I have a Safari App Extension which allows users to switch between last open tabs with a shortcut option+tab in the same way it's possible to switch between last open apps with command+tab. Here is how i do it: I inject a content script on all websites which has the only thing – key listener for option+tab presses. When a user presses option+tab, that keyboard listener detects it and sends a message to the Safari Handler. Then Safari Handler sends a message to the containing app and it shows a panel with last open tabs. This approach has a problem: it shows a message to a user in settings: "Can read sensitive info from web pages, including passwords..." Which is bad, because in reality i don't read passwords. If i remove SFSafariContentScript key in the Safari App Extension target's Info.plist, then this message about reading sensitive data disappears, but then i loose the ability to open the tabs panel. How can I open my app window with a shortcut without frightening a user? It's possible to listen to global key presses, but that would require a user to grant the app permissions of Accessibility (Privacy & Security) in macOS system settings, which also sounds shady. I know an app which does not require an Accessibility permission: https://apps.apple.com/ua/app/tabback-lite/id6469582909 and at the same time it does not tell a user about reading sensitive data in the extension settings. Here is my app: https://apps.apple.com/ua/app/tab-finder/id6741719894 It's open-source: https://github.com/kopyl/safari-tab-switcher

Failed to verify your identity. Try again.

Recently we started noticing in Safari v18.2 browser an unexpected useragent set for https://www.espn.com pages like - "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/110.0.0.0 Safari/537.36" (recognized as Chrome by most UA detection logic) and breaks our video playback in some scenarios. When digging into this we came across site specific quirks and the "Disable site-specific hacks" setting which fixed the playback functionality and set a more expected UA - "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/18.2 Safari/605.1.15". Is the unexpected UA being set somewhere in Safari/Webkit? Can this be removed so site functionality works without needing to find the "Disable site-specific hacks" setting?

Hi, I'm developing an extension and I need to debug console logs that are logged in the Service Worker. The worker is configured in the manifest and is generally working as expected: However, when I open the browser, go to any site, and open Develop -> Service Workers or Develop -> Web Extension Background Content it is not visible there, so I can't really access the logs: But then I noticed that if I go out of focus from the browser for some time (and probably let the SW die), it becomes visible and I can open it without an issue: So, a couple of questions: Why isn't it instantly accessible? The extension Service Worker dev tools should be accessible regardless of what is happening to the tab or the browser, even if the SW terminates. Why does it eventually appear under Web Extension Background Content instead of the Service Workers when it is in fact an SW?

I am using GTM in my project, and the OneTrust cookie consent script is loading in GTM. After ruling out iOS 18.3.1, I am unable to click on the OneTrust preference center since the GTM script is not loading. This issue i faced for safari browser

A DNR rule with lower priority is being applied before a DNR rule of higher priority on Safari. Specifically, a low-priority DNR block rule that matches a request is being applied before a high-priority DNR redirect rule that matches the same request, preventing the redirect from occurring. The only way to get the high-priority redirect rule to occur is to remove the DNR block rule. This does not occur on other browsers. I have already submitted a Feedback Assistant report about this bug: FB16535579 How to reproduce: Create/install a web extension on Safari with the declarativeNetRequest and declarativeNetRequestWithHostAccess permissions Open the Web Extension Background Content console and add a redirect rule with a high priority number. For example: await chrome.declarativeNetRequest.updateDynamicRules({addRules: [ {id: 5000, condition: {urlFilter: "||www.google-analytics.com*/ga.js", resourceTypes: ["script"], domainType: "thirdParty"}, priority: 80, action: {type: "redirect", redirect: {url: “http://www.apple.com/”}}} ]}) Add a block rule of lower priority for the same urlFilter: await chrome.declarativeNetRequest.updateDynamicRules({addRules: [ {id: 5001, condition: {urlFilter: "||www.google-analytics.com^", domainType: "thirdParty"}, priority: 1, action: {type: "block"}} ]}) Visit https://efforg.github.io/privacybadger-test-fixtures/html/ga_surrogate.html Check the network tab and see that neither a request to Google Analytics nor apple.com appear. This means that the request to Google Analytics was blocked instead of being / before being redirected Remove the block rule: await chrome.declarativeNetRequest.updateDynamicRules({removeRuleIds: [5001]}) Reload https://efforg.github.io/privacybadger-test-fixtures/html/ga_surrogate.html. Check the network tab and confirm that there is a request to apple.com, showing that the redirect rule is only applied if the lower-priority block rule is removed. The priority of the DNR rules should handle this without having to remove a DNR rule. I have confirmed that the incorrect application of DNR rule priority happens on other top level domains, with other urlFilters, and with other redirect URLs. I confirmed that this is happening while I’ve granted my extension permissions on all websites.

