Bazel is here to stay. Sooo many companies are jumping on this bandwagon [1].

Bazel is better than Maven, Pants... you name it.

Think of this like Protobuf or Kubernetes. It's an open source tool used to build things at scale. There will be lots of users and contributors. A cottage industry will spring up, and this will grow well beyond Google.

[1] https://github.com/bazelbuild/bazel/wiki/Bazel-Users

If only Android team cared to use it instead of Gradle.

Android is probably the only platform where almost at every conference there is a regular talk on how to improve build times.

Not even C++ conferences talk so much about build times.

I find gradle is something like doing thing diligently what should not be done in first place. It seems clear though Oracle/OpenJDK have no interest in putting some fucking options in `javac` itself to compile project/module or even package. Users must wade through maven/gradle crap every time even to compile pure Java projects.

No idea what you're talking about here.

I used to compile Java projects since 1996 until 2004 pretty alright.

And Maven runs circles around Gradle's performance, without requiring a background daemon sucking up at least 2GB, minimum.

Gradle is the only reason Groovy is still somehow relevant and it has to thank Google's Android team for it.

As for the Oracle/OpenJDK remark, it is an open source project, under the GPL with Classpath exception license, it is up to the ones that keep saying that Oracle doesn't do a good job to contribute.

The cottage industry is already springing up. There are a few firms offering Bazel consulting and training. https://bazel.build/experts.html

There is also a considerable work done by Angular team to integrate it on front end builds.

It’s OSS software not a SaaS.

If people use it there will be no deprecation date. You can just fork and continue.

Bazel is too complex to make this practical for most people. It would be like forking Go.

However, it's unlikely to be abandoned.

I’m sorry if that sounds harsh, but that’s at best a naive view. In practice you cannot just fork the codebase and continue maintaining it, that has a high cost that a lot of organization cannot afford. It’s technically true that you can fork the project but in practice forking something that big doesn’t really happen without a big player leading the fork (or some motivated people with enough time and money to dedicate to the fork, which is possible but unlikely).

Hence the “if people use it” qualifier.

Companies relying on Bazel would step in and support it. We’ve seen this play out plenty of times before, e.g. Hudson to Jenkins or OpenOffice to LibreOffice, where a popular OSS product loses a corporate sponsor and continues.

You’re ignoring all those times where that didn’t actually happen. It seems quite risky for an organization to rely on the idea “you can just fork and continue the development yourself in case of issues”. That’s not what happens in practice, taking over the development of something like bazel has a non-negligeable cost. Now for sure you can find some (cherry-picked IMHO) success stories but that’s the exception as far as I can tell.