The HBAC details page has been enhanced to support Undo and Reset operations.
The functionality is implemented in the base widget class so the behavior
will be more consistent across widgets. A <span> tag now used to define the
field boundary in the HTML doc. The tag contains the visual representation
of the field which include the input tag and optionally the undo link.
The Update method on HBAC details page has been modified so that it executes
several operations using a batch command. The operations being executed
depends on the changes made to the fields. These operations may include:
- removing access time if access time is changed to any time
- removing memberships if member category is changed to all
- modifying rule attributes if description or rule type is changed
- enabling/disabling the rule if rule status is changed
The behavior of the Add & Remove buttons also has been changed such that
it adjust the category attribute properly in addition to adding the
memberships using batch command. For example, if category is initially
set to all, adding a new member will also change the category to empty.
The ipa_command have been modified to store the on_success and on_error
handlers as properties. When the command is executed as a part of batch
operation, the result of each command will be passed to the appropriate
handler.
The unit tests and test data have been updated as well.
The HBAC Service search and details pages have been added under the HBAC
tab. This requires some changes to the framework.
Currently the navigation framework doesn't support multiple entities under
one tab. As a temporary solution, an 'entity' URL parameter is used to
determine the entity to be displayed. This parameter is now only used by
HBAC tab, but its use might be expanded later. The navigation framework
needs be redesigned to provide more flexibility.
The search page in all entities except DNS records have been changed to
use the ipa_search_widget. The Select/Unselect All checbox and Delete
button now work correctly and consistently.
The Add dialog has been enhanced to render and work in a more consistent
way while still supporting custom widgets & layouts. For the search page,
the Add button will refresh the search results and clear the fields in
the dialog box.
The framework now provides some extension points which can be overriden
by the subclasses:
- init(): for initialization and configuration
- create(): for creating the layout dynamically or from template
- setup(): for setting the look and feel
- load(): for loading the data
Entity and facet initialization is now done after IPA.init(). This is to
ensure the metadata is loaded first so the entities and facets can use
localized messages/labels/titles.
The group entity has been partially converted to use the new framework.
The unit tests have been updated accordingly.
IPA commands now can be defined in these classes:
- ipa_command: a single IPA command
- ipa_batch_command: a batch command for executing multiple commands
on the server side using the new batch plugin
The dialog boxes for adding and removing entries have been refactored:
- ipa_dialog: base class for dialog boxes
- ipa_adder_dialog: generic adder dialog box
- ipa_deleter_dialog: generic deleter dialog box
- ipa_association_adder_dialog: adding entity association
- ipa_association_deleter_dialog: removing entity association
Dialog boxes for adding/deleting HBAC users, hosts, services, and
sourcehosts are implemented using the association dialog boxes.
The dialog box for adding access time is implemented using ipa_dialog
and currently contains only a text field. This will be replaced with a
custom dialog box in a separate patch.
The dialog box for removing access time is implemented using the
generic deleter class because it's not an association. Removing multiple
access times is implemented using batch operations.
New test data files for access times have been added.
THis patch handles Kerberos ticket expiration in the UI. Additionally it removes the mod_atuh_kerb authorization for elements in the static directory, cutting down on the number of round trips required for initializing the web app
Conflicts:
install/static/ipa.js
The UI framework has been extended to include a collection of widgets:
- ipa_widget: base class
- ipa_text_widget: text field
- ipa_radio_widget: radio button
- ipa_textarea_widget: textarea
- ipa_button_widget: button
- ipa_column_widget: column for table
- ipa_table_widget: table
These widgets can be used to create input controls. They can also be
extended to create custom controls.
The framework has also been enhanced to support custom layouts. This
can be used to change the look of the application without changing
the code. Initially this is only available in details section.
Layout consists of a collection of HTML templates. Each template is a
complete and valid HTML file representing a portion of a page. The
template will be loaded and initialized by the code, then filled with
the data from the server. The layouts are located in
install/static/layouts/<name> folder.
