[bmon]

I can't agree with having a source file include the tag. It is an eternal source of merge conflicts and pain, unless you take steps to automate it. And in that case, does the file add much value anymore?

My personal experience with go modules and versioning has been really positive. In that ecosystem - you only define your major version in the mod file and rely on the VCS for everything else.

[WorldMaker]

Yeah, I often prefer, when possible, to automate in CI dumping the git describe output to a .gitignored source file and letting tags themselves be the "source of truth".

Related to that "tags as source of truth" is using tags as a deployment trigger. A release manager applying a tag can be a signal or gate for a version to go to later environments. (For instance: CI builds from main branches stop at Dev environments, CI builds triggered from new tags automatically move on to UAT and Staging environments.)

Also, another tip I've found useful for people with more "monorepos": tag name restrictions match branch name restrictions and you can use a "folder structure" of tags. You can name tags things like subcomponent/v1.0.2. Some git UIs even present such tags as a folder structure. Doing that can confuse git describe, of course, so finding an arrangement that works for your project may be a balancing act. I've used lightweight tags for subcomponents so that git describe doesn't "see them" by default and then you can use the git describe --tags that also takes lightweight tags into account if you need a "subcomponent version" for subcomponent tag triggered deployments (and then you just need to remove to remove the folder prefix).

[aarchi]

> you can use a "folder structure" of tags. You can name tags things like subcomponent/v1.0.2. […] Doing that can confuse git describe

Using the --match option, `git describe --match='subcomponent/*'` fixes this problem. It filters the tags that are considered to only those matching the pattern, so that a later tag for another subcomponent will not be used.

[zbuf]

Yes, having a version number centralised in source code is exactly the thing Git helped us many times to avoid.

Also ties in with the author's recommendation to begin tags with "v". My experience is that excluding it is better. Then a simple "git describe" readily gives the version number in scripts with no sed or reprocessing.

I've seen many conventions with Git. It's interesting to hear some rationale, but a stretch to describe these suggestions as the "proper" way.

[CameronNemo]

The v prefix really helps us only run ci on version tags instead of random tags as well. If you are strict and always use semantic versions for tags, you can just keep doing what you are doing. But if you ever want to create some random tag later and don't want your automation to try to use it as a version number, the v prefix helps.

[jonnycomputer]

How is the version information included in the software if you don't include it in the source? Do you have a deploy script that modifies source based on the git tag, or ?

[WorldMaker]

I sometimes dump git describe to a JSON file rather than modifying a source file and let the build bundle it as an embedded resource. You can .gitignore the JSON file to keep it from accidentally getting checked in (and causing merge conflicts).

As also pointed out, many build tools that want or need version numbers often also have command line flags or can take environment variables instead of using source files.

[jonnycomputer]

There seems a lot to like about that approach.

[t0astbread]

Some languages and build systems also allow you to set constants from compiler flags (Go for example). Other systems make the entire build configuration an executable program (Gradle).

[jonnycomputer]

I did not know that. I mostly use interpreted languages (Python, JavaScript, etc.)

[formerly_proven]

For Python there are tools like setuptools_scm to completely automate the setup of this.

[jonnycomputer]

Good to know.

[zbuf]

The build process uses the tag.

In practice that's a tiny bit more complicated to do it really well, with the version as a dependency in the build process. When developing, you don't want to trigger a full rebuild if the version number changes, but you do have rebuilds to do when it does.

[da-x]

I don't think that it poses an issue with merge conflicts. It is likely not an issue as long as the tagging is only made on the main branch by the release managers.

I know it is working well for Linux kernel hackers. Linus does all the tagging, and I don't recall issues with merge conflicts on the top Makefile with the part that holds the version.

[zbuf]

I think you're assuming that all software results in a single "release manager".

In many environments, multiple branches of software are deployed and maintained at the same time.

Even your example with Linux; I don't think that is correct -- Linus doesn't tag the stable branches, and others. There is a centralised agreement of how the version numbers are maintained though.

[da-x]

My example is relevant to tagging in branches that may merge back into the main one. The stable branches in Linux are cherry-pick branches that don't get merged back and therefore creating tags there is harmless.

[howinteresting]

How does it work if a single repository contains more than one independently usable library?

[WorldMaker]

Tags in git allow just about the same naming options as branches, and "v" prefix is just a convention not a requirement. So rather than applying a tag v1.3.2 you could use a "folder structure" such as library1/v1.3.2 and library2/v3.1.2. Today I learned that you can use `git describe --match="library1/v*"` to get the version relative to just "library1" if you are versioning this way (in a monorepo, for instance).

[howinteresting]

It seems like what you describe is what golang supports: the folder structure in the tag name must match the folder structure in the repository. That's pretty cool.

[wnoise]

Well, if they're actually independent, clearly those shouldn't be in the same repository...

[t0astbread]

"Independently usable" does not have to mean "totally independent". The libraries could be related to the same product using the same infrastructure.

[howinteresting]

That doesn't follow. Why should they not be in the same repository? They may be related but independently usable.

[tricky777]

mono repo approach (or something to that direction)

[jrochkind1]

Do people use a mono repo approach with different parts of the repo having different version numbers? The way I thought of it, the whole point of mono repo is not that.

[howinteresting]

Sometimes, yes. Monorepos use monolithic versioning internally, but they may use semantic versioning when publishing libraries.