Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Develop into central-develop #3156

Merged
merged 322 commits into from
Mar 9, 2015
Merged

Develop into central-develop #3156

merged 322 commits into from
Mar 9, 2015

Conversation

jamalex
Copy link
Member

@jamalex jamalex commented Feb 28, 2015

No description provided.

mrpau-dev and others added 30 commits February 11, 2015 16:42
…ads list of screenshots to take from `screenshots.json` file.

Added `/data/screenshots` folder and `ghostdriver.log` to `.gitignore` file.
…he form, take screenshots between inputs.

Make re-usable the reset and delete sqlite database methods.
More efficient selenium browsing code by checking if browser's current url is already the same as the `start_url` or if the last user is the same user for the screenshot so no login is needed.
Change keys on .json for consistency and make these keys constants.
Specify field names on the `actions` array - "Explicit is better than implicit."
Needs to be refactored. Probably good to decouple from
TestCase framework. Are there other unnecessary parts that
could affect performance? Don't want to slow down sphinx
build process.
Add two command line options, --output-dir which
lets you specify an output directory for the screenshots
relative to the project root url, and --from-str so that
we can just pass a JSON string to the command (instead of
reading from a file).
Also added the ability to specify verbosity, namely
now -v 0 gives no output (except for some mysterious
TestCase stuff???).
Entered some sort of import hell...
At least it's possible to invoke the management command
by an os.system call, but it's not pretty.
And some other minor changes:
* Handle empty arg_str for _parse_nav_steps
* In _parse_nav_steps change return type of `args` keyword
in order to conform with the invocation in `run` method
* Removed _commands_handler as it was rendundant and folded
the case of multiple commands into _command_handler
In other words, there's no difference between 1 and many commands now
* Prettify _command_handler and _login_handler by moving some shared
info into instance variables set in `run`
* Added some implementation-specific helper functions to the end of
the file

Is this file getting too cluttered?
screenshot.py wants to `from selenium import` something,
so add ka-lite/python-packages to conf.py (for `make` commands)
and to the test script (so tests run without import errors).
This is dependent on the current directory structure.
The screenshots are collected in a build environment variable
and executed as a batch at a certain point in the build process
This will significantly decrease the build time.
Prevents a potential error for multiple shots if one fails
and the browser remains open.
rtibbles and others added 21 commits March 3, 2015 20:36
…lates_script

Remove the css classes mistakenly translated.
…ng-is-missing-left-border

fix for #2980 - Answer box highlighting is missing left border.
Returns keypad to its former stylistic glory
fixed points not shown when logged in as learner. note: html id cannot c...
…able-to-select-one-input

Software keypad only able to select one input
docs - screenshot directive (and management command) can take screenshot as though the device was registered
@jamalex
Copy link
Member Author

jamalex commented Mar 9, 2015

This seems like an appropriate place to ask what the advantage is to having a separate develop and central-develop branch.

In an ideal world, these could be the same. But we've often needed to do extra stuff on the central-* branches that shouldn't (yet/ever) be in the master/develop branches. For example, the nalanda branch added some models (e.g. AttemptLog) that weren't part of 0.12, but since the nalanda code was being used in production, and needing to sync against the central server, the nalanda branch needed to be merged to central-master (e.g. see master...central-master)

Longer-term, we want to have:

  • all projects/partnerships to be working on the mainline release cycle, which would reduce the need for this
  • every large project to have its "own" central server, if it needs to have project-specific stuff enabled (this may only be feasible after we have Zone-specific databases/certificates)

To what extent does the central server depend on the implementation details of the distributed server (including our own packages in python-packages like securesync)

Not sure whether this is exactly what you're asking, but yeah, the stuff inside ka-lite is used extensively within ka-lite-central, in particular:

  • securesync itself
  • all the SyncedModels
  • the coach reports
  • the user management stuff

MCGallaspy and others added 5 commits March 9, 2015 14:22
…h-reports

Move `subtitle` css styles to be used by all coach reports templates instead of just the tabular report.
Add workflow for contributing to docs
jamalex added a commit that referenced this pull request Mar 9, 2015
@jamalex jamalex merged commit e6f34b2 into central-develop Mar 9, 2015
@jamalex jamalex removed the has PR label Mar 9, 2015
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.