[spoiler]­[/spoiler]
-
[spoiler]
[/spoiler]<=There's an ­ there ...
IE can't handle the softness of that hyphen!
-
Something else missing?
-
Maybe it actually needs line breaks. Fixed!
-
Indeed.
It's also only when you scroll up to place it in view
-
Here's what I see on chrome. What did you expect to see? What do you see in IE now?
-
What do you see in IE now?
-
@boomzilla said:
What do you see in IE now?
A headache inducing border around the window?
-
Humorously enough, IE10 also crashes, but putting it into IE9 Browser and Document modes and the topic is fine.
Also opening it without the OP in view in IE10 normal was fine, but when I scrolled up to the OP I got:
Hitting "Don't Allow" crashed the tab. Hitting "Allow" froze IE for a minute or so, then crashed the tab.
-
... is dw20.exe?!
-
Dr Watson; it's some sort of diagnostic and crash reporting tool MS package in Windows
-
what is......
#HOW‽‽
-
dw20.exe
Then it already crashed, windows however is still trying to decide that and/or create a crashdump
-
Oh fuck me.
Accidentally hit edit instead of reply
-
Does IE run tabs in their own processes? Could be how the parent process interprets Dr Watson's appearance.
-
Does IE run tabs in their own processes?
Yes.
Accidentally hit edit instead of reply
TL4ABUSE
-
Doesn't help that the editor is initially empty when you select edit ...
-
-
I think it's more Discourse slowness, and not with editing, but with populating the editor in general.
I've had points where hitting Edit gives me a blank post, but right now it's showing content fine when I click Edit. I've also had points where clicking "Quote Reply" gives me a blank editor for awhile too, which I attributed to server cooties.
-
That is...a very impressive level of fail on the part of IE. Is the topic that triggers this in a place visible to random browsing?
-
Is the topic that triggers this
You mean this very topic we're in right now?
in a place visible to random browsing
No lock or key icon on the category, so yes, this topic can be viewed by the world.
-
Yes. If you scroll to the OP of this thread, your IE tab will crash.
-
Confirmed on IE 11.0.17 on W7 x64.
More info: crash is a NULL pointer deref, apparently due to
this
beingNULL
in the call toCDXRenderTarget::FillGeometry()
(it looks to be a COM interface implementation internal tomshtml.dll
which makes it__stdcall
instead of__thiscall
). Backtrace:[code]
mshtml.dll!CDXRenderTarget::FillGeometry() Unknown
mshtml.dll!CDispSurface::DrawCharactersWithFillAndStroke(class CPoint const &,unsigned short const *,int,long const *,class CBaseCcs const &,enum IDispFont::DominantBaseline,class CRect const *,bool,class IDispBrush const *,float,class IDispBrush const *,struct IDispStrokeStyle const *) Unknown
mshtml.dll!CTextRenderList::RenderTextShadowLayers() Unknown
mshtml.dll!CTextRenderList::Render(class CTextShadow const *) Unknown
mshtml.dll!Layout::LineBoxBuilder::FlushShadowRendering(int) Unknown
mshtml.dll!Layout::LineBoxBuilder::DrawTextRun() Unknown
mshtml.dll!Layout::LineBoxBuilder::DrawTextRunsTogether() Unknown
mshtml.dll!Ptls6::DrawCharRange() Unknown
mshtml.dll!Ptls6::LsDisplayTextChunk(long,struct Ptls6::lschnke *,struct Ptls6::tagLSPOINT const *,unsigned int,unsigned long,struct Ptls6::tagLSRECT const *,int,struct Ptls6::LSDEVRES const &) Unknown
mshtml.dll!Ptls6::LsDisplaySublineCore() Unknown
mshtml.dll!Ptls6::LsDisplayLine() Unknown
mshtml.dll!Layout::LineBox::Draw() Unknown
mshtml.dll!Layout::FlowBox::DrawFlowItems() Unknown
mshtml.dll!Layout::FlowBox::DrawClientContainerContent() Unknown
mshtml.dll!CDispContainer::DrawChildrenInActiveLayer() Unknown
mshtml.dll!CDispNode::DrawInternal() Unknown
mshtml.dll!CDispNode::Draw() Unknown
mshtml.dll!CDispContainer::DrawChildren() Unknown
mshtml.dll!CDispContainer::DrawChildrenInActiveLayer() Unknown
mshtml.dll!CDispNode::DrawInternal() Unknown
mshtml.dll!CDispNode::Draw() Unknown
