dbaron@dbaron.org
f4264ddfc8
Rename {nsIDocument,nsIDocumentObserver,nsIFrame}::ContentChanged to CharacterDataChanged, since that's the only thing it's used for. Replace its nsISupports* parameter with a PRBool, since we don't need to allocate an object to pass a boolean parameter. Remove some (but probably not all) unneeded implementations on frames whose content is never a text node. b=234960 r+sr=jst
2004-02-20 21:38:31 +00:00
dbaron@dbaron.org
de18c7a5d5
Replaced elements should not support generated content. b=169334 r+sr=bzbarsky
2004-02-19 18:04:15 +00:00
bryner@brianryner.com
d9d66e38d5
DeCOMtaminate nsCSSFrameConstructor (bug 231157). This gets rid of the nsIStyleFrameConstruction and nsICSSFrameConstructor interfaces. The frame constructor now has the same lifetime as the pres shell. r+sr=dbaron.
2004-01-21 23:05:10 +00:00
dbaron@dbaron.org
719ccc2fad
Refactor code that calls ComputeStyleChangeFor and handles the result. b=231027 r+sr=bzbarsky
2004-01-17 22:43:14 +00:00
dbaron@dbaron.org
50a661cc7a
Do style resolution of later siblings required by the CSS + combinator on event state and attribute changes by changing the result of Has(Attribute|State)DependentStyle into a bitfield and setting and handling the restyle-later-siblings bit appropriately. b=15608 r+sr=bzbarsky
2004-01-14 01:36:35 +00:00
bryner@brianryner.com
a90ec6ff21
Eliminate nsIStyleSet as an XPCOM interface, making the implementation internal to layout. Frame constructor ownership moves to the pres shell, and the style set methods that forward to the frame constructor are gone. Rewrote style set sheet add/remove to reduce code size. For more details, see bug 64116. r+sr=dbaron.
2004-01-07 22:30:53 +00:00
dbaron@dbaron.org
9446ce805f
Use the correct term "float" instead of "floater". b=218654 r+sr=bzbarsky
2003-10-13 21:51:02 +00:00
dbaron@dbaron.org
660339f971
Remove hint parameter from AttributeChanged methods and simplify style hint mechanism. b=211308 r+sr=bzbarsky
2003-07-11 21:16:12 +00:00
dbaron@dbaron.org
da6a024b2e
Ensure relatively positioned floats act as containing block for absolutely positioned elements by using (and adding a parameter to) nsCSSFrameConstructor::ConstructBlock. Remove ProcessBlockChildren, which no longer needs to be separate from ProcessChildren. b=210873 r+sr=bzbarsky
2003-06-30 21:48:07 +00:00
dbaron@dbaron.org
07b8818b23
Handle state and attribute changes on elements without frames more efficiently, and keep the undisplayed map up to date. b=209733 r+sr=bzbarsky
2003-06-18 21:54:51 +00:00
dbaron@dbaron.org
99d81c4dbc
Make CSSStyleRuleImpl immutable by changing mutation code (including some merging of code between nsDOMCSSAttributeDeclaration and DOMCSSDeclarationImpl) so that it creates a new rule for the modified declaration. Change StyleRuleChanged notification signatures, and change pres shell's handling of style rule change, addition, and removal to just do reresolution. Remove fine-grained ways of clearing style data. Change PeekStyleData so it doesn't do more work than it needs to, and can be relied on not to access rules that no longer match any content. b=188803 r+sr=bzbarsky
2003-06-18 01:59:57 +00:00
scott@scott-macgregor.org
e59b260fff
Bug #201891 --> lazily instantiate CSSFrameConstructor::gXBLService in order to break a weird cycle
...
with the component manager that can cause two instances of the xbl service to be created.
r/sr=dbaron
2003-06-03 04:23:56 +00:00
dbaron@dbaron.org
4a04463107
Remove some code left over from old native form controls. b=205575 r+sr=roc a=asa
2003-05-14 22:36:33 +00:00
bzbarsky@mit.edu
7dda42fc18
Pass the content parent frame to ConstructTableFrame so the view that's created
...
goes in the right place in the z-ordering. Bug 203086, r+sr=roc, a=asa
2003-04-26 01:08:35 +00:00
bzbarsky@mit.edu
c4f180a7f9
Make sure that Init() failures for image frames don't mean no alt text. Also
...
fixes a crasher when reframing fixed-pos frames.
