Closed Bug 980875 Opened 11 years ago Closed 8 years ago

Github JS breaks on Firefox nightly (NS_ERROR_FILE_CORRUPTED)

Categories

(Core :: Storage: IndexedDB, defect)

Product:
Core
Component:
Storage: IndexedDB
Version:
30 Branch
Platform:
x86_64
macOS
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED INCOMPLETE

People

(Reporter: seocam, Unassigned)

References

Details

I was trying to use Github today using Nightly and I've realized that do Javascript in the website was working. Checking the the console for JS errors I got the message: NS_ERROR_FILE_CORRUPTED: File error: Corrupted At first I thought this could be their mistake but them I've tested using Firefox 27.0.1 and Chrome 33.0.1750.146 and works perfectly on both. Follow my about:support Application Basics ------------------ Name: Firefox Version: 30.0a1 User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:30.0) Gecko/20100101 Firefox/30.0 Crash Reports for the Last 3 Days --------------------------------- All Crash Reports Extensions ---------- Name: Automatic Token Selection Version: 2.0 Enabled: false ID: AutoTokenSelection@morpho.com Name: Middleware ypsID Version: 1.0.2 Enabled: false ID: MorphoMWypsID@ypsid.morpho.com Important Modified Preferences ------------------------------ browser.cache.disk.capacity: 358400 browser.cache.disk.smart_size_cached_value: 358400 browser.cache.disk.smart_size.first_run: false browser.cache.disk.smart_size.use_old_max: false browser.places.smartBookmarksVersion: 6 browser.sessionstore.upgradeBackup.latestBuildID: 20140306030201 browser.startup.homepage_override.buildID: 20140306030201 browser.startup.homepage_override.mstone: 30.0a1 extensions.lastAppVersion: 30.0a1 network.cookie.prefsMigrated: true places.database.lastMaintenance: 1393263616 places.history.expiration.transient_current_max_pages: 104858 plugin.disable_full_page_plugin_for_types: application/pdf plugin.importedState: true privacy.sanitize.migrateFx3Prefs: true storage.vacuum.last.index: 1 storage.vacuum.last.places.sqlite: 1392291624 Graphics -------- Device ID: 0x 8a0 GPU Accelerated Windows: 1/1 OpenGL (OMTC) Vendor ID: 0x10de WebGL Renderer: NVIDIA Corporation -- NVIDIA GeForce 320M OpenGL Engine windowLayerManagerRemote: true AzureCanvasBackend: quartz AzureContentBackend: quartz AzureFallbackCanvasBackend: none AzureSkiaAccelerated: 0 JavaScript ---------- Incremental GC: true Accessibility ------------- Activated: false Prevent Accessibility: 0 Library Versions ---------------- NSPR Expected minimum version: 4.10.4 Beta Version in use: 4.10.4 Beta NSS Expected minimum version: 3.16 Basic ECC Beta Version in use: 3.16 Basic ECC Beta NSSSMIME Expected minimum version: 3.16 Basic ECC Beta Version in use: 3.16 Basic ECC Beta NSSSSL Expected minimum version: 3.16 Basic ECC Beta Version in use: 3.16 Basic ECC Beta NSSUTIL Expected minimum version: 3.16 Beta Version in use: 3.16 Beta
Sergio, what exactly were you trying to do on GitHub that didn't work? Can you please provide some steps to reproduce? Does the failure persist when you restart Firefox in safe-mode (with all add-ons disabled)?
Tim Taubert there is no Javascript working on Github. If I try to clone a repository for example I click on the organization where I want to clone it and nothing happens. Also if a send a comment it doesn't show up in the web page unless I reload it. Also if a try to delete a branch by clicking in the button "Delete Branch" after a merge nothing happens. I'm gonna try in safe mode.
Same problem in safe mode. My build ID: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:30.0) Gecko/20100101 Firefox/30.0 ID:20140306030201 CSet: 8122ffa9e1aa
I can't reproduce this on OS X 10.9. Navigating to https://github.com/mozilla/zippy and then forking it to my GH accounts works instantly.
I have this problem (or a very similar one) with Aurora 29.0a2 (2014-03-17). I see "NS_ERROR_FILE_CORRUPTED: File error: Corrupted" in the web console when I open any Github page, and I can reproduce the following symptoms: - When I comment on or close an issue, the button greys out, but the page doesn't update. Reloading the page shows that the comment was added / the issue was closed as intended. - When I open a project page (e.g. https://github.com/mozilla/addon-sdk) and click the "branch" dropdown menu, the menu opens but no branches or tags are listed. At least the latter symptom goes away if I try it from a private browsing window, so it seems like there is some bad state data that doesn't go away. Is there something I could try doing to further diagnose the problem?