mshtml.dll!CDispContainer::DrawSelfContentFullStackingContext() Unknown
mshtml.dll!CDispContainer::DrawSelfContent() Unknown
mshtml.dll!CDispContainer::DrawSelf() Unknown
mshtml.dll!CDispNode::DrawInternal() Unknown
mshtml.dll!CDispNode::Draw() Unknown
mshtml.dll!CDispContainer::DrawChildren() Unknown
mshtml.dll!CDispContainer::DrawSelfContentFullStackingContext() Unknown
mshtml.dll!CDispContainer::DrawSelfContent() Unknown
mshtml.dll!CDispContainer::DrawSelf() Unknown
mshtml.dll!CDispNode::DrawInternal() Unknown
mshtml.dll!CDispNode::Draw() Unknown
mshtml.dll!CDispProxyNode::DrawSelf() Unknown
mshtml.dll!CDispProxyNode::Draw() Unknown
mshtml.dll!CDispNode::DrawIfNotProxied(class CDispDrawContext *,class CDispNode *,long,bool) Unknown
mshtml.dll!CDispParentNode::DrawChildrenInActiveLayer() Unknown
mshtml.dll!CDispStructureNode::Draw(class CDispDrawContext *,class CDispNode *,long,bool) Unknown
mshtml.dll!CDispNode::DrawIfNotProxied(class CDispDrawContext *,class CDispNode *,long,bool) Unknown
mshtml.dll!CDispParentNode::DrawChildrenInActiveLayer() Unknown
mshtml.dll!CDispStructureNode::Draw(class CDispDrawContext *,class CDispNode *,long,bool) Unknown
mshtml.dll!CDispContainer::DrawChildren() Unknown
mshtml.dll!CDispContainer::DrawSelfContentFullStackingContext() Unknown
mshtml.dll!CDispContainer::DrawSelfContent() Unknown
mshtml.dll!CDispContainer::DrawSelf() Unknown
mshtml.dll!CDispNode::DrawInternal() Unknown
mshtml.dll!CDispNode::Draw() Unknown
mshtml.dll!CDispContainer::DrawChildren() Unknown
mshtml.dll!CDispContainer::DrawSelfContentFlow(class CDispDrawContext *,class CDispInfo &,class CDispNode * *) Unknown
mshtml.dll!CDispContainer::DrawSelfContent() Unknown
mshtml.dll!CDispContainer::DrawSelf() Unknown
mshtml.dll!CDispNode::DrawInternal() Unknown
mshtml.dll!CDispNode::Draw() Unknown
mshtml.dll!CDispContainer::DrawChildren() Unknown
mshtml.dll!CDispContainer::DrawSelfContentFlow(class CDispDrawContext *,class CDispInfo &,class CDispNode * *) Unknown
mshtml.dll!CDispContainer::DrawSelfContent() Unknown
mshtml.dll!CDispContainer::DrawSelf() Unknown
mshtml.dll!CDispNode::DrawInternal() Unknown
mshtml.dll!CDispNode::Draw() Unknown
mshtml.dll!CDispRoot::DrawEntire() Unknown
mshtml.dll!CDispRoot::DrawRoot() Unknown
mshtml.dll!CPaintHandler::RenderInternal() Unknown
mshtml.dll!CPaintController::Render(void *,class CRenderState const &,unsigned int,bool &,bool &,unsigned long &) Unknown
mshtml.dll!CRenderTaskDrawInPlace::Execute() Unknown
mshtml.dll!CRenderTaskQueue::AddRenderTask() Unknown
mshtml.dll!CView::RenderInPlace() Unknown
mshtml.dll!CDoc::PaintWorker(bool,bool) Unknown
mshtml.dll!CDoc::PaintInPlace() Unknown
mshtml.dll!CPaintController::RunRenderingLoop() Unknown
mshtml.dll!CPaintController::OnUpdateBeat() Unknown
mshtml.dll!CPaintBeat::OnBeat() Unknown
mshtml.dll!CPaintBeat::OnVSyncMethodCall() Unknown
mshtml.dll!GlobalWndOnPaintPriorityMethodCall(void) Unknown
mshtml.dll!GlobalWndProc(struct HWND__ *,unsigned int,unsigned int,long) Unknown
user32.dll!_InternalCallWinProc@20() Unknown
user32.dll!_UserCallWinProcCheckWow@32() Unknown
user32.dll!_DispatchClientMessage@24() Unknown
user32.dll!___fnDWORD@4() Unknown
ntdll.dll!_KiUserCallbackDispatcher@12() Unknown
user32.dll!_NtUserDispatchMessage@4() Unknown
user32.dll!_DispatchMessageWorker@8() Unknown
user32.dll!_DispatchMessageW@4() Unknown
ieframe.dll!CTabWindow::_TabWindowThreadProc() Unknown
ieframe.dll!LCIETab_ThreadProc() Unknown
iertutil.dll!_IsoThreadProc_WrapperToReleaseScope() Unknown
IEShims.dll!NS_CreateThread::DesktopIE_ThreadProc(void *) Unknown
kernel32.dll!@BaseThreadInitThunk@12() Unknown
ntdll.dll!___RtlUserThreadStart@8() Unknown
ntdll.dll!__RtlUserThreadStart@8() Unknown
[/code]
-
IE 11.0.19, Win 8.1 x64
<img src="/uploads/default/original/3X/9/6/965416deb0d591f16901b2df351c3909493dc5ea.png" width="690" height="371"
EDIT
-
Anyone tried this with Edge?
