Need to remove the user's "all users" group read permission during unsetup.
This shortcoming became apparent during testing when "is_invited" flag is checked.
Added integration tests for create new user, setup and unsetup existing user. The tests use selenium driver without headless. Hence, we see the browser while these tests are running.
Next step is to use headless to hide the browser during test execution.
Brett Smith [Wed, 2 Apr 2014 20:55:24 +0000 (16:55 -0400)]
docker: Install staged source on Docker images.
A separate build step to stage source makes it easier for us to fix
permissions and do other clean-up before installing it in a Docker
image. It also reduces code redundancy; for example, builders don't
have to worry about excluding log and tmp directories.
Brett Smith [Mon, 31 Mar 2014 13:46:50 +0000 (09:46 -0400)]
doc: tutorial-job1 updates and style tweaks.
The major change here is that we set no_reuse on the job, to avoid
confusing people by using results from previous tutorial runs. These
updates also reflect some minor changes in the Workbench UI.
Brett Smith [Fri, 28 Mar 2014 19:47:52 +0000 (15:47 -0400)]
doc: Make running-pipeline-command up-to-date.
* @arv pipeline run@ now requires @--run-here@.
* Pipelines are now under Activity in Workbench navigation.
* The format of the example output has changed, along with all the
Keep hashes.
Brett Smith [Fri, 28 Mar 2014 14:31:59 +0000 (10:31 -0400)]
doc: Highlight "you" in reference JSON.
Since this reference provides users with raw example JSON, and doesn't
tell them to write it with @cat@, we can't rely on shell variable
expansion here.
Brett Smith [Mon, 31 Mar 2014 13:51:42 +0000 (09:51 -0400)]
doc: Use $USER in tutorial JSON.
The tutorial encourages users to copy and paste command text.
However, most of the JSON needs to have a user-specific value for
"repository", which is easy to overlook. Since we're telling users to
save JSON with @cat@, we can take advantage of shell variable
expansion to make the right thing happen automatically.
I added accompanying notes to help explain what's going on for people
who aren't copying instructions so literally.
Tom Clegg [Tue, 1 Apr 2014 22:39:12 +0000 (18:39 -0400)]
Add a system_group, and ensure it has can_manage permission on every
existing and future user.
This lets us give a user access to all users' data (on read-only or
read+write basis) without also giving them system level privileges
with the is_admin flag (which would also let them override business
logic like UUID assignment).
Tom Clegg [Tue, 1 Apr 2014 19:16:18 +0000 (15:16 -0400)]
Support group-wide administrator privileges.
If an admin user or group has permission to read/write/manage a target
user, the admin's permission extends to all other objects on which the
target user has permission.
Brett Smith [Mon, 31 Mar 2014 13:46:50 +0000 (09:46 -0400)]
doc: tutorial-job1 updates and style tweaks.
The major change here is that we set no_reuse on the job, to avoid
confusing people by using results from previous tutorial runs. These
updates also reflect some minor changes in the Workbench UI.
Brett Smith [Fri, 28 Mar 2014 19:47:52 +0000 (15:47 -0400)]
doc: Make running-pipeline-command up-to-date.
* @arv pipeline run@ now requires @--run-here@.
* Pipelines are now under Activity in Workbench navigation.
* The format of the example output has changed, along with all the
Keep hashes.
Brett Smith [Fri, 28 Mar 2014 14:31:59 +0000 (10:31 -0400)]
doc: Highlight "you" in reference JSON.
Since this reference provides users with raw example JSON, and doesn't
tell them to write it with @cat@, we can't rely on shell variable
expansion here.
Brett Smith [Mon, 31 Mar 2014 13:51:42 +0000 (09:51 -0400)]
doc: Use $USER in tutorial JSON.
The tutorial encourages users to copy and paste command text.
However, most of the JSON needs to have a user-specific value for
"repository", which is easy to overlook. Since we're telling users to
save JSON with @cat@, we can take advantage of shell variable
expansion to make the right thing happen automatically.
I added accompanying notes to help explain what's going on for people
who aren't copying instructions so literally.