Moving backwards and then forwards again results in previous pages in the history. Example: Go to www.mozilla.org Click on "At A Glance" Click Back Click on "Feedback" Look at the history in the "Go" menu: Feedback At A Glance mozilla.org It should be: Feedback mozilla.org The "Feedback" never got removed. Checking the "Go" history immediately after clicking on Back shows Feedback as the current page (checkmarked) even though the mozilla.org page is displayed.

confirmed with an 20min old CVS build on win2k

*** Bug 107128 has been marked as a duplicate of this bug. ***

B.14 Back, Forward - Click Back button twice and Forward button twice. This is a smoketest. Upping to blocker.

What seems to be happening here is that hitting back doesn't update our memory of our current position in session history. In other words, given a history A B C ^ (i.e, back dropdown shows B, then A, page C is displayed, and forward dropdown is empty) hitting back shows page B but we still think the history is what is above rather than what it should be, which is: A B C ^ (i.e., B is displayed, A is in the back dropdown, and C is in the forward dropdown) This means hitting back again also goes to page B, and going to a new page D yields the session history A B C D ^ instead of A B D ^

I have noticed that if I go back one or two pages, then hit reload (optionally with shift), it sometimes loads the page I was at before hitting back x times. This occured when I was awaiting review on a bug, I hit reload on the bug page and it kept loading the patch in another bug I had been looking at.

*** Bug 107255 has been marked as a duplicate of this bug. ***

*** Bug 107256 has been marked as a duplicate of this bug. ***

*** Bug 107258 has been marked as a duplicate of this bug. ***

I suspect this is a regression caused by rpott's checkin for bug 99627. nsDSURIContentListener::DoContent() overwrites the load type in docshell. AIUI it should be LOAD_HISTORY for navigating in session history.

Backing out rpotts fixes this for me. Reassigning. Commands used to back out: cvs update -j1.62 -j1.61 mozilla/security/manager/ssl/src/nsNSSComponent.cpp cvs update -j1.10 -j1.9 mozilla/xpfe/components/xremote/src/XRemoteService.cpp cvs update -j1.57 -j1.56 mozilla/mailnews/compose/src/nsURLFetcher.cpp cvs update -j1.65 -j1.64 mozilla/mailnews/base/src/nsMsgWindow.cpp cvs update -j1.40 -j1.39 mozilla/mailnews/absync/src/nsAbSyncPostEngine.cpp cvs update -j1.279 -j1.278 mozilla/editor/composer/src/nsEditorShell.cpp cvs update -j1.11 -j1.9 mozilla/xpfe/browser/resources/content/nsBrowserContentListener.js cvs update -j1.3 -j1.2 mozilla/xpfe/components/xremote/src/XRemoteContentListener.cpp cvs update -j1.90 -j1.89 mozilla/uriloader/exthandler/nsExternalHelperAppService.cpp cvs update -j1.18 -j1.17 mozilla/uriloader/base/nsURILoader.h cvs update -j1.89 -j1.88 mozilla/uriloader/base/nsURILoader.cpp cvs update -j1.25 -j1.24 mozilla/uriloader/base/nsIURILoader.idl cvs update -j1.21 -j1.20 mozilla/uriloader/base/nsIURIContentListener.idl cvs update -j1.10 -j1.9 mozilla/embedding/browser/photon/src/WebBrowserContainer.cpp cvs update -j1.10 -j1.9 mozilla/embedding/browser/gtk/src/EmbedContentListener.cpp cvs update -j1.47 -j1.46 mozilla/embedding/browser/activex/src/control/WebBrowserContainer.cpp cvs update -j1.110 -j1.109 mozilla/docshell/base/nsDocShell.h cvs update -j1.370 -j1.369 mozilla/docshell/base/nsDocShell.cpp cvs update -j1.49 -j1.48 mozilla/docshell/base/nsDSURIContentListener.cpp

We need to make sure that this patch doesn't cause http://bugzilla.mozilla.org/show_bug.cgi?id=47636 to re-occur!! -- rick

I've checked in the patch... because the regression from bug #99627 was SO bad!!

FIXED in 2001102903 on Win2k. Marking as such.

*** Bug 107394 has been marked as a duplicate of this bug. ***

*** Bug 107541 has been marked as a duplicate of this bug. ***

VERIFIED Fixed