bd8feefd91
Fix invalid optimization that prevented style changes on inline elements that caused a reflow from applying to more than the first line. b=98783 r=attinasi sr=hyatt
2002-02-16 16:31:49 +00:00
e47cc4c2bc
Do restoration of information in the layout history state immediately after frame initialization rather than using separate code within PresShell::ContentAppended that calls GetPrimaryFrameFor (which is slow and forces every frame into the primary frame map, inefficiently). b=109428 r=rods sr=waterson
2002-01-11 20:10:46 +00:00
40437d1fd6
Fix tabbing through image maps and fix Windows crash right-clicking on image maps by reverting my changes to nsImageMap from bug 110911 and by giving the primary frame map larger entries to accomodate mapping from a content node to a frame with a different |mContent|. b=115481, 114377 r=bryner sr=attinasi
2002-01-05 15:22:52 +00:00
0ec3f045d8
Fix for 112980, nsITheme API and dynamic skin switching re-enabling. r=ben, sr=waterson
2001-12-17 22:51:39 +00:00
e46c895524
Fix blocker bug 114675 with null check around assertion code that I meant to check in yesterday (bug 114220).
2001-12-11 19:16:15 +00:00
743e67725c
Backing out changes I didn't mean to check in.
2001-12-11 05:41:43 +00:00
dc85afb6b7
Prevent input elements from adding frames back into the primary frame map during their destruction (and leaving the map with dangling pointers) by passing the frame to functions that will be called during frame destruction to avoid calls to GetPrimaryFrameFor that will repopulate the primary frame map. b=114220 r=jkeiser sr=jst
2001-12-11 05:39:36 +00:00
01f0380aa0
Changing noisy warning I added yesterday from |#ifdef DEBUG| to |#ifdef DEBUG_dbaron|. b=110911 / 113810
2001-12-06 19:53:37 +00:00
c3d6b55ca5
Remove nsDST and convert its users (PresShell::mSubShellMap, FrameManager::mPrimaryFrameMap, and FrameManager::PropertyList::mFrameValueMap) to PLDHash. Also convert FrameManager::mPlaceholderMap to PLDHash. Give image map area structures a pointer back to their image frame rather than (mis)using the primary frame map. Change API for changing placeholder map to reflect hash entry that accesses key through value by passing only the value and not the key. Give the root frame a proper mContent. b=110911 r=attinasi sr=brendan
2001-12-06 05:45:07 +00:00
7e8137a340
Bug 41119 - Don't waste time notifying the frame manager about frame destruction if it is happening from nsFrameManager::Destroy. r=dbaron, sr=waterson.
2001-12-01 00:22:00 +00:00
cec98cc49a
Perf and footprint opt. Make style re-resolution cheaper. r=glazou, sr=hewitt, bug 109261
2001-11-14 08:19:13 +00:00
6d64ded4f2
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
a041591297
Ensure that we don't re-resolve placeholders to the style context that should be on their out-of-flow frame. Fixes autocomplete popup. b=107744 r=waterson sr=hyatt
2001-11-02 04:14:58 +00:00
c13d4f0b0c
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
4b9b701faf
two more nsUnicharUtils.h stragglers, I think I got r=jag, sr=scc on these in a previous patch - again for bug 100214
2001-10-26 17:51:22 +00:00
5574ddbbfc
No need to be this paranoid, checking for pointers being null once is enough! Checking twise in a row is just overkill.
2001-10-26 03:31:49 +00:00
157c017816
Bug 86947:
...
Make the line list doubly linked and access it through a list class and iterators.
Stop recomputing margins on all of the children of each block in the reflow chain (which causes O(N^2) state recovery during incremental reflow). Instead, add a second dirty bit to the lines and walk backwards through the line list to recompute vertical margins only when either dirty bit is set and the previous line was not reflowed. Add nsIFrame::IsEmpty to identify frames through which margins collapse.
Fix O(N^2) propagation of float damage by maintaining a set of intervals damaged by floats (bug 61962) and be sure to damage the correct areas (bug 48138).