This change adds an assertion in ConstructHTMLFrame that InitAndRestoreFrame
should succeed (unless it returns NS_ERROR_FRAME_REPLACED). This is an interim
change; eventually we want to bail out on errors, but there may be frame impls
that return errors in non-error conditions from Init() (eg nsFrameSetFrame did
that) and we're a little close to freeze.
Bug 202506, r=rbs, sr=roc+moz
2003-04-21 23:57:47 +00:00
bryner@netscape.com
4682cd8077
Remove nsScrollFrame and frame constructor support for it. Remove dlldeps.cpp for layout, it's not necessary. Bug 201457, r+sr=roc.
2003-04-11 08:26:42 +00:00
bryner@netscape.com
4486f8d13d
Make --disable-xul work (bug 69995). We build enough XUL to support
...
scrollbars unconditionally, but can now leave out heavyweight widgets like
grid, listbox, and tree. r=bzbarsky, sr=alecf.
2003-04-11 00:56:27 +00:00
roc+@cs.cmu.edu
7143d14503
Bug 126263. Whack comboboxes/listboxes to use Gfx scrollbars. Also, make scrollbars attach their XBL content only when they become visible, to avoid pageload hit. r+sr=dbaron
2003-04-08 20:50:57 +00:00
dbaron@dbaron.org
e9466869d1
Handle inline style changes by creating new style rules to wrap the modified nsCSSDeclarations, thus moving towards style rule immutability and away from property-based change hints. Patch from Boris Zbarsky <bzbarsky@mit.edu>. r=dbaron sr=roc b=171830
2003-03-06 19:07:22 +00:00
roc+@cs.cmu.edu
a4211a4217
Bug 186754. Make sure scrolled elements honor the content hierarchy for z-index. r+sr=bzbarsky
2003-02-23 02:03:43 +00:00
bryner@netscape.com
ba58162e96
Remove nsIStyleContext (bug 114713), r/sr=dbaron. Changed lots of methods to use nsStyleContext* or already_AddRefed<nsStyleContext> as return values, depending on whether a reference is returned. Turn on nsRefPtr (bug 104346), and change nsDerivedSafe to not declare AddRef and Release explicitly virtual, r=jag, sr=dbaron.
2003-02-22 00:32:13 +00:00
karnaze@netscape.com
a9a3087c48
bug 179683 - implement printing of fixed positioned elemements, clean up ViewportFrame. sr=roc, r=bzbarsky.
2003-01-20 18:04:34 +00:00
dbaron@fas.harvard.edu
8de079b0b8
Don't set NS_BLOCK_SPACE_MGR and NS_BLOCK_WRAP_SIZE on relatively positioned blocks, by default. Use |ConstructBlock| for them rather than duplicating it, slightly incorrectly. b=180711 r+sr=roc
2002-12-13 20:13:17 +00:00
karnaze@netscape.com
69749a7a87
bug 180805 - Added GetAdjustedParentFrame to handle fieldsets. Moved inner/outer table logic there. Added fieldset logic to IsValidSibling. sr=bzbarsky, r=jkeiser.
