[meta] don't use deprecated methods, features

[bill]

General ticket for refactors to avoid using deprecated methods like dojo.connect() in source code.

Excludes changes that would break backwards compatibility, and fixes to modules that are themselves deprecated.

[bill]

In [30378]:

stop using dojo/_base/connect, refs #16585 !strict

[bill]

In [30379]:

oops, remove "debugger" statement, refs #16585

[bill]

In [30380]:

Stop using dojo/_base/event.stop(), refs #16585 !strict. It would be nicer though if dojo/on had a similar convenience method.

[bill]

In [30385]:

use dojo/sniff not dojo/_base/sniff, refs #16585 !strict.

[bill]

In [30386]:

Stop using dojo/_base/json, refs #16585 !strict. Turns out I can't use dojo/json::parse() because the parser can take non-JSON for things like data-dojo-props.

[bill]

In [30388]:

Use dojo/json instead of deprecated dojo/_base/json, refs #16585 !strict. Passes dojo/tests/_base unit tests on FF18/mac.

[bill]

In [30393]:

remove unnecessary uses of dojo/ready, refs #16585 !strict.

[bill]

In [30394]:

Use doh.run() rather than doh.runOnLoad(), since the latter depends on the deprecated dojo/ready module. Refs #16585 !strict. Not updating some tests for deprecated modules.

[bill]

In [30395]:

Use dojo/domReady rather than dojo/ready, refs #16585 !strict.

[bill]

In [30401]:

Use data-dojo-config instead of djConfig (as attribute passed to <script src=dojo.js>), and data-dojo-id instead of jsId. Demos, dijit, dojo, and util. Refs #16585.

[bill]

In [30402]:

Use data-dojo-config instead of djConfig (as attribute passed to <script src=dojo.js>), and data-dojo-id instead of jsId. DojoX. Refs #16585.

[bill]

In [30403]:

event="..." --> data-dojo-event="...", args="..." --> data-dojo-args="...", refs #16585

[bill]

In [30408]:

Stop using deprecated API's in themeTester, including:

• dijit/dijit and dijit/dijitAll rollup files
• ItemFileReadStore and dijit/tree/ForestStoreModel
• dojo and dijit globals
• dojo/ready

Refs #16585

[bill]

In [30410]:

Remove dependency on dojo/data/util/filter, since dojo.data is deprecated. I had to duplicate the patternToRegExp() function but when written concisely it's only a few lines (excluding comments). Refs #16585 !strict.

[bill]

In [30420]:

use dojo/query not dojo/_base/query, refs #16585

[bill]

In [30421]:

When possible without breaking backwards compatibility, use dojo/Deferred instead of dojo/_base/Deferred. I could actually change the other dojo/_base/Deferred usages too if I used a shim like dijit/Tree does, but it didn't seem worth it. Refs #16585 !strict

[bill]

In [30448]:

remove usages of deprecated LinkPane from tests, refs #16585

[bill]

In [30461]:

Monitor touchend in addition to mouseup, for mobile. Also, use this.own() with dojo/on instead of deprecated this.connect() . Refs #16622, #16585 !strict.

[bill]

In [30463]:

For _TextBoxMixin too, monitor touchend in addition to mouseup. Also, use this.own() with dojo/on instead of deprecated this.connect() . Refs #16622, #16585 !strict.

[bill]

In [30465]:

For _TemplatedMixin, use this.own() with dojo/on instead of deprecated this.connect() (which uses the deprecated dojo.connect()). Perhaps in the future we should setup a more standard way to map from strings to synthetic events, for example data-dojo-attach-event="dojo/mouse.enter: _onMouseEnter" instead of data-dojo-attach-event="mouseenter: _onMouseEnter". Refs #16585 !strict.

[bill]

In [30466]:

For Editor, use this.own() with dojo/on instead of deprecated this.connect() (which uses the deprecated dojo.connect()). But still using dojo/_base/connect._keypress synthetic event. That will be removed in a later checkin. Refs #16585 !strict.

[bill]

In [30467]:

For non-deprecated dijit widgets, use this.own() with dojo/on or dojo/aspect, instead of deprecated this.connect() (which uses the deprecated dojo.connect()). Refs #16585 !strict.

