deploy stuff from development/test to production.
I recently started a job where developers can go directly from their machines to production with no middle man. Are there reasons that developers should not be able to do this?
What I have so far:
You are more careful about deploying
something if it has to go through
someone else. As a young programmer
it sometimes took me several tries to
get a working deployment out. Since
the NetOps guys were pissed I learned
to make sure it was right the first
time.
There is some accountability if something goes wrong and more than one person knows what's going on. Boss: "The site just went down!", Everyone else in the office: "Abe just did a deploy, it's his fault!"
When someones sole responsibility is the production server, it's less likely that they will do something stupid.
There will (hopefully) be more information on the deploy and roll back capabilities. Logs, backups that can be rolled back to, automated features...
Are there any other good reasons? Am I just being a control freak?
A few that come to mind (there may be overlap with yours):
A developer can tweak something until it works. This shouldn't be done in Production. If that developer is hit by a bus the next day, nobody will know the system. A documented and repeatable-by-someone-else deployment process helps ensure that such business knowledge is captured.
As a developer, I don't want that kind of access. If something fails, it's far less likely that it's my fault. I'll come in and help, we're all on the same team after all, but I like to know that someone else had to review my work and agree with it. (The same is true of my DB delta scripts. I want a more qualified DBA whose sole responsibility is to the database to review my work. If all they do is run what I tell them when I tell them, then that's essentially no different than giving me direct access. It's just slower.)
Developers often make quick fixes to simple things. We all know that it's often not as cut and dry as the developer thought, and that quick fix either didn't fix it or broke something else. No matter how small the change/fix, there should still be a QA process. (For some shops where uptime isn't so critical that QA process can actually be Production, but that's a rare exception. It shouldn't be that way, from a purist perspective, but as with anything it's a risk/reward ratio. If the risk is low (as in a Production failure doesn't incur much penalty if any at all) and the cost of QA is comparatively high, then it's fine.)
Regulatory needs. PCI compliance, etc. often mandates clear separation of tasks between jobs. This is often misconstrued as "developers can't access production" and treated very black and white. But it does mean that developers should be able to access only what they need in order to do their job. If you don't need production data, and that data is sensitive, you shouldn't have it.
Because many developers are congenitally incapable of thinking they make mistakes - the same reason good dev groups have dedicated test teams.
"I'll just make this small config change in Prod, that won't break anything."
OOP developers should understand separation of responsibilities, I would have thought. You break it, you own it. Avoid the problem with a separate Ops team.
In some environments (e.g. finance) large sums of money (and sometimes the law) are also at risk from ill-advised or ill-intentioned changes in an uncontrolled Production environment.
In small teams, I can see a case for developers having production access, but that has to be controlled and auditable so that you ALWAYS know what is in Production. In that sense, it does not matter who pushes the deploy and rollback buttons, but that they exist and are the only way to change the Production environment.
I for one do not want that to be a large part of my job. You may find that your own devs agree once they see how much more time they can spend coding.
The main reason is because allowing a dev to deploy directly to production cuts out the QA process. Which introduces risk. Which management types don't like.
So another bullet point for you is massive increase in RISK.
Security - By having one gatekeeper (with a backup) only one person is accessing production data and servers. This means fewer access points.
Ease of management - You don't need to create as many accounts in your production environment to keep track of - or even worse, share one account among many. (assuming your prod environment is separated from your dev environment.
Practice makes perfect - one person who builds a routine and sticks to it has fewer chance for screw ups.
If there is a way to make a mistake it will eventually happen. Law of big numbers. It is unreasonable to put the burden on developers to be perfect, if you also want them to be productive.
Change management
Accountability
QA
One button builds / deployment
Unit tests
Code stability - suppose you push, right when someone else just checked in code?
Now, the amount of overhead / difficulty to change should be directly related to your up time requirements. Restated: the more costly downtime is, the more you should invest in preventing downtime.
By deploying directly to the production environment, there is a good chance that no QA was involved (i.e. nothing was tested).
Because there needs to be ONE person you can go to who knows what's deployed on the site. If every developer can deploy, you don't know who deployed what when somebody notices something wrong.
