514751 - malformed URI exception during session restore

Status: RESOLVED FIXED

(Firefox :: Session Restore, defect, P1)

Description

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Saw this in the error console when restoring a bunch of tabs:

Error: uncaught exception: [Exception... "Component returned failure code: 0x804b000a (NS_ERROR_MALFORMED_URI) [nsIIOService.newURI]"  nsresult: "0x804b000a (NS_ERROR_MALFORMED_URI)"  location: "JS frame ::
file:///C:/Program%20Files%20(x86)/Namoroka/components/nsSessionStore.js ::
          sss_deserializeHistoryEntry :: line 2130"  data: no]

Unfortunately, I didn't think to save my session store file :/

Comment 1

[Dietrich Ayala (:dietrich)]

note: vlad also said he lost his session, which makes this serious indeed.

this particular exception, if related to the session loss, should be easy enough to wallpaper, so that the rest of the session isn't lost when this happens.

brendan said he'd seen similar behavior, and only in the last few days. adding qawanted to get better STR and a regression range.

marking blocking for investigation, due to dataloss.

Comment 2

[Simon Bünzli]

Likely related: bug 509315 (SeaMonkey); bug 514370 (about:sessionrestore).

Comment 3

[Dietrich Ayala (:dietrich)]

Brendan's seeing this on a mac, changing OS to All.

Comment 4

[Paul O'Shannessy [:zpao] (not bugmail, email directly)]

Attachment: Wallpaper Patch v0.1

For some reason we're getting an extra history entry that's completely empty {}. The we try to add that to history by accessing it's .url, which doesn't work.

So this is the quick wallpaper. I'll look into why this is happening in the first place.

Comment 5

[Paul O'Shannessy [:zpao] (not bugmail, email directly)]

Attachment: Reduced Test Case

This is a reduced guilty sessionstore.js

Comment 6

[Dietrich Ayala (:dietrich)]

Comment on attachment 399321 [details] [diff] [review]
Wallpaper Patch v0.1

r=me for wallpaper to land asap. please followup w/ unit test.

Comment 7

[Paul O'Shannessy [:zpao] (not bugmail, email directly)]

Pushed http://hg.mozilla.org/mozilla-central/rev/1ad1932defe6

Will followup with a test & continue to look into why we're getting into this state in the first place.

Comment 8

[Paul O'Shannessy [:zpao] (not bugmail, email directly)]

Attachment: Wallpaper Patch Test v0.1

Followup test

Comment 9

[Dietrich Ayala (:dietrich)]

Comment on attachment 399530 [details] [diff] [review]
Wallpaper Patch Test v0.1

looks ok from here. r=me given you've confirmed it fails w/out the wallpaper patch.

Comment 10

[Paul O'Shannessy [:zpao] (not bugmail, email directly)]

Pushed the test in http://hg.mozilla.org/mozilla-central/rev/59f211597f4a.

So do we push these (as a single patch) to branch too or do we try to figure out the actual problem first? And how does that work with resolving the bug? Is this now fixed and we figure out the actual problem in a followup bug?

Sorry, I'm not sure how the whole wallpaper thing fits into our workflow.

Comment 11

[Pablo Greco]

Is this related to bug 467248 or bug 494917 ?

Comment 12

[Paul O'Shannessy [:zpao] (not bugmail, email directly)]

Pushed http://hg.mozilla.org/releases/mozilla-1.9.2/rev/ad284b242e98 and filed
bug 516073 to followup with the proper fix when we find it.

Comment 13

[Henrik Skupin [:whimboo][⌚️UTC+1]]

Do we still need a regression window? And is 1.9.1 affected too?

Comment 14

[Paul O'Shannessy [:zpao] (not bugmail, email directly)]

(In reply to comment #13)
> Do we still need a regression window? And is 1.9.1 affected too?

A little late here, but a regression window would be great. I feel like it's going to be hard to find though. That probably better belongs in bug 516073 though since that's the actual problem.

1.9.1 would be affected if the browser ever saved the empty history entry, but I haven't seen any reports of this.

Comment 15

[Ryan VanderMeulen [:RyanVM]]

Making the executive decision that finding a regression window here after almost 6 years is pretty unlikely to happen.