[bill]

In [30468]:

Intermediate checkin for StackContainer to stop using deprecated this.subscribe(). In a later checkin I'd like to get rid of using topics altogether. Also need to stop using deprecated synthetic keypress event. Refs #16585 !strict.

[bill]

In [30470]:

Move StackContainer/StackController to use native keydown/keypress events, rather than synthetic dojo/_base/connect::_keypress event. The onkeypress data-dojo-attach-event's in the templates technically still map to dojo/_base/connect::_keypress, but the code is no longer depending on evt.charOrCode, or for keypress events to fire for non-printables. Also clearing dangling uses of deprecated this.connect(). Refs #16585 !strict.

[bill]

In [30471]:

Remove uses of this.connect() from Editor plugins, but leave definition of _Plugin::connect() method for backwards compatibility.
Plugins (and Editor) still using synthetic dojo/_base/connect::_keypress event in a few places. Refs #16585 !strict.

[bill]

In [30480]:

Deprecate typematic's charOrCode option, and un-deprecate keyCode and charCode, making keyCode use on(node, "keydown", ...), and charCode use on(node, "keypress", ...). Dijit only uses keyCode.

This reversal of the original plan is because dojo/_base/connect itself is deprecated, so if typematic is carried over to version 2, it cannot support charOrCode. Using on(node, "keydown", ...) solves the problem mentioned in #6667 about the left parenthesis being treated as the down arrow.

Refs #6667, #16585 !strict.

[bill]

In [30483]:

Seems "keydown" is better than "keypress" for catching Ctrl-W (to close a tab), refs #16585 !strict.

[bill]

In [30484]:

Update varioius widgets to use native keydown event and evt.keyCode rather than synthetic keypress event and evt.charOrCode. Didn't touch the code in _TextBoxMixin that calls this._onInput() with a homegrown synthetic event though.

Refs #16585 !strict.

[bill]

In [30485]:

Make Editor use native keydown instead of synthetic dojo/_base/connect._keypress event, refs #16585 !strict.

[bill]

In [30502]:

fix themeTester breakage from [30408], fixes #16642, refs #16585

[bill]

In [30503]:

Stop using deprecated dojo/_base/event::stop(). This leaves lots of places calling both evt.stopPropagation() and evt.preventDefault(), somewhat bloating the code, but I think many of those cases can be changed to just calling either evt.stopPropagation() or evt. preventDefault().

Also doing coding style fixes.

Refs #16585 !strict.

[bill]

In [30508]:

Rollback spurious indentation changes from [30483], refs #16585 !strict. I'd rather check them in separately instead of mixed with real changes.

[bill]

In [30509]:

Rollback spurious indentation changes from [30503], refs #16585 !strict. I'd rather check them in separately, rather than mixed with real changes.

[bill]

In [30510]:

Rollback spurious indentation changes from [30503], refs #16585 !strict. I'd rather check them in separately, rather than mixed with real changes.

[bill]

In [30511]:

Rollback spurious indentation changes from [30474], refs #16585 !strict.

[bill]

In [30512]:

For dojo/store, use dojo/when in preference to deprecated dojo/_base/Deferred::when, and also change one usage of dojo/_base/Deferred to dojo/Deferred (without breaking backwards compatibility), refs #16585 !strict.

[bill]

In [30554]:

use aspect.after() rather than deprecated this.connect() method, refs #16585.

[bill]

In [30672]:

remove use of deprecated dojo/_base/unload::addOnWindowUnload, refs #16585 !strict

[bill]

In [30768]:

Fix data-dojo-attach-event on widgets (as opposed to DOMNodes), fixes #16798, refs #16585 !strict.

[bill]

In [30769]:

Fix data-dojo-attach-event on widgets for DTL. Also stop using deprecated signature of _attachNodes() where you pass an array of widgets rather than a DOMNode. Refs #16798, #16585 !strict.

[bill]

In [30793]:

use _WidgetBase rather than _Widget in tests, refs #16585

[bill]

In [30800]:

remove calls to deprecated this.disconnect() method, refs #16585 !strict

[ben hockey]

Replying to bill:

In [30484]:

Update varioius widgets to use native keydown event and evt.keyCode rather than synthetic keypress event and evt.charOrCode. Didn't touch the code in _TextBoxMixin that calls this._onInput() with a homegrown synthetic event though.