SOC-1 compliance may (unnecessarily) suggest or require that the developer be a separate person than the one deploying to production so that controls are in place to prevent malicious intent.
We're developing a web site. One of the development tools we're using has an alpha release available of its next version which includes a number of features which we really want to use (ie they'd save us from having to implement thousands of lines to do pretty much exactly the same thing anyway).
I've done some initial evaluations on it and I like what I see. The question is, should we start actually using it for real? ie beyond just evaluating it, actually using it for our development and relying on it?
As alpha software, it obviously isn't ready for release yet... but then nor is our own code. It is open source, and we have the skills needed to debug it, so we could in theory actually contribute bug fixes back.
But on the other hand, we don't know what the release schedule for it is (they haven't published one yet), and while I feel okay developing with it, I wouldn't be so sure about using it in production so if it isn't ready before we are then it may delay our own launch.
What do you think? Is it worth taking the risk? Do you have any experiences (good or bad) of similar situations?
[EDIT]
I've deliberately not specified the language we're using or the dev-tool in question in order to keep the scope of the question broad, as I feel it's a question that can apply to pretty much any dev environment.
[EDIT2]
Thank you to Marjan for the very helpful reply. I was hoping for more responses though, so I'm putting a bounty on this.
I've had experience contributing to an open source project once, like you said you hope to contribute. They ignored the patch for one year (they have customers to attend of course, although they don't sell the software but the support). After one year, they rejected the patch with no alternative solution to the problem, and without a sound foundation to do that. It was just out of their scope at that time, I guess.
In your situation I would try to solve one or two of their not-so-high priority, already reported bugs and see how responsive they are, and then decide. Because your success on deadlines will be compromised to theirs. If you have to maintain a copy of their artifacts, that's guaranteed pain.
In short: not only evaluate the product, evaluate the producers.
Regards.
My personal take on this: don't. If they don't come through for you in your time scale, you're stuck and will still have to put in the thousands of lines yourself and probably under a heavy time restriction.
Having said that, there is one way I see you could try and have your cake and eat it too.
If you see a way to abstract it out, that is to insulate your own code from the library's, for example using adapter or facade patterns, then go ahead and use the alpha for development. But determine beforehand what the latest date is according to your release schedule that you should start developing your own thousands of lines version behind the adapter/facade. If the alpha hasn't turned into an RC by then: grin and bear it and develop your own.
It depends.
For opensource environments it depends more on the quality of the release than the label (alpha/beta/stable) it has. I've worked with alpha code that is rock solid compared to alleged production code from another producer.
If you've got the source then you can fix the any bugs, whereas with closed source (usually commercially supported) you could never release production code built with a beta product because it's unsupported by the vendor who has the code, and so you can't fix it.
So in your position I'd be assessing the quality of the alpha version and then deciding if that could go into production.
Of course all of the above doesn't apply to anything even remotely safety critical.
It is just a question of managing risks. In open source, alpha release can mean a lot of different things. You need to be prepared to:
handle API changes;
provide bug fixes and workarounds;
test stability, performance and scalability yourself;
track changes much more closely, and decide whether to adopt then yet;
track the progress they are making and their responsiveness to patches/issues.
You do use continuous integration, do you?
Which best practice involved in software configuration management to help in improving project management?
It mitigates a whole bunch of project risks, including:
The risk of making a change which is found to be incorrect: SCM software allows you to see the change and roll back
The risk that you could lose all your source code (much less likely since everyone has a copy on their machine)
The risk that two people could make incompatible changes: good SCM will allow you to merge the two and get the best of both worlds.
Also, these days SCM is so easy and cheap to set up that embarking on a software project without it is madness.
Assuming you're really focused on best practices, I can outline a couple of possibilities.
Using the best (SCM) tools available. While this might depend on your specific goals and constraints, Mercurial and Git are hard to beat (distributed, excellent branch/merge capabilities, multiplatform, FOSS, really fast, flexible workflow etc.).
You can analyse the data in your source repository using a tool like PanBI (disclaimer: I wrote it). A short screencast shows off what you can learn from repository contents analysis. In brief:
general work dynamics on the codebase
breakdown per developer
daily work dynamics
type of changes to the codebase (add/remove/modify), part of the source tree
...and much more.