By default, if no templates are used, the fields in the details page
are rendered vertically using dd/dt/dd tags. For pages that require
different layout, a custom UI needs to be developed. There are two ways
to do that:
- write a custom widget to generate the UI dynamically
- create an HTML template and write the initialization code
For components that are quite complex or used frequently, it's might
be better to use the first method. For simple pages that are used only
in one location or need to support customization, the second method
might be preferable. Other benefits of templates:
- cleaner code and UI separation
- more flexibility in customization
- new pages can be developed quickly and require less coding
- multiple templates can be used with the same initialization code
- easier to maintain
The HBAC details page has been implemented using both methods. By
default it will use custom widgets to generate the page. To use a
custom layout, add the following parameter to the URL, then reload
the page:
&layout=<name>
Currently the only available layout is 'default' which produces the
same look as the custom widgets.
The HBAC details page is usable, but it still needs additional work.
The access time is not working yet. There is no undo button, hint,
or validation yet.
The table in the association facet has also been changed to use
ipa_association_widget which is derived from ipa_table_widget.
The Makefile has been updated to include the layouts. The unit tests
have been updated as well.
The ipa_cmd() has been modified to identity the type of the error
it has received and display the error using the right dialog box.
The dialog box can be customized further to display the appropriate
amount of information for each type of error.
This patch introduces a new framework for implementing custom UI.
It consists of the following classes:
Main:
- IPA: global namespace and object repository
- ipa_entity: base class for entities
- ipa_facet: base class for facets
Add dialog:
- ipa_add_dialog: default add dialog
- ipa_add_field: the fields used in the dialog
Search facet:
- ipa_search_facet: default search facet
- ipa_search_column: the columns in the search result
Details facet:
- ipa_details_facet: default details facet
- ipa_details_section: the sections in the details facet
- ipa_details_field: the fields in the details facet
Association facet:
- ipa_association_facet: default association facet
- ipa_association_config: the association configurations
To use this framework, create a class extending the ipa_entity (e.g.
ipa_hbac). Use the create_* methods to create add dialog, search facet,
details facet, and association facet. The fields/columns for the dialog
and facets can be specified using the init() function. Custom UI can be
defined by overwriting the base methods (e.g. setup, save, load).
The entity must be added into the repository using IPA.add_entity().
The original ipa_entity_setup() has been generalized by moving facet-
specific codes into the corresponding facet. Some facet names are still
hard-coded. This will be fixed in follow-up patches.
Some global variables have been removed because their function has been
replaced by the object repository:
- ipa_entity_add_list
- ipa_entity_search_list
- ipa_entity_details_list
- window_hash_cache
Some functions and variables have been moved into IPA namespace:
- ipa_json_url -> IPA.json_url
- ipa_use_static_files -> IPA.use_static_files
- ipa_ajax_options -> IPA.ajax_options
- ipa_objs -> IPA.metadata
- ipa_messages -> IPA.messages
- ipa_dialog -> IPA.error_dialog
- ipa_init() -> IPA.init()
Initially the HBAC and Service entities have been rewritten to use the
new framework. The DNS is partially converted, the ipa_records_facet
is used to define custom records facet.
Other entities can still work using the old framework. The old framework
has been modified to be a wrapper for the new framework. Eventually all
entities will be converted to use the new framework.
Some unit tests have been modified to use the new framework.
The ipa_error_handler() has been modified to display the AJAX URL
that is having a problem. The ipa_cmd() error handler is now invoked
using call() to pass 'this' object which contains the URL.
The ipa_entity_set_association_definition() has been added to configure
the association between 2 entitites. By default the associator is
BulkAssociator and the method is add_member. The entities have been
updated to use the right configurations.
The ipa_cmd() has been modified to detect IPA errors and invoke the
error handler.
A bug in refresh_on_success() has been fixed as well.