-
Oh fuck me.
OK ...
I mean no problem I was wondering what I posted that required a spoiler
-
Does IE run tabs in their own processes?
Yes only the tdwtf tab crashed. SharePoint kept running.
-
Followup:
CDispSurface::DrawCharactersWithFillAndStroke
pushes a bunch of args on the stack, and then callsCDBindMethodsSimple::~CDBindMethodsSimple() {}
(yes, it's a no-op dtor) right before callingCDXRenderTarget::FillGeometry()
.However, even though what looks like a valid
this
ptr from the caller's side is being pushed onto the stack right before the two calls -- it never makes it to the callee, who instead proceeds to read the word above it on the stack frame, which is aNULL
from a local variable pushed by the caller, and die. This all smells of two subroutines which can't agree on parameter passing (either mismatched argument lists, or worse yet, a calling convention mismatch), but how could that sort of bug slip past the gates of Redmond?
-
-
I know!
-
That's a strange use of shy; it's supposed to be used in words to indicate the location a hyphen should be inserted if a word break is needed at EOL.
verylongveryverylong verylongveryverylong verylongveryverylong verylongveryverylong verylongveryv straight
verylongveryverylong verylongveryverylong verylongveryverylong verylongveryverylong verylongveryve straight
(Having some problems with baking. Finally got an example.)
Notice that where it appeared, the hyphen was put between g and h or between t and r...both very odd places to hyphenate the single-syllable straight. Both words have indicated shy points: st-raig-ht.
It may be the unusual solitary use that's causing it.
-
It may be the unusual solitary use that's causing it.
It might be awful and semantically meaningless, but it's still a legal sequence in HTML. It shouldn't make the browser's renderer keel over!
-
It might be awful and semantically meaningless, but it's still a legal sequence in HTML. It shouldn't make the browser's renderer keel over!
Oh, sorry, didn't mean to imply otherwise. I was just discussing thoughts on why they might have overlooked such an egregious fail.
Browser must not fail even when input violates every sensible rule.
We had a discussion about another aspect of that--what happened when you nest tags to more than 150 levels or something like that. While I argued that ignoring levels was acceptable, my reasoning was because it was better to ignore levels than to fail outright.
Otherwise you get user panics, "Oh noes! I've been hacked! My browser failed!"
-
Browser must not fail even when input violates every sensible rule.
There's fail and then there's fail. Crashing comes into the latter category, especially when we're talking a relatively simple combination of HTML, an (unusual) entity, and CSS. Just not drawing it at all (or, conversely, always drawing it) would be more in the former category.
-
I've also had points where clicking "Quote Reply" gives me a blank editor for awhile too, which I attributed to server cooties.
That's better attributed to quoting requiring a server round-trip. I'm pretty sure the UI makes a ton of unnecessary round-trips to the server.
Good work on finding an IE bug! Is it reported yet?
Oh yeah, Microsoft doesn't have a public bug tracker.
-
quoting requiring a server round-trip
Don't we already have both the raw and the cooked ? Otherwise the "view raw" wouldn't work, would it ?
-
No, that round-trips.
-
ah, i see. It seems to do so way faster than the edit text appears tho.
-
Probably because it's fetching the whole post.
https://what.thedailywtf.com/posts/460872
-
Good work on finding an IE bug! Is it reported yet?
Isn't IE discontinued?
-
-
@ChaosTheEternal said:
I've also had points where clicking "Quote Reply" gives me a blank editor for awhile too, which I attributed to server cooties.
That's better attributed to quoting requiring a server round-trip.
And the round trip sucks because of cooties.
-
Dr Watson; it's some sort of diagnostic and crash reporting tool MS package in Windows
Wait, this thing still exists? I thought it got superseded by Error Reporting in like XP...
For the record: Modern IE blanks the page, then loads the bottom of the topic again.
-
For the record:
ModernIE Edge blanks the page, then loads the bottom of the topic again.FTFY?
-
He's probably talking about Windows 8 IE
-
Yep, I'm one of those two guys that actually like that browser.
-
Wouldn't surprise me if it is a new thing, but the EXE name is the same