2002-12-03 16:02:42 +00:00
bzbarsky@mit.edu
878f8d36d9
ReframeContainingBlock needs to work on <body> and such too. Bug 74951,
...
r=caillon, sr=dbaron
2002-11-06 02:26:52 +00:00
roc+@cs.cmu.edu
458e434a36
Bug 160936. Add type nsChangeHint for style hints. Make it a set of bits. Make sure an NS_STYLE_HINT_REFLOW always also does a repaint of the frame. r=dbaron,sr=kin
2002-08-24 14:41:28 +00:00
jkeiser@netscape.com
a807eea307
Make form within table or tr not create a cell frame and screw up layout. bug 149282, r=dbaron@fas.harvard.edu, sr=waterson@netscape.com
2002-06-14 06:47:35 +00:00
dbaron@fas.harvard.edu
1c8634debe
Fix crashes related to editing of HR elements by wrapping generated content for leaf frames in an extra frame so :before and :after are always the first/last children of the primary frame. Disable generated content for floating and absolutely positioned leaf frames. b=141054 sr=waterson r=kin
2002-06-12 03:27:48 +00:00
dbaron@fas.harvard.edu
712b354488
Clean up FindFrameWithContent. No need to search children of special frames too (hang bug 140275). Use special sibling when next-in-flow isn't available, and use the hint even when we have special frames. Convert gotos into additional while loop. Turn useless :after check into assertion. Don't modify aParentFrame and aParentContent for when the hint fails. b=140275 r=attinasi sr=waterson
2002-05-16 13:39:44 +00:00
karnaze@netscape.com
4e7faac33d
bug 113235 - avoid reframing when content is inserted inside an inline frame which contains a block. sr=waterson, r=attinasi.
2002-04-30 21:34:15 +00:00
karnaze@netscape.com
109d19539b
bug 136848 - consider a sibling of type col group, col, row group as a valid siblings of a child only if the child is the same type. sr=attinasi, r=kin
2002-04-19 03:11:29 +00:00
dbaron@fas.harvard.edu
4b63835325
Fix up and enable hierarchical :hover and remove the hacks that made :hover work for anchors without it. Pass mask rather than pseudo-class to ContentStatesChanged notifications. Optimize event state changes in the style system by passing state mask to HasStateDependentStyle and using it in HasStateDependentStyle implementations. b=5693 r=bryner, joki sr=hyatt
2002-04-11 03:49:30 +00:00
hewitt@netscape.com
ffbc59afea
bugs 110156, 110155
...
Removing support for <outliner> tags
Removing <tree> layout code and moving <tree> tags to outliner layout
Convert all usage of <outliner> to tree tags
Convert all usage of <tree> tags to new <tree> syntax or <listbox>
r=cmanske,varga sr=hyatt,sspitzer a=asa
Also includes fixes by Jan Varga (varga@utcruk.sk ) for bugs 132020, 133451, 131393, 115894, and 129327
2002-03-29 02:46:01 +00:00
karnaze@netscape.com
9a3780e93f
bug 24000 - implement page-break-before/after:always using new nsPageBreakFrame. Tables handle page breaks internally. a=asa, sr=attinasi, r=alexsavulov
2002-03-17 21:35:08 +00:00
dbaron@fas.harvard.edu
1eda44576d
Attempt to fix topcrash bug 118014 by calling DeletingFrameSubtree in a bunch of places where it should be called to ensure that we don't leave dangling pointers in the undisplayed content map. sr=attinasi r=kin a=brendan
2002-03-13 13:51:20 +00:00
bryner@netscape.com
367cc93c8f
- Implement the :checked CSS pseudoclass which maps to the "selected" property on option elements.
...
- Eliminate the _moz-option-selected attribute; move the actual selected state into the option content node.
- Change all users of _moz-option-selected to use :checked.
- Add a third parameter to nsIDocument[Observer]::ContentStatesChanged to indicate which pseudoclass changed, this is used for optimizing handling of :checked state changes.
Bug 128947, r=dbaron, sr=jst, a=asa.
