How do I get the commands executed by Bazel answers a similar question for bazel build, and --subcommands works well for that. But --subcommands does not show the command that's executed for bazel run.
It's just running the executable built by the target you give it, plus anything else that isn't a bazel flag. Determining what's a bazel flag follows common semantics: anything that doesn't start with - anywhere, and anything at all after --. Some examples:
bazel run //foo:bar runs bar.
bazel run --subcommands //foo:bar and bazel run //foo:bar --subcommands are both just running bar. There are no flags after -- because there is no --.
bazel run //foo:bar xyz runs bar xyz, because xyz doesn't start with -. bazel run //foo:bar -- xyz also does the same thing
bazel run //foo:bar -- --subcommands runs bar --subcommands, without bazel interpreting --subcommands.
You can also use --script_path to output a shell script with the full command and environment setup written out.
Related
given a bazel command, for example bazel build //..., is it possible to see what build command this expands to based on the bazelrc(s)?
bazelrcs are great to set default flags, like the default platform to build for setting some variables based on profiles.
They also allow for nested bazelrcs, that you can import and try-import leading to a rather complex expansion of bazel commands via bazelrcs.
To verify that the bazelrcs did indeed configure the correct things and did not override some flags accidentally, I would love to see the output of what bazel would expand a bazel command to.
Is that possible?
I tried setting --logging=0 and --logging=6 with bazel 5.3.0 but see no such output.
You can run your build using the --explain flag.
Something like
bazel build --explain explain.log //...
should produce a file explain.log where the first line contains the expanded flags.
From my tests, it doesn't appear that bazel run runs in a sandbox. For example, I ran bazel run //:some_target --spawn_strategy=darwin-sandbox --sandbox_debug, and it didn't generate a new directory in <outputBase>/sandbox/darwin-sandbox.
Am I correct that bazel run doesn't use sandboxing? If so, why not?
bazel run will build the target you pass with sandboxing and caching the same way as bazel build, and then it will run it outside the sandbox. That's the whole point of bazel run.
If you want to run a command inside the sandbox, write a genrule and then bazel build the genrule target.
I'm wondering if there's a way to run an action or script or program after a bazel run command finishes. Running such a script/program before the bazel run is (in a way) possible via --run_under.
Here's an example use-case:
bazel run //path/to/my:target
Running target generates an output file out in it's cwd (somewhere in the bazel .cache directory)
Run a script that uses the output file out
If it's not possible to get run that script afterwards, is there a way to get the cwd of the binary that was just run? Then it would also be possible to access out from "outside" without bazel.
--run_under is the correct approach to this. You can simply run the command in the script after running what is passed as args to your script that you pass to --run_under e.g.
#!/bin/bash
set -euo pipefail
./prefix_command
bash "$#"
./postfix_command
I have added an annotation processor as a java_plugin and have added this into the plugins section of my java_library rule. I was wondering what are the bazel options to step through the annotation processor code and the javac compiler's code?
One way to do this is to run bazel build with --subcommands. Bazel will then print out all the commands it executes during a build. You can then find the javac invocation you're interested in, copy the command line (including the cd part so you're in the correct directory), modify the command line to include the debugging options, and run it manually. Then you can debug it like you would any java program.
One thing to note is that bazel will print only the commands that it actually runs in that build, so if the action you're interested in is already up-to-date, you may have to delete one of its outputs (e.g. the jar output of that library) to get bazel to re-run the action.
running bazel build //... \
--aspects some-aspect.bzl%some_aspect \
--output_groups=some_new_output,default
does not create test jars outputs.
on the other hand running bazel test does create the test jar outputs:
bazel test //... \
--aspects some-aspect.bzl%some_aspect \
--output_groups=some_new_output,default
How come?
This question was updated to reflect use of aspects:
The original question:
running bazel build //... does not add test code to output jar.
on the other hand bazel test //... builds the test code but also
runs it.
Is there a way in bazel to build the test code without running the
tests?
I had a mistake in the values I gave the --output_groups flag.
It should have been --output_groups=+some_new_output,+default
The default can even be omitted:
--output_groups=+some_new_output
This flag is not documented at all. There is an open issue on this in bazel github repo.
You may be looking for --build_tests_only.