User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; hu-HU; rv:1.9.2.13) Gecko/20101203 SUSE/3.6.13-3.1 Firefox/3.6.13 Build Identifier: Mozilla/5.0 (X11; U; Linux x86_64; hu-HU; rv:1.9.2.13) Gecko/20101203 SUSE/3.6.13-3.1 Firefox/3.6.13 myBrowser.docShell.QueryInterface(Components.interfaces.nsIDocShellHistory).useGlobalHistory = false; stopped working. It worked all right in 3.6 but doesn't stop using Global History in FF4 betas. Reproducible: Always Steps to Reproduce: 1. open a window from ff4 by an extension 2. define a browser with useGlobalHistory = false in the docShell 3. see FF's history. Actual Results: FF 4 GlobalHistory contains the unwanted URL's Expected Results: stop pushing URL-s into global history.

Confirming from newsgroup discussion and quick code perusal.

Indeed. This is a regression from the IHistory changes. Shawn, I think we need to fix this... It breaks embedding cases and extension cases of various sorts. Do you have time to take it, or should I?

(In reply to comment #2) > Indeed. This is a regression from the IHistory changes. Shawn, I think we > need to fix this... It breaks embedding cases and extension cases of various > sorts. Do you have time to take it, or should I? I won't have time until next week.

Shawn, have you had a chance to look at this yes. Can we figure out if it's a blocker or not so we can get it out of the nomination queue?

Estimating is hard. I likely won't have time to look at this until bug 606966 is resolved. bz - how bad does this break embedders?

It depends on the embeddor. If nothing else, this looks like it could make embeddors privacy features not work right, no? I guess I should just take this. I do think this should block. It's not hard to fix.

Alright, I can do the review then if that's OK with you (I'm not a peer, but you are the owner so...)

Comment on attachment 507025 [details] [diff] [review] Make useGlobalHistory work again on docshells. r=sdwilsh Should we care about a test here?

> Should we care about a test here? Probably, but I'd have to do some thinking to figure out how to write one...

(In reply to comment #10) > > Should we care about a test here? > > Probably, but I'd have to do some thinking to figure out how to write one... I can post my extension to you if it helps

No, I understand how to reproduce the problem in general. I'm just not sure how to write an automated test for it. "See history" doesn't work for an automated test... ;)

Pushed http://hg.mozilla.org/mozilla-central/rev/5e7903dbbd28

(In reply to comment #12) > No, I understand how to reproduce the problem in general. I'm just not sure > how to write an automated test for it. "See history" doesn't work for an > automated test... ;) Write a reftest that contains a link to itself. Compare two copies of the reftest (different URLs linking to the same backend page?), one running in a <browser type="content" disablehistory="true"> and one running in a <browser type="content">. The latter's link will be visited ;-)

> The latter's link will be visited ;-) Sadly, that won't work because there's no way to detect when the visited styling has been applied.

(In reply to comment #15) > Sadly, that won't work because there's no way to detect when the visited > styling has been applied. You could write a chrome test that opens different xul windows with the right attributes and then tests if things are visited with nsIGlobalHistory2.