Refs #16585 !strict.

umm... what happened to the templates during this commit? they have some strange whitespace now.

[bill]

Oh, that's annoying. My IDE did that; I accidentally clicked a weird checkbox and it updated the files during checkin. I'll revert that. Actually, I reverted most of the spurious reformatting already but didn't notice the templates.

[bill]

In [30879]:

revert unintended spacing changes from [30484], refs #16585 !strict

[bill]

In [30934]:

Move code to remove attach points and events to separate method, so it can be called independently of destroyRendering(), fixes #16889 !strict.

Also removing redundant call to this.own(), from [30465], refs #16585.

[bill]

In [31168]:

after [30480] kernel dependency no longer used, refs #16585

[bill]

In [31172]:

commenting some code for deletion in 2.0, refs #16585 !strict.

[bill]

In [31173]:

Add tests for LayoutContainer's new API: LayoutContainer.design and children using region and layoutPriority. Note that for backwards compatibility, "design" shouldn't affect children that set layoutAlign instead of region. This is basically "accidentally" working due to the way _getOrderedChildren() happens to be written. Refs #16770 !strict.

Also convert uses of deprecated layoutAlign in calls to utils.layoutChildren() to region, refs #16585 !strict.

[bill]

In [31398]:

don't depend on programmatic instantiation picking up id from DOMNode, refs #16585

[bill]

In [31404]:

don't use domConstruct.create() to setup event listeners, refs #16585 !strict

[bill]

In [31405]:

use data-dojo-attach-point rather than dojoAttachPoint, refs #16585

[bill]

In [31406]:

use dojo/dnd/common.getCopyKey() rather than dojo/_base/connect.isCopyKey(), refs #16585 !strict

[bill]

In [31407]:

Modules are supposed to list their NLS resources in the define() dependencies. RangeBoundTextBox is missing "dojo/i18n!./nls/validate". I guess it gets pulled in by a superclass but it's standard practice to list it explicitly. Plus necessary, starting in 2.0. Refs #16585 !strict.

[bill]

In [31408]:

use data-dojo-attach-point rather than dojoAttachPoint, refs #16585

[bill]

In [31410]:

Various themeTester cleanup:

1. use boilerplate.js rather than _testCommon.js
2. don't list expected globals as stray globals
3. don't use parser's auto-load
4. don't use deprecated dijit/layout/LinkPane

Refs #16585, #17022.

[bill]

In [31416]:

stop using _WidgetBase.connect() and other deprecated methods in test files, refs #16585

[bill]

In [31420]:

Fix TimeSpinner to use keydown rather than keypress, fixes #30484 and refs #16585 on 1.9 branch !strict.

[bill]

In [31421]:

Fix TimeSpinner? to use keydown rather than keypress, fixes #30484 and refs #16585 on trunk !strict.

[bill]

Oops, above two changes form #17101, not #30484.

[bill]

In [31424]:

Fix TimeSpinner? to use keydown rather than keypress, fixes #30484 and refs #16585 on trunk !strict.

[bill]

Oops, [31424] actually fixed #17033 and had nothing to do with this ticket.

[Bill Keese <[email protected]…>]

In e154822ddab4c272332f9fb8ffddac56560322b3/dijit:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled? 

[Bill Keese <[email protected]…>]

In 0bd911d395e08c929f5b252a2fec18d688fb27b6/util:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled? 

[Bill Keese <[email protected]…>]

In 5d4d2d35032f3ff8d173320b1d2df317cea775ce/dijit:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled? 

[Bill Keese <[email protected]…>]

In 16f55ebe526ecd839f672a4affbbf3f674b24175/dijit:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled? 

[Bill Keese <[email protected]…>]

In c535128334b5a130b78c79b87d2b236ab0490d9b/dijit:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled? 

[dylan]

meta ticket, moving to 1.12.

[dylan]

Ticket planning... move current 1.13 tickets out to 1.15 to make it easier to move tickets into the 1.13 milestone.