Introduce nsCollapsingMargin to do correct collapsing of combinations of positive and negative margins (bug 50142).
Clean up some odds and ends and fix another smaller O(N^2) problem in nsBlockFrame::AddFrames.
r=attinasi, rbs sr=waterson
2001-10-25 01:08:40 +00:00
0c8c1348b3
Fix for 104336, deCOMtaminate rulenode, rulewalker, frames, r=dbaron, sr=waterson
2001-10-24 00:01:09 +00:00
b5ba390898
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
ed869c764e
Replace uses of overloaded |fputs| and most uses of |ns[C]String::ToCString|. b=104763 r=jag rs=scc
2001-10-16 03:53:44 +00:00
4966d85f91
Fixes for printing bugs 84017 and 89151. The fix for bug 84017 removes the need for a resize reflow to get form frame values to print. The restoration of these values into the new frame model is done as part of the initial reflow. The fix for bug 89151 gets rid of the extra frame model built and forces printed images to come out of the decoded image cache. r=darinf, heikki, sr=rpotts
2001-09-29 17:50:15 +00:00
d76d3fdf0d
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
422834eff1
Bug 66835 - fix method hiding warnings by renaming some methods on nsIContent. r=jag, sr=jst.
2001-08-17 08:14:14 +00:00
6bf6b98abb
Bug 93972: Don't save form control values in session history if they have an autocomplete attribute set to off, or if their containing form has an autocomplete attribute set to off. This is accomplished by returning an empty 'state key' which indicates that the state should not be stored in session history's hash table. r/sr=jst@netscape.com,vidur@netscape.com
2001-08-16 07:18:24 +00:00
a3c1dd1b09
Use NS_PTR_TO_INT32 macros to do 64-bit safe pointer conversions.
...
Bug #20860 r=Roland.Mainz@informatik.med.uni-giessen.de sr=brendan@mozilla.org
2001-08-14 07:59:59 +00:00
1e20984ef0
fixing bug 78690. removing old imagelib. r=jst sr=waterson
2001-08-09 20:22:39 +00:00
84f1ff48ed
Fix for 90081, r=dbaron, sr=attinasi
2001-07-25 06:33:06 +00:00
6c393ae2f4
Fixes for 80512 and 83707. r=blake, sr=waterson, a=asa
2001-06-04 00:22:43 +00:00
b135c5b55a
Fix for 78695 (rule matching improvements). r/sr=attinasi, jst, waterson
2001-05-31 22:19:43 +00:00
1ac7a3715e
Bug 77834: Fix problem where form control state is being restored to the wrong form control or not restored at all when going back and forth through session history. The problem was that the hash key we used was the content ID, which was not consistent from page load to page load. The fix was to key instead by index of form in document / index of control in form, with logic to fall back to content ID if we are unable to do this. r=nisheeth@netscape.com, sr=jst@netscape.com, a=drivers@mozilla.org (blizzard@mozilla.org)
2001-05-30 11:26:21 +00:00
b0f9d44ead
Bug 79508. If the computed style change is going to cause a reflow (or worse), then don't touch any continuing frames. This keeps us from queuing an incremental reflow to a continuation, but is not ideal (e.g., there may be cases where we don't properly recompute the style contexts for continuations). r=rbs, sr=attinasi
2001-05-23 06:39:34 +00:00
ee02375ced
Bug 76407. Handle second CantRenderReplacedElement() request for same frame: check to see if we've already posted an event for the frame. r=dbaron, sr=attinasi, a=chofmann
2001-04-24 00:05:33 +00:00
06d55ad850
bug 71370; author=simon@softel.co.il; r=attinasi; sr=erik; implement bidi
...
(bidirectional text) for languages such as Arabic and Hebrew; from IBM bidi
project; not part of build yet (due to #ifdef IBMBIDI)
2001-03-09 03:29:00 +00:00
5496fd604a
Remove attempted workaround for CantRenderReplacedElement crash that didn't fix the crash (which is now fixed the right way).
...