Connecting an SCM tool with an issue tracker can also add value. Developers place issue ID's in commit messages, e.g. "[#1455]: improved performance a bit" and the issue tracker relates the issue with the changes in the code repository. From a project management perspective, this allows you to loosely track the time spent on individual issues, project phases or complete projects. A simple commit hook refusing commits without an issue number can go a long way in ensuring data consistency. Such "measured" data can be compared to the baseline to understand what's working and what isn't.
Building official releases on a build server from a tagged source version pulled from the repository could also be considered beneficial from a project management perspective because it's a way to control quality. Building software this way detaches the build process from any dependencies or specifics of developer machine environments, provide reproducibility, allows robust automatic/semiautomatic publishing of the build etc., i.e. streamlines and shields parts of the deployment process.
These are just some of the possibilities, it doesn't stop here.
Later this year I want to release a PHP framework that I've been working on as open source. I do use source control (SVN), but it's on an extremely limited basis. I'm self-taught, I develop by myself and don't have the experience of working with large teams. I have some ideas about what can help make a project successful, but I'm fuzzy on some of the details. Since it's not yet released, I want to do everything I can to set up the right infrastructure from the beginning. What do I need to know in order to setup and manage a successful project?
Some ideas that I have to make it successful (beyond marketing it):
Good documentation and tutorials
Automated unit tests and builds to
push update to the website
A clear roadmap
Bug Tracking integrated with the
source control
A style guide to keep the code
consistent
A forum for the community to get
support, share ideas, etc.
A good example application built with
the framework
A blog to keep the community informed
Maintaining backwards compatibility
wherever possible
Some of my questions:
How do I setup and automate a one
step submit-test-commit-generate API
docs-push update to website process? Edit: Would Ant or Maven be good candidates for this? If so, do you know of any resources for setting up a PHP project using them?
How do I handle (technically)
submissions from other users? How can
I ensure that those submissions must
be approved before being integrated?
What are some of the pitfalls that
can be avoided in terms of the
project community? I'd prefer to have
it be as friendly and helpful as
possible without a lot of drama.
I'd love to learn from your experience on any of these points. If you think I'm missing anything big, please share that as well. Any resources (preferably geared toward a beginner) that you could point me towards would also be greatly appreciated.
I'm just getting started in community projects, but I'll give you some advice on what I know.
How do I setup and automate a one step submit-test-commit-generate API docs-push update to website process?
I've never implemented it as one process. You could just have a checklist, and possibly even create some scripts to do certain tasks. I've never worked with any source control that automates the uploading and such to be done by a script. Most of the time, there is some web interaction to be involved.
You don't want to push API changes until it's an official release.
EDIT: Working Environment
For PHP, most of the time, I either edit directly on the server and test it there, using a beta.example.com, or similar, before pushing to example.com. You could also set up an web environment on your home PC (using XAMPP for Windows, or the standard LAMP installation on Linux). You would probably just use a mirror of your repository here, so you'd do svn commit, or whichever is appropriate for the VCS or DVCS you choose.
The fun part is testing this with different PHP versions. I've not done this myself, but you could probably use a .htaccess file to run a different PHP binary in order to test it out. I'm not really sure what the best option is for this is.
I've not done much with API, as I've never created a library, but just doing a quick search I found http://www.phpdoc.org/. It looks like a mature project, so that might be a starting point.
As far as creating releases go, I generally create a script that only includes the files that are part of the distribution (it will filter out any VCS files, and anything that you don't want in the distributed file). You could write a script around find on linux (which is what I do most of the time), or there may be other better options.
How do I handle (technically) submissions from other users? How can I ensure that those submissions must be approved before being integrated?
This is mostly handled by the bug tracker, and limited access in the Version Control System. Usually, you, and the people you allow, can commit to the VCS. Other users can submit patches, but then you might have someone review the patch, test the patch, and commit. You could split these tasks up as a team, or assign a patch to one person and have them do it all.
