171286 - Save As... crashes the second time [@ CHeaderSniffer::PerformSave ]

Status: RESOLVED WORKSFORME

(Core Graveyard :: Embedding: Mac, defect)

Description

[Dan Gilliam]

Steps to reproduce:

1. Go to a website
2. do a Save As... and click Save.
3. (important) close the browser window and open a new one 
4. Go to a website
5. do a Save As... and click Save.

result: crash

Reproduced in 1.0.2 PPEmbed 09/25/02 build.

Comment 1

[Olivier Cahagne]

can you post stack trace for this crash ?
http://developer.apple.com/qa/qa2001/qa1057.html

Comment 2

[Dan Gilliam]

relevant part of stack trace follows:
----------

Date/Time:  2002-09-27 17:16:42 -0400
OS Version: 10.1.5 (Build 5S66)
Host:       localhost

Command:    PPEmbed
PID:        622

Exception:  EXC_BAD_ACCESS (0x0001)
Codes:      KERN_PROTECTION_FAILURE (0x0002) at 0x00000008

Thread 0 Crashed:
#0   0x004f1bd0 in Init__9CDownloadFP6nsIURIP12nsILocalFilePCwPCwxP20nsIWebBrowse
#1   0x004f1bc0 in Init__9CDownloadFP6nsIURIP12nsILocalFilePCwPCwxP20nsIWebBrowse
#2   0x004fa74c in 0x4fa74c
#3   0x004f94f8 in CHeaderSniffer::PerformSave(nsIURI *, ESaveFormat)
#4   0x004f7f84 in CHeaderSniffer::OnStateChange(nsIWebProgress *, nsIRequest *)
#5   0x019fd000 in OnStartRequest__19nsWebBrowserPersistFP10nsIRequestP11nsISuppo
#6   0x01bf801c in nsHttpChannel::ProcessNormal(void)
#7   0x01bf7898 in nsHttpChannel::ProcessResponse(void)
#8   0x01c01480 in nsHttpChannel::OnStartRequest(nsIRequest *, nsISupports *)
#9   0x01bf2618 in nsOnStartRequestEvent::HandleEvent(void)
#10  0x01bf1990 in nsARequestObserverEvent::HandlePLEvent(PLEvent *)
#11  0x006719c0 in PL_HandleEvent
#12  0x0067182c in PL_ProcessPendingEvents
#13  0x00617d9c in nsEventQueueImpl::ProcessPendingEvents(void)
#14  0x01f50d1c in nsMacNSPREventQueueHandler::ProcessPLEventQueue(void)
#15  0x01f50bc0 in nsMacNSPREventQueueHandler::RepeatAction(EventRecord const &)
#16  0x000e8b14 in Repeater::DoRepeaters(EventRecord const &)
#17  0x004ee9f4 in CEmbedRepeater::SpendTime(EventRecord const &)
#18  0x0050953c in 0x50953c
#19  0x00504918 in 0x504918
#20  0x00504688 in 0x504688
#21  0x004e304c in main

Comment 3

[Olivier Cahagne]

Didn't find dupes, marking NEW.

Comment 4

[Greg K.]

Really marking NEW per comment 3.

Comment 5

[Simon Fraser [no longer active]]

Attachment: Patch

With the patch in bug 171745, the one additional change I had to make to fix
this was to fix the loop in CHeaderSniffer that looks for update or activate
events. For some reason, I get stuck in that loop (certainly while debugging).

Comment 6

[Michael Buckland]

jaime nominated this via email so marking adt1.0.2
I did not see r= or sr= but correct me if I am wrong.

Comment 7

[Jaime Rodriguez, Jr.]

This is a blocker for a major embedding customer, so I am marking it as
nsbeta1+/topembed+.

Comment 8

[Jaime Rodriguez, Jr.]

Michael: There is no nomination needed at this time, as we still need to get
reviews. Trust ME, we will nominate, and track down approvals, when ready. But.
if the ADT would like to pre-approve this one, that would be kewl. ;-)

Comment 9

[Mike Pinkerton (not reading bugmail)]

as stated in 171475, i think it needs to be |while|. what part of that patch
fixes this bug? i don't see it.

Comment 10

[Simon Fraser [no longer active]]

After fixing bug 171475, I was not able to reproduce this.

Comment 11

[Michael Buckland]

Discussed in team meeting.  Minusing for adt.  Please renominate if you can
reproduce.

Comment 12

[Conrad Carlen (not reading bugmail)]

Will try to repro.

Comment 13

[Conrad Carlen (not reading bugmail)]

Moving milestone.

Comment 14

[Conrad Carlen (not reading bugmail)]

-> Future. Focus is shifting from PowerPlant to Cocoa.

Comment 15

[Michael Buckland]

Discussed in topembed triage.  Topembed minusing.

Comment 16

[Juan José Mata]

After talking to Samir, removing bogus keyword and status whiteboard monikers.

Lowering to minor severity given that PPEmbed is no longer a priority.

This might not even be an issue anymore according to comment 10, regardless.

Comment 17

[Matthias Versen [:Matti]]

marking wontfix, the file is gone according to mxr.
( CHeaderSniffer.cpp )