I am posting here because we have an urgent issue affecting the operation of our service and are in need of a solution after our own analysis has come up with few answers. Beginning in iOS 18.2.x, we experienced exactly the same issue as the author of this thread, as we are also operating a service that allows for device certificate login for users configured to require one: https://developer.apple.com/forums/thread/767374 The author seems to have resolved the issue but the fix mentioned in the thread did not resolve our problem for iOS devices with iOS 18.2.x installed and the contents of that private support ticket are, of course, not visible to us. Furthermore, we have a different issue that surfaced with the release of iOS 18.3.x. Namely, the issue in iOS 18.3.x is more severe than the one in iOS 18.2.x, in that instead of simply taking a long time for the certificate/identity selection dialog to appear, it simply fails immediately and is returning a “no certificate selected” response to our server. One thing to note here is that, curiously, if we wait for several seconds (about 10-15 seconds) this behavior is not replicated. So, it seems there is potentially something going in the background, and the certificate selection process will only occur successfully like before if we wait. This is a very unideal workaround. After entering user credentials, we have the user navigate to a dedicated certificate authentication page. On the BIG IP side, upon users visiting this page, we have it configured to apply an SSL profile that contains appropriate CAs for the given user, and then requests to the browser that a new connection requiring a certificate be made. We are investigating this by checking logs in in a variety of places: We can verify in BigIP logs that a response is being returned to the server without a certificate included. For the sake of our application, this is handled as a “user did not select a certificate” event, and thus the attempted login is failed. Using the MacOS “Console” application, we are able to see the following logs from the “trustd” process of the target iOS 18.3.x device: Failure case: debug 11:19:49.648581+0900 trustd XPC [com.apple.WebKit[1034]/1#25 LF=0] operation: trust_evaluate (8) debug 11:19:49.648766+0900 trustd complex trust settings anchor Successful case (after waiting 10-15 seconds after initial login page load/before moving to certificate page): debug 11:26:02.803153+0900 trustd XPC [MobileSafari[1031]/1#169 LF=0] operation: trust_evaluate (8) debug 11:26:02.804219+0900 trustd non ev score: 121 <private> There appears to be no attempt by MobileSafari to initiate the display of a certificate selection window in the failure cases. The iOS device is swift to return a response with no certificate selected to Big IP, and the result of “no certificate selected” is thus propagated through Big IP and ultimately to our web service. Does anyone have any advice or information on the following? Recommended tools to gather more data that may be pertinent. Any ideas on changes in iOS 18.2.x+ that could have resulted in the behavior changing as described above? If more information is necessary, I will do my best to supply it. Thank you in advance!

We are currently running a lightweight server within our iOS mobile app to pass a unique device ID via localhost for device-based restrictions. The setup works by binding a user's email to their device ID upon login, and later, when they attempt to log in via a browser, we retrieve this ID by making a request to http://localhost:8086/device-info. However, we're encountering an issue when making this request. Here’s the error message: Error fetching device info: TypeError { } r@webkit-masked-url://hidden/:27:166011 value@webkit-masked-url://hidden/:27:182883 @webkit-masked-url://hidden/:27:184904 We are making this request from an HTTPS website, and we suspect this could be related to mixed-content restrictions. Could you guide us on how to properly make localhost requests over HTTPS, especially in a production environment with the necessary security measures? Any insights or best practices on resolving this issue would be greatly appreciated.

I reported this bug one year ago in https://developer.apple.com/forums/thread/746406, but as it is not been fixed yet, I'm going to try by opening this new incident report. iOS is not working for the Unicode Variation Selector-15 (U+FE0E) for all the characters. Can you please apply that variation selector to all your Unicode characters? I) Steps to reproduce the issue: navigate in safari to the page https://eurovot.com/vs.htm II) Expected result: as the 1st column of characters have the Variation Selector-15 (U+FE0E) applied, and the 2nd column have the Variation Selector-16 (U+FE0F) applied, the first column should always display text characters (in orange) and the second column emoji characters. III) Error result: some characters are working fine in the 1st column and displayed as text (in orange colour), but some other aren't displayed as text, they wrongly displayed as emojis instead.

I'm referring to the following: https://developer.apple.com/documentation/automaticassessmentconfiguration I can't find this information anywhere. I'm wondering if I can enable and disable this function in Safari or other browsers on my website.

I'm posting a question here as I have encountered an issue while seeking help from engineers in the thread. thread773837 If the "Not Secure Connection Warnings" is enabled in Settings > App > Safari, are HTTP connections not allowed under any circumstances? I also posted a question about NSAllowsLocalNetworking not being applied, and I was informed that ATS (App Transport Security) is not related to SFSafariViewController. If that's the case, what feature causes the error "Safari cannot open the page. Error: Failed to navigate to an HTTP URL with HTTPS-only mode enabled"? I am currently working to resolve this issue.

Hi, now we could try WebGPU by manually enabling it in feature flags in no matter Safari or Safari technology preview on macOS. But, do we know when this WebGPU feature would be enabled by default or any plan to enable it by default in Safari? Thanks!