What are some of the pitfalls that can be avoided in terms of the project community? I'd prefer to have it be as friendly and helpful as possible without a lot of drama.
I would just make sure to keep it as positive as possible with the project members and community. There's going to be some disagreements, and it will drive a few people away, but as long as you have a stable product that meets the needs of most people, I think that's all that anyone can expect.
One minor suggestion that's worked well for me: start using first-person plural pronouns, rather than singular ones. That is, talk about "we" and "us" rather than "I" and "me." It encourages other people to participate when they feel like part of team, rather than when they feel like they're contributing your own self-aggrandizement.
The most important thing you have to do is to attract users. Without users, you won't get any contributions and developers helping you out. Because developers are users first, and then they decide to extend/fix something they use and might become contributors.
So to get users, you should consider
describe what your framework does in one or two sentences at the top of your project page
mention how your framework can be used and for what, what situations it is most useful for
add a lot of examples on how to use it
mention whether your framework is stable, beta or alpha. That's important because user need to know that before they start using it
also mention whether you want to keep improving it and keep working on it - most users don't want to use a framework that's abandoned (also keep in mind that a lot of users check your commits to see whether you really are working on it - if your last commit to the repository was months ago then you're not really working on it, so cheating isn't possible)
If you got all this, and people start submitting patches, you can use a patch tool to apply those to your source. Depending on your version control system, you can either use the GNU patch, a diff/patch tool that comes with your version control or maybe even a GUI tool that helps you with this. SVN doesn't have a patch tool (yet), but 'svn diff' will create a patchfile which you can then apply with the GNU patch tool, or in case you're using TortoiseSVN, right-drag the patchfile to your working copy and have TortoiseMerge apply it for you.
And on how to best deal with the community:
answer questions in time, don't wait more than two or three days to answer questions
try to be nice, even with upset and angry people. Only if they keep bothering tell them to (still in a nice way if possible) go elsewhere
always keep discussions about the project on a mailing list. You don't want to repeat the same discussions over and over again - if you have a mailing list, just point users to the archives before the discussion starts all over again
And you should watch the talk "How Open Source Projects Survive Poisonous People (And You Can Too)" - it's really good and tells you a lot on how to deal not just with 'poisonous people' but also how to deal with all people involved in your project.
I'd like to add that you should make it as easy as possible for your users to get the whole thing running and modify the code - these 'power users' can be 'converted' into developers or at least people who send smaller patches.
Don't try to do it all yourself - for open source projects there are several hosting providers that solve most of the problems. I recommend codeplex or google code.
Setting up build scripts will depend a certain amount on what platform you set up, but in general it's easy to add any tool you want into the script once you start using any sort of build script.
If you really need the one step process you describe, you need a build server. I use TeamCity, which I have set up to watch for any changes in svn and trigger build/test whenever something is checked in. The build server will generally be able to perform any steps that you put into the build script.
Read up on Git as an alternative to SVN
free public repository/bug tracker/wiki/fork-enabled community in Github (which hosts symfony and PHPUnit amongst others)
"How do I handle (technically) submissions from other users? How can I ensure that those submissions must be approved before being integrated?" - with Git, pull what you/your closest team finds most interesting to the master branch
Consistent API
be inspired of other public API:s
only change in major versions
guessable
Interesting for both users & developers
clear goal (your roadmap - excellent)
useful, contra everything else available
easy to use, but still not easy-enough-to-write/maintain-yourself
You could check out either Ant or Phing to build your project. Include CodeSniffer in the build and you'll save time checking for basic formatting errors/differences.
These are all technical tips, about the soft part... treat humans with respect, a lot of interest and be overly excited about their contributions and make them feel that they're not wasting their time. That would appeal to me.
Take a look at Karl Fogel's book on Producing Open Source Software. It probably has everything that you asked.
You should also plan for engaging the community. I'd recommend reading Jono Bacon's The Art of Community [http://www.artofcommunityonline.org/].
You have a great set of ideas to start. You might have to start by trimming them down! Ask yourself what's necessary for a first release.
For automating the builds and tests, the scripting can be done with ant, maven or phing for PHP projects.
You'll probably need a host so you can demo the product. For PHP that is pretty easy to find.
