ctx.actions.write generates executable files even when asked not to

[endobson]

Code:
https://gist.github.com/endobson/1fa9fe0ad27eeaa8296c17187165d642

  ctx.file_action(ctx.outputs.object, "object", False)

Output:

endobson@yggdrasil () ~/tmp/bazel-bug4 % ls -l bazel-bin/demo_rule.o
-r-xr-xr-x  1 endobson  wheel  6 Apr 25 10:03 bazel-bin/demo_rule.o

Using self built release 0.4.5:
Build label: 0.4.5- (@non-git)
Build target: bazel-out/local-opt/bin/src/main/java/com/google/devtools/build/lib/bazel/BazelServer_deploy.jar
Build time: Tue Apr 25 06:47:27 2017 (1493102847)
Build timestamp: 1493102847
Build timestamp as int: 1493102847

[Profpatsch]

See #6530

[endobson]

Even if this is desired behavior, then the docs should probably be fixed.

https://docs.bazel.build/versions/master/skylark/lib/actions.html#write

is_executable | bool
Whether the output file should be executable.

If this bit is not controlling the execute bit on the file, what is it controlling?

[EricCousineau-TRI]

Relates #3359 (still)?

[brandjon]

Confirmed this is still the case.

I'm confused why we have these attributes at all if we always make outputs executable. This is both a documentation issue and a design issue.

[sgowroji]

Hi there! We're doing a clean up of old issues and will be closing this one. Please reopen (or ping me to reopen) if you’d like to discuss anything further. We’ll respond as soon as we have the bandwidth/resources to do so.

[endobson]

Can you explain why this is being closed? Your description makes it sound like that this is an old issue and 'stale' and those are the reasons for being closed. Those sound like a bad reasons, things don't stop becoming defects just because they are old.

#2888 (comment) showed that it reproduced still ~4 years after I initially filed it.

The open/closed bit should represent whether or not the issue is still impacting users/is a defect not whether or not there is bandwidth to fix the defect.

[georgegiovanni]

I agree with endobson.