1. IsEditable is much less hacky. It basically says:
1. if the node is a special bogus text node, it's not editable
2. if the node is a text node with only newlines, then it's not editable
3. if the node has no frame associated with it, then it is not editable
4. otherwise, it is editable.
As part of this, I had to make IsEditable non-static. This had some cascading effects on static methods that called
IsEditable, which likewise had to be made non-static.
2. SetTextProperties now works when the endpoints are leaf nodes (like images), not just text. Charlie, you should
put this through some stress cases.
nsIHTMLEditor, nsITextEditor, and nsIEditorShell and enabled
the "Start Log" and "End Log" menu items in the editor test bed.
* Modified makefiles for Unix and Win32 to turn on Editor API Logging.
* Added action argument to the DeleteSelection() method in nsIEditorShell.
* Modified nsJSTxnLog to use nsJSEditorLog when writing it's output.
for HTML and CSS atoms, since we can't get these from layout yet. The old
use assumed one editor, but as we know many editors can exist simultaneously.
Now, the editors share a singleton which is fetched from a factory method,
saving memory and allocation overhead.
This checkin adds the implementations for encoding
HTML and plain text:
nsTextEncoder::EncodeToString(..)
nsHTMLEncoder::EncodeToStream(..)
Hooked these services into the nsHTMLEditor implemenation
We probably want to update the clipboard code to
use the same pattern as in nsHTMLEditor.cpp
* Added some code to restore selection after RemoveTextProperty.
* Made a tiny change to the way the editor holds onto the transaction manager, so client (like the ender text control) can
turn undo on and off effectively
* Fixed a bug in IsNodeInline that was causing GetBlockParent to return the wrong parent node (and who knows what
other mischief) where atom compares were being done on strings of different case, for <B>, "b" vs. "B". The DOM
method nsIDOMElement::GetTagName forced the returned tag name to upper case. I don't know how long that has
been true, but I assume it's a fairly recent change (or something on our side related to it changed recently) or this method
never would have worked. I also found a few cases of nsString::Equals that I changed to EqualsIgnoreCase
* Made some general improvements to the interaction between editors and rules, so that the aCancel out param is always
intialized, and all rules with Will... are matched with a Did... call. Added Will/DidSetTextProperty() and
Will/DidRemoveTextProperty(). This helps enable plain text mode. Added a skeleton for max length support in text editor.
* fixed some warnings.
* added focus listener. Doesn't do much yet, but when focus notifications start appearing, we'll be ready for them. The code is in
place to hide selection when we lose focus and paint selection when we get focus. That's probably not quite right, but it's a start.
We will need to be able to determine the distinction between losing focus to another control within our app, and losing focus to
another app.
* added support for disabled and readonly states in the editor. This is accomplished by having flags set by the client, and letting the
rules system deal with those flags. The flags I added are:
TEXT_EDITOR_FLAG_PLAINTEXT 0x01 // only plain text editing is allowed
TEXT_EDITOR_FLAG_SINGLELINE 0x02 // enter key and CR-LF handled specially
TEXT_EDITOR_FLAG_PASSWORD 0x04 // text is not entered into content, only a representative character
TEXT_EDITOR_FLAG_READONLY 0x08 // editing events are disabled. Editor may still accept focus.
TEXT_EDITOR_FLAG_DISALBED 0x10 // all events are disabled (like scrolling). Editor will not accept focus.
* added WillInsertBreak/DidInsertBreak into text rules, so flags could be checked. This gets us readonly, disabled, and single line
behavior.
* cleaned up the code that allocates, registers, and destroys event listeners. Thanks to Kin and Simon for cleaning up the
ownership model on the listeners, it was a big help.
* added support for a max text length. You can now tell the text editor, be no bigger than n characters.
features like set paragraph style, set list type, indent, etc. I'm sure there's holes in my
implementation that these high level features will make obvious.
2. I've factored a lot of utility methods from several modules into nsEditor as public static methods.
This makes them easily accessable to all, and will help Joe and I remove redundant methods.