I'm not getting this anymore.
I'm having this issue but it only starts appearing after a while of usage. If I quit and restart Firefox everything goes back to normal. Disabling all the extensions doesn't solve the issue either. So far I'm having issues on Github, Trello, Persona, Discourse My user agent is "Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:28.0) Gecko/20100101 Firefox/28.0"
Same issue
The issue remains after updating to 29.0.1. Is there anything we can do to help tracking the issue ?
Having the same issue in Github, Trello and a few other high profile sites. Using 29.0.1. Only started happening recently. Need to use Chrome in the meantime :(
(In reply to zimbatm from comment #9) > The issue remains after updating to 29.0.1. Is there anything we can do to > help tracking the issue ? Yes, it would be great to have some steps that are reproducible, if not always then at least most of the times. So far we unfortunately haven't been able to confirm the issue on our side although multiple users seem to be affected.
I tried finding time patterns but didn't find anything obvious. Here's a little script that writes to the local storage and updates the input with the current date every second, when it fails the input field is not updated anymore: http://zimbatm.com/lab/ff-localstorage-bug/ The input would work for half a day or just 20 min. I should have kept a strict record though. After having this test tool I installed the profilist extension ( https://addons.mozilla.org/en-US/firefox/addon/profilist/ ) and ran a second profile in parallel to see if it would freeze globally or not. The new profile never had an issue while the old profile kept breaking at random interval. In the end I destroyed my `~/Library/Application Support/Firefox` folder. Thanks to Firefox Sync I was able to restore all my settings and now things have been fine for the last couple of days.
I ran my regular profile in Nightly for a day or so and was getting this error consistently. The result is that sites like Slack don't work at all. If I switch back to FF Dev Edition the same profile works correctly. The feedback in this failure case is really bad - I didn't even realize it was localStorage related when I looked at this bug. It doesn't provide a stacktrace or a meaningful text message explaining what's corrupt, just the error code and a .js file name (which made me assume the javascript file was corrupt somehow...) Slack for example, from the developer console: 2016/1/22 01:22:40.754 session_load_count_client: 1 rollup-core_required_ts.js:1:2182 NS_ERROR_FILE_CORRUPTED: rollup-secondary_a_required.js:1:0 A failure case like this really needs better messaging so it's possible for a user to deal with it. Otherwise, the only solution to this is to blow away your entire profile and lose anything that wasn't synced. In this case, the profile is valid in FF Dev Edition so there is probably an actual bug here, but regardless, the error handling is *super bad*. Is there a specific file I can attach that will contain the relevant bits of localStorage goo so you can reproduce the problem?
This is now reliably reproducing in Developer Edition (I guess whatever's responsible rode the trains from Nightly in the last day or two). As a result, YouTube video uploads don't work, Slack doesn't work, etc. NS_ERROR_FILE_CORRUPTED: www-upload.js:1780:0 TypeError: yt.www.metadataeditor is undefined <anonymous> upload:4563 upload:4563:5
I have also been having this problem for several months now. It feels like I have to go and delete the offline cache at least twice a day to work around this, with varied success. Currently on Firefox 50.1.0
Forgot to add: this is not just happening on Github, but seems to be any site that makes use of the indexdb/offline storage
These may be related to bug 1341070. Asking reporter if they are still experiencing this issue.
Component: General → DOM: IndexedDB
Flags: needinfo?(seocam)
Product: Firefox → Core
I've not heard back from the reporter as to if the bug is still occuring, closing.
Status: NEW → RESOLVED
Closed: 8 years ago
Flags: needinfo?(somethingrandom)
Flags: needinfo?(seocam)
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.