usage.csv escapes only limited characters

[Eric Peterson]

horizon / openstack_dashboard / dashboards / admin / overview / templates / overview / usage.csv

Has escapes in there for limited CSV characters. However, there are still some characters (like line breaks) that can cause this template to not render csv data correctly.

Suggestion:
Using the native python csv library (http://docs.python.org/2/library/csv.html) and a more generic template like this:
{% autoescape off %}{{ csv }}{% endautoescape %}

Seems to work with a much larger set of csv data combinations.

[Gabriel Hurley (gabriel-hurley)]

I have to admit, I personally never touch the CSV feature in Horizon so I tend not to think about it much. That's a very good suggestion though. We'll definitely get that in H. Thanks!

[Lin Hua Cheng (lin-hua-cheng)]

As part of this change, we should also look into StreamingHttpResponse added in Django 1.5. This is useful for generating large CSV file.

https://docs.djangoproject.com/en/1.5/ref/request-response/#django.http.StreamingHttpResponse

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/29443

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/29443
Committed: http://github.com/openstack/horizon/commit/a4e583cffbd2dbe2e431c306c389a0b0815073e1
Submitter: Jenkins
Branch: master

commit a4e583cffbd2dbe2e431c306c389a0b0815073e1
Author: Tihomir Trifonov <email address hidden>
Date: Wed May 22 11:02:04 2013 +0300

    Improvements in csv export for usage data

    Added a csv writer using the 'csv' library to format
    properly exported data - escaping, encoding etc.
    Added a HttpResponse-based class to handle csv generation
    Added translation for the CSV columns and template.
    Improved consistency for exported data - now passing project name
    instead of project id for csv export. Also added both project name/id
    in the header of a project usage export.

    Fix bug #1158383

    Renamed few occurencies of 'tenant' to project.

    Also added a new 'project' in nova_data.py, which required
    some refactoring of few tests, that didn't consider the current project
    for project-based calls.

    Note: I've added a StreamingHttpResponse example,
    which is introduced in Django 1.5+ and being advised in the ticket.
    However, my opinion is that at the moment we don't need this - it is
    too complicated for the current usage.

    Change-Id: Ic00626b273921fa5c6c89704b3a9e08b252aaeb0

[Julie Pichon (jpichon)]

Reopening as the commit was reverted in ef26414b8d77827dc4d174acfe11699544a397e4 after a few regressions were discovered.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/33947

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/33947
Committed: http://github.com/openstack/horizon/commit/0c0153a037594c6c98025f698e8593a01ed0ea2d
Submitter: Jenkins
Branch: master

commit 0c0153a037594c6c98025f698e8593a01ed0ea2d
Author: Tihomir Trifonov <email address hidden>
Date: Fri Jun 21 13:48:43 2013 +0300

    Improvements in csv export for usage data

    Added a csv writer using the 'csv' library to format
    properly exported data - escaping, encoding etc.
    Added a HttpResponse-based class to handle csv generation
    Added translation for the CSV columns and template.

    Fix bug #1158383

    Renamed few occurencies of 'tenant' to 'project'.

    Also added a new 'project' in nova_data.py, which required
    some refactoring of few tests, that didn't consider the current project
    for project-based calls.

    Note: I've added a StreamingHttpResponse example,
    which is introduced in Django 1.5+ and being advised in the ticket.
    However, my opinion is that at the moment we don't need this - it is
    too complicated for the current usage.

    Change-Id: I18bd70a23b7b8389c7c0f96dbf8794fea5e1e75b