3. I changed the HTML tags to lower case, and made all string compares case-insensitive. No, this
isn't quite the right thing to do, but we don't have atoms from layout yet. The Right Thing is for us
to reuse those atoms.
nsI*Editor::OutputText(nsString& aOutputString);
nsI*Editor::OutputHTML(nsString& aOutputString);
These methods always returns back a Unicode version of whatever is in the content model. It is the
responsibility of the caller then to call whatever converter is required to convert to the appropriate
charset.
Added:
nsI*Editor::OutputText(nsIOutputStream* aOutputStream, nsString* aCharsetOverride = nsnull)
nsI*Editor::OutputHTML(nsIOutputStream* aOutputStream, nsString* aCharsetOverride = nsnull)
These methods output the the current content model to aOutputStream. The document is encoded using the
document defined charset or if the user passes in a non-null value for aCharsetOverride then this
encoding overrides the encoding used by the document.
1. fix for bug 5796, crash on exit. This was a bad, bad memory smudge on my part, easily fixed by doing the right ref counting in the
right places.
2. some preliminary code for M6 block transformations has leaked into this checkin. It's safer than trying to re-code the fix above into
a fresh tree. Unless you're making calls to do block transformations, you won't see any difference.
* added TextEditorTest.cpp, a unit test module for nsTextEditor. It's use is actually commented out since my checkin is happening so late due
to all-day build bustage, and I don't have a Mac handy to verify. With someone's Mac help tomorrow, I can turn it on.
* some minor bug fixes to property handling
Split nsEditor::DeleteSelectionAndCreateNode into two parts,
so that routines like InsertImage which need to set attributes
on the new node before inserting into the tree can do so.
This is somewhat temporary, since the current method isn't undo-able;
undo will be implemented after M4.
added aFirst out param to GetTextProperty, so the caller can know if the first character has the property in the case of aAny=true and aAll=false.
fixed a bunch of places where result was being used incorrectly as a return val from do_QueryInterface
some minor undo/redo fixes to split and join of interior nodes.
anywhere any more.
* Cleaned up split and join undo/redo.
* Added TypeInState, a data struct that remembers things about text properties for collapsed selections, so you can type
* Ctrl-B with an insertion point and the next character will be bold.
* Added all the logic to handle inline vs. block elements when setting text properties.
* Added some support for italic and underline as well. Adding these things is pretty easy now. Ctrl-B, Ctrl-I, Ctrl-U for testing bold, italic, underline.
* Added all the logic to make sure we only add style tags where they're needed, so you should never get the same style tag nested within itself, except as needed for block elements.
* Added methods for testing a node to see if a particular style is set. This isn't 100% done yet, but with very little work we could have toolbar buttons that respond to selection changed notification that show the state of bold, italic, underline, etc. in real time. Supports tri-state: whole selection is bold, some of selection is bold, none of selection is bold, ...
* Fully undoable and redoable.
* Added some debug printfs to transactions and editors. all controlled by a gNoisy static in each module. helps me track down undo/redo problems. if the output bugs people enough, I'll shut it off and re-enable it in my local tree.
Noticably missing: make un-bold, make un-italic, etc. This is coming soon.
trivial to add more properties
finished first cut at SetTextProperty. This triggers lots of crashes in
range/selection code where we're holding onto a stale frame pointer (at
least, that's my best guess.)
synched with Charlie's change-o-rama
added an assert in DeleteTextTxn::Init() checking aNumCharsToDelete vs.
0
NS_ASSERTION(0!=aNumCharsToDelete, "bad arg, numCharsToDelete");
if the number of chars to delete is 0, we shouldn't be creating a
transaction at all. I had never seen this condition arise before
Charlie's checkin, I don't know if he introduced it or if it was a
latent bug I just never tripped over before.
added nsEditProperty
beginning of implementation for SetTextProperties. Currently, it only works if the selection is entirely within a single text node. Currently only
supports bold and italic (mapped to ctrl-b and ctrl-i for testing purposes.)
changed a bunch of interfaces to make things easier, like CreateElement now returns (as an out-param) the element created.
