Open Bug 1472626 Opened 4 years ago Updated 2 years ago

Error: [Exception... "Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIFaviconService....

Categories

(Firefox :: Tabbed Browser, defect, P2)

defect

Tracking

()

Tracking Status
firefox63 --- wontfix

People

(Reporter: standard8, Unassigned)

References

(Blocks 2 open bugs, )

Details

(Keywords: steps-wanted, Whiteboard: [nightly-js-sentry:4426746])

This bug was automatically filed from Sentry: https://sentry.prod.mozaws.net/operations/nightly-js-errors/issues/4426746/

Error: [Exception... "Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIFaviconService.replaceFaviconDataFromDataURL]"  nsresult: "0x80004005 (NS_ERROR_FAILURE)"  location: "JS frame :: resource:///modules/PlacesUIUtils.jsm :: loadFavicon :: line 192"  data: no]
    at loadFavicon(resource:///modules/PlacesUIUtils.jsm:192:7)
    at loadFavicon(resource:///modules/PlacesUIUtils.jsm:324:5)
    at setIconFromLink(chrome://browser/content/browser.js:3769:7)
    at receiveMessage(chrome://browser/content/browser.js:3743:9)

This is new as of a couple of days ago, blame points to bug 1453751.
Flags: needinfo?(dtownsend)
Priority: -- → P2
We're catching this and reporting it so I guess it is just expected here. Not sure what else to do here or why it is showing up in sentry.
Flags: needinfo?(mkelly)
Flags: needinfo?(mak77)
Why are we reporting this if it's an expected exception?
Hm, what's failing exactly in replaceFaviconDataFromDataURL? the API is supposed to complain if it gets invalid input, otherwise we'd hide bugs. It's possible there's just a missing check in loadFavicon and we should not invoke replaceFaviconDataFromDataURL with some invalid input.
Flags: needinfo?(mak77)
Keywords: steps-wanted
And yes, if it's really expected we should probably analyze the exception and filter out the expected ones.
(In reply to Dave Townsend [:mossop] from comment #1)
> We're catching this and reporting it so I guess it is just expected here.
> Not sure what else to do here or why it is showing up in sentry.

If it's expected and you want to log it just as a note without collecting, log it as a warning and it won't be collected[1].

[1] https://searchfox.org/mozilla-central/source/browser/modules/BrowserErrorReporter.jsm#249
Flags: needinfo?(mkelly)
(In reply to Marco Bonardo [::mak] from comment #3)
> Hm, what's failing exactly in replaceFaviconDataFromDataURL? the API is
> supposed to complain if it gets invalid input, otherwise we'd hide bugs.
> It's possible there's just a missing check in loadFavicon and we should not
> invoke replaceFaviconDataFromDataURL with some invalid input.

Difficult to say what is failing but we could add more information to the error log to see that in sentry. Looking at the C++ code it looks like we do various checks on the size of the decoded image that we can only really do in C++. We could also change the C++ result in that case to be something more descriptive, maybe NS_ERROR_FILE_TOO_BIG.
Flags: needinfo?(dtownsend)

Hi,
I've encountered this behaviour on Windows 10 with Firefox Nightly 74.0a1 (2020-01-13) (64-bit) in the following page http://webglsamples.org/aquarium/aquarium.html

Regards.

Hi, thank you but there is no need to keep re-verifying this for releases. If it goes away that might be interesting, but we don't need comments to say this is still occurring.

Blocks: 120352
You need to log in before you can comment on or make changes to this bug.