We are building a Safari web extension utilising native messaging, to send messages to the Swift native part of the app. We sometimes experience, that the beginRequest handler is executed multiple times, at the same time. We have a special part of the code in the handler, that must be run only once. Because it uses NS defaults storage, and also because it calls our servers. We have tried to use a serial dispatch queue, as well as other locking and mutex techniques, to no success. We suspect that the instances of the handler are isolated in a way, that these locks don’t work (maybe they don’t share memory?). But we are not sure. When looking at os_logs from the handlers, they all share the same PID. Has anyone experienced anything similar and can shed some light on what's going on?

Hi all, Chrome has it already - animation-timeline aka scroll-animations. I can nowhere find any informations on what's the status in Safari/Webkit. Seems like they do not have it on the agenda at all? Does anyone know anything - I wanted to push a feature request for that - but also seem there is no feature request list anymore for webkit. See: https://www.w3.org/TR/scroll-animations/ Cheers and kind regards!

How can i add Authorization header to a wkwebview. I checked https://developer.apple.com/documentation/foundation/nsurlrequest#1776617 which says Authorization header is a reserved http header and shouldn’t be set. I want to set it when requesting a url to the server which will be used for verification. How can i do that?

I'm not sure if this is even the right place to ask this, or voice this complaint if there's no fix. If not, please direct me where to go. I'm trying to switch all my development over to Safari, but the tab sidebar is driving me absolutely insane. Let's say I have a bunch of tab groups expanded, and I collapse the sidebar to focus on the webapp I am building. When I toggle to show the sidebar again, every single tab group is collapsed. Why is it 2 clicks to see the tabs in the sidebar? Even with zero tab groups, and just tabs in the workspace, if you collapse the sidebar it'll collapse the list of tabs. Is there some hidden setting for this? Who do we have to complain to?

Hi, We have a native application with a webview part. When we open the app, we load webview resources in a false foreground. The webview communicates with the native part to query local variables, among other things. We have noticed that for the past few weeks, when the biometric component is presented to enter the fingerprint or faceId, while the component is displayed, the webview does not respond. We do not see the calls made by the webview, nor can we respond to them. Even if you open the Safari debugger, the webview does not respond UNTIL you close the biometric component in the native part. This has created a big problem in our app, and we wanted to know if it is an issue you had on your radar. Thank you very much.

I can open an app (e.g. tiktok) from a website of mine by using the deeplink “tiktok://open”, if the app is installed on my iphone. What's the deeplink to launch TestFlight? Thanks in advance

Hello! I've made a Safari extension that supports command "ReTab", and a couple of month ago, adding a customized macOS shortcut for Safari with menu title "ReTab" did trigger the extension. However, it's not working anymore and I'm not sure if it's from macOS/Safari update or because I changed manifest from v2 to v3 - could you help check if there's anything wrong with either the manifest.json or background.js? (the default Cmd+E still works) Thank you in advance! Xun manifest.json: { "manifest_version": 3, "default_locale": "en", "name": "ReTab", "description": "Go to the last active tab with Cmd+E!", "version": "1.4", "homepage_url": "https://LycheeIsle.com", "background": { "service_worker": "background.js" }, "action": { "default_icon": "images/toolbar-icon.svg" }, "permissions": [ "commands", "tabs", "storage" ], "commands": { "ReTab": { "suggested_key": { "default": "Command+E" }, "description": "Go to the last active tab" } }, "options_page": "options.html" } in background.js, I have this line which should listen to the command, and Cmd+E works but any customized shortcut for "ReTab" in Safari doesn't: browser.commands.onCommand.addListener(async (command) => { if (command === "ReTab" || command === "retab") { await retab() } });

Safari 18.0.1 on macOS 15.01 doesn't support the Passkey PRF extension during cross-device WebAuthn authentication when using QR code scanning, while it works correctly with iCloud passkeys. Steps to Reproduce: Clone and setup: git clone https://github.com/quocle108/passkey-prf-test yarn yarn start Test iCloud Passkey Flow: Open http://localhost:3000 in Safari Open DevTools (Cmd+Option+I) Click "Register" Choose "Passkey on iCloud" Expected console output: PRF supported: true Test Cross-Device Flow: Click "Register" Choose "Phone/Tablet" Scan QR with mobile device Expected: PRF supported: true PRF extension should be supported in cross-device flow, matching iCloud passkey behavior. Actual: PRF supported: false Cross-device flow returns empty extension results. Verify in Chrome Repeat steps 2-3 in Chrome Both flows return proper PRF extension results: PRF supported: true Test Environment: Browser: Safari 18.1.1 , Chrome 131.0.6778.70 OS: macOS 15.01 Mobile: iOS 18.x / Galaxy Note9 Android 10 Test repo: https://github.com/quocle108/passkey-prf-test