2002-03-07 03:34:29 +00:00
attinasi@netscape.com
04335c106c
Fixed problems with reframing recursively due to block-in-inline frames and content appended notifications. r=karnaze sr=waterson b=56894
2002-01-14 22:05:26 +00:00
hewitt@netscape.com
77b8bdf68d
70858 - [XUL 1.0] Implement <listbox>, r=evaugahn, sr=hyatt
2002-01-12 01:18:10 +00:00
bryner@netscape.com
2e1743aab5
Initial checkin of XBL select widget (listbox) code (bug 112713).
...
Summary of changes:
- Removed old XBL form control cruft.
- Made nsOutlinerContentView accessible from content scripts.
- Created a new interface, nsIDOMNSXBLFormControl which exists to allow a
select element to have an outliner box object.
- Factored out nsChildIterator into its own file.
- Added support to the outliner content view for having option elements
directly inserted in the outliner.
- Cleaned up pixel/twip conversions in outliner.
- Added support for intrinsic width for outliner.
r=jag, sr=hyatt. Note that the new select widget is NOT turned on by
default.
2002-01-05 07:17:10 +00:00
dbaron@fas.harvard.edu
4fa6450214
Split ConstructFrameByTag into ConstructTextFrame and ConstructHTMLFrame to make it easier to reorganize frame construction by tag and by display type and to speed up text frame construction. b=109788 r=bryner sr=hyatt
2001-12-09 04:28:41 +00:00
dbaron@fas.harvard.edu
2507c79cf4
Backing out most of my checkin for bug 109788. b=111710
2001-11-24 23:28:45 +00:00
dbaron@fas.harvard.edu
0438bea566
Split ConstructFrameByTag into ConstructTextFrame and ConstructHTMLFrame to make it easier to reorganize frame construction by tag and by display type. b=109788 r=hyatt rs=waterson
2001-11-23 04:52:24 +00:00
attinasi@netscape.com
355d4dfc00
Export GetAltTextfor as static member fcn. b=102281 r=kmcclusk sr=waerson
2001-11-07 04:32:25 +00:00
jst@netscape.com
af4947ee24
Checking in John Keiser's <jkeiser@iname.com> fix for *tons* of form submission and <select> related bugs, see tracking bug 34297 for details. r=rods@netscape.com, sr=jst@netscape.com
2001-11-02 07:40:01 +00:00
dbaron@fas.harvard.edu
6b4c727ec3
Avoid attempting to do CSS selector matching on text nodes, comments, or processing instructions by adding a 'NonElement' style resolution method that assumes that no rules match (temporarily, until we don't have style contexts for text nodes) and by cleaning code that was using textPseudo (the usual case) or passing the text node directly (only a few unusual cases). b=56117 r=hyatt sr=attinasi
2001-10-30 06:02:05 +00:00
hyatt@netscape.com
3bc745993d
Fix for 106212, make ContentAppended work right with XBL insertion points. r=brendan, sr=waterson
2001-10-25 07:31:18 +00:00
hyatt@netscape.com
f89422ce49
Fix for 104336, deCOMtaminate rulenode, rulewalker, frames, r=dbaron, sr=waterson
2001-10-24 00:01:09 +00:00
dougt@netscape.com
2605bce49e
Extraneous Symbol Removal. Patch by Jon Smirl jonsmirl@mediaone.net. r=dougt@netscape.com, sr=alecf@netscape.com. Bug 104172 , 104316 , 104420 , 104426 , 104461 , 104511 , 104527
2001-10-23 01:37:21 +00:00
gerv@gerv.net
58d409fba9
Relicensing Round 1, Take 2. Most C-like NPL files -> NPL/GPL/LGPL. Bug 98089.
2001-09-28 20:14:13 +00:00
waterson@netscape.com
f9091f0d03
Bug 84645. Move XBL insertion point logic into nsCSSFrameConstructor, where it can interact nicely with continuing frames. r=dbaron, sr=hyatt
2001-09-12 20:19:41 +00:00