Clean up the FrameManager a bit, mainly by moving the destruction of the frame tree into the |Destroy| method rather than doing it in the destructor. This would make it safer to convert nsIFrameManager to IDL and use the frame manager from JS. It also reduces the risk of leaks caused by having something owned by a frame own the frame manager. Breaking cycles in a destructor can be risky because the objects entrained by the cycle could include the object whose destructor would break the cycle.
b=65800 r=waterson@netscape.com sr=attinasi@netscape.com
2001-03-06 01:46:03 +00:00
7efd062521
Bug 62713: Only save scroll position in SH if offset from origin, only put entries in state table if state was created, cleanups, r=jst@netscape.com, sr=vidur@netscape.com
2001-02-06 02:32:04 +00:00
45c08e6635
Fix for 55292, r=ben, sr=brendan
2001-02-02 00:54:47 +00:00
6a1cef6f9a
Remove trailing semicolons after all uses of MOZ_DECL_CTOR_COUNTER to quiet build warnings on Mac. b=60415 a=brendan@mozilla.org
2001-01-27 23:06:33 +00:00
6a93e50850
bug 56432 ()
...
the problem was a factorial algorithm for finding primary frames.
fixed by added a hinting mechanism that takes advantage of existing
primary frame cache, greatly reducing the cost for the common case.
r=rods
sr=waterson
2001-01-16 22:10:04 +00:00
edb0c6270c
removing most all NS_DEFINE_IID and replacing them with NS_GET_IID macros
...
Bug 61274, r=buster sr=waterson
2001-01-04 20:44:42 +00:00
09cf5ac2fb
Fixes regression due to Style Data sharing: forcing shared contexts to be unique now when they have to be reresolved. b=63118 r=karnaze@netscape.com a=buster@netscape.com
2000-12-21 22:16:38 +00:00
3f9e12209c
Fix insertion point reparentin gissue with blocks. r=hixie, a=ben
2000-11-28 10:10:41 +00:00
5af6f6e116
Fix for 51264. r=danm, a=ben
2000-11-27 22:00:33 +00:00
cd56c0575b
Bug 47207. Backing out logging/PRINTF changes until we can fix stopwatch.h, introduce double parens, etc.
2000-10-28 22:17:53 +00:00
9a6d92a433
Bug 47207. Changing printf to PRINTF to use new logging facility. r=valeski,sr=waterson
2000-10-27 22:43:51 +00:00
1fd5ba1fd2
Added default implementation for frames to allow a frame to specify which frame provides its parent style context. Also reworked ReResolveStyleContext and VerifyStyleTree to allow the style context tree to diverge from the frame tree. b=45210 r=karnaze
2000-09-12 22:47:09 +00:00
84f381ec78
Fix blocker bug #52039 (crash when resolving pseudo-style for anonymous content).
2000-09-11 03:58:22 +00:00
7aa7ed7f6d
Fixing nsbeta3+ bug 28277. Add a callback to nsHTMLImageElement so that we know when a image created in JavaScript has loaded (or fails to load), this is needed to support onload and onerror event fireing and also to support the '.complete' property on images create in JavaScript. r=nisheeth@netscape.com
2000-09-01 07:05:08 +00:00
3f1d49a015
Bug 32770: Save scroll position in session history r=nisheeth
2000-06-15 00:35:46 +00:00
d264987b1c
Removed erroneous assertion: b=38768 r=dcone
2000-06-07 23:05:30 +00:00
e1de241d59
bug 38665
...
r=rickg a=rickg
fixed an optimized-only crash that looks like a race condition
where we send a message containing a pointer to a frame back to
the frame manager after the frames have been deleted.
2000-06-05 03:37:33 +00:00
dbaron%fas.harvard.edu
hyatt%netscape.com
bryner%netscape.com
jst%netscape.com
alecf%netscape.com
dougt%netscape.com
vidur%netscape.com
waterson%netscape.com
pollmann%netscape.com
cls%seawood.org
pavlov%netscape.com
erik%netscape.com
buster%netscape.com
rods%netscape.com
attinasi%netscape.com
warren%netscape.com
pierre%netscape.com