Some errors (e.g. server down) are reported as AJAX success with
empty data. The ipa_cmd() has been modified so that it will detect
such errors and invoke the error handler.
The ipa_cmd() has been modified such that when an error occurs a
dialog box will appear showing the error message with 2 buttons:
Retry and Cancel. If Retry is clicked, it will attempt to execute
the same operation again. If Cancel is clicked, the operation will
be canceled and the control is returned to the caller.
New unit tests have been added to test ipa_cmd() on successfull
and unsuccessfull cases.
The associate.js, details.js, entity.js, search.js, and webui.js
have been modified to display the error message inside the page.
This behavior can be changed in the future (e.g. redirect to error
page).
The navigation.js and webui.js have been modified to render only
the visible tabs. This improves the performance and reduce hidden
errors. The navigation unit test has been modified to reflect this
behavior.
Some variables/functions also have been renamed for consistency.
Performing I18N completely on the server, to leverage the
existing gettext architecture.
Also, the browser does not have access to the Language header.
Added the additional po files for a set of required languages
conflict with install/static/ipa.js was resolved.
Note that the addition of the .po files in this patch is necessary.
In order to get Transifex support, we need to update the LINGUAS
file with the languages for which we want support. If we don't
add the .po files in, they get automatically generated by the rpmbuild
process. Our implementation of gettext has a bug in it (It might
be F13 thing) where the the Plurals line is not getting correctly
transformed, which causes a build failure. However, since the
RPM would have the .po files anyway, we should revision control
the ones we have, even if they are empty.
Fixed the Bug reporting url to the original value.
Corrected the Chartype encoding for UK
Test framework for Web UI has been created using qUnit. The test files
are located in install/static/test. The main page is index.html which
contains links to all test suites (xxx_tests.html). The test cases are
stored in xxx_tests.js. All test suites can be executed at once using
all_tests.html. The test data is stored in data folder. This patch
includes test suites for ipa.js and entity.js.
Some variables and functions in ipa.js have been modified to accomodate
testing (e.g. JSON URL, error handler, synchronous operation). The
sampledata has been moved to test/data. The develop.js and webui.js also
have been modified accordingly.
Quick summary:
- use jQuery UI and jQuery BBQ libraries
- code restructuring
The patch has so many changes they can't be listed here. Many parts
of the code have been rewritten from scrach.
See freeipa-devel mailing list:
webUI code restructuring [wall of text, diagrams, ... you've been warned!]
2010-09-07
-Refactored the associations code into a set of objects that are configured by the entities
-Added support for associations that can be done in a single rpc
-hostgroup to host and group to user associations working
-Restructed sampledata so that the file is matched automatically by the RPC method name
-The new ipa_cmd/sampledata scheme insists on there being sample data for any commands or the ipa_command fails.
-Added sampledata files for all the calls we make
-renamed several of the sampledata files to match their rpc calls
-Started a pattern of refactoring where all the forms for the entity fall under a single object
We now catch the hashchange event and use that to drive most of the site.
To trigger page transitions, modify location.hash.
Params start with # not ?.
Removed user-group.inc.
converted tabs to spaces
trivial imlementation of add and details for netgroup and hostgroup
lots of bug fixes based on routing problems and the refactorings.
to hash params ( starting with # ). User Details are now part of
index.xhtml, ao one more .inc file has been removed.
Updated commit to catch a few things that had been left out, including
sampledata handling and updateing Makefile.am
What it means?
Well, first I removed some development control variables from ipa.js.
Namely useSampleData and sizelimit. I moved useSampleData to the top
of index.xhtml. This way we won't forget about it when we don't need
it anymore. sizelimit has nothing to do in ipa.js and be hardcoded
for ALL commands! Some don't have this parameter and could fail.
Since ipa_init now loads meta-data over JSON-RPC, we need to wait for
it to finish its job. That's why I put a second parameter to ipa_init:
on_win. ipa_init will call on_win when all data is loaded properly and
we can start building the page.