You need an open source hosting provider-- especially github (but also google code, source forge, etc). Github provides bug tracking, default licenses, blog and great mechanisms for accepting changes from the community. Built on git, it facilitates distributed projects quite well.
Although it's good to have a one-step build and install in place, automating integration of others changes probably isn't important (or desirable) off the bat.
Good luck!
Context: I work at a small software company that has traditionally done research-type work, and does not have much experience in the commercial space. We are now trying to push into the commercial world. Due to our origins in research we are used to a very rapid development cycle and very little structure in terms of maintaining proper versions of projects.
Problem: The lack of structure is now proving to be somewhat of a hindrance, as every developer has a slightly different view of the code base. A problem one developer discovers is not reproducible by another developer, and problems found in one build may disappear in the next (or worse, new problems may appear). This makes for a very frustrating experience for someone who is responsible for integrating all the projects and ensuring quality and performance standards are met - i.e. myself.
Potential solution: Personally I am convinced we need to enforce better structure via fixed version numbers and regular releases. It should be self-evident how proper versioning would help with many of our problems, but of course it is not without problems - developers need to do extra work to perform and test releases, and will no longer be able to use the latest versions of everything.
Question: To come to a point - what sorts of strategies do you recommend for ensuring the process and effort required for releases occurs as smoothly as possible? We are using git for version control, maven for our build system, and we have bug tracking and continuous integration systems running, so I believe the tools are there. I am simply unsure about what a proper release process should look like.
You have the big three in place: version control, one-click build via Maven and your continuous build server, and bug tracking. It sounds like you guys are gravitating towards Agile methodologies, and so you ought to be trying to keep the trunk version of your product in a near deliverable state at all times.
When you decide to make your first release, create a branch off of your trunk version for that release. Decide on a labelling scheme and be sure to label the branch version. For example, your first release could be 1.0.4530, where the 1 means first version, the 0 means it's the first release candidate, and the 4530 is the version control change number. You test this release branch and fix important bugs on it. After a while you issue another release candidate, say 1.1.4807. This process iterates a couple more times (say), your release becomes good enough, and you ship version 1.3.5167.
Meanwhile, your new development occurs only in the trunk version, and from time to time you'll need to merge bug fixes from the 1.x release branch back to the trunk. Later, you'll split off a 2.x branch from the trunk to repeat the process for your second release. You'll generally have several active branches (plus the trunk), with development limited to the trunk and each branch kept pristine and independent from development.
You guys will get the hang of things and your developer coordination problems will become less frequent. But these problems are nearly all going to be limited to the trunk, not the release branches.
A problem one developer discovers is
not reproducible by another developer,
and problems found in one build may
disappear in the next (or worse, new
problems may appear). This makes for a
very frustrating experience for
someone who is responsible for
integrating all the projects and
ensuring quality and performance
standards are met - i.e. myself.
Potential solution: Personally I am
convinced we need to enforce better
structure via fixed version numbers
and regular releases.
I don't think you need to have very frequent releases just to coordinate internally. You can do that through version control. Just have people talk about specific git revisions when reporting issues. Also note that you will have to coordinate any external dependencies/libraries too. Some kind of vendor branches could help with this.
It sound like the developers need to use "test branches" and respect the "stable/production branch" a little bit more.
Sell in the concept of "do your wild west stuff in this branch", and when you are happy with the results then you merge it into this "boring stable production branch"....
(or something like that)
There are books written about the general topic; Amazon search even returns three titles for specialized "version control with git."
I think you will benefit from defining a canonical view of the code base. Call it Test. A problem is a problem if it appears in Test. If a problem does not appear in some developer's view, it is up to that developer to figure out what is the important difference; and likewise for a problem that appears in a developer's view, but not in Test.
One convention is for Test to be re-built from sources on a nightly basis. A more strenuous convention is for Test to be re-built upon every update. If your team is small (five or fewer) and not dispersed over great distances or multiple timezones, a reasonable first approximation is to make Test a git workspace on a server upon which your toolchain has been installed along with some cron jobs so that this workspace is updated and rebuilt every night (usually).