Remove a few legacy prefs #228

Merged
celenity merged 78 commits from any1here/Phoenix:dev into dev 2026-01-24 02:19:29 +01:00
Contributor

I am going through old issues and found some prefs that are not used anymore.

toolkit.telemetry.coverage.opt-out -> toolkit.coverage.opt-out

lightweightThemes.getMoreURL

browser.newtabpage.activity-stream.feeds.section.topstories.options -> pocket is dead

I removed network.trr.allow-rfc1918 being set to true a while back

I am going through old issues and found some prefs that are not used anymore. toolkit.telemetry.coverage.opt-out -> toolkit.coverage.opt-out [lightweightThemes.getMoreURL](https://bugzilla.mozilla.org/show_bug.cgi?id=2003172) browser.newtabpage.activity-stream.feeds.section.topstories.options -> pocket is dead I removed network.trr.allow-rfc1918 being set to true a while back
celenity requested changes 2026-01-20 08:02:55 +01:00
Dismissed
celenity left a comment

Thanks for the PR!

It looks like browser.newtabpage.activity-stream.feeds.section.topstories.options is still used in ESR, so let's keep it for now, but add a [ESR] tag at the end, so that we can remember to remove it for next ESR cycle.

(In general, I don't really recommend or go out of my way to support ESR, but I don't mind keeping prefs around for it, so long as they don't cause issues for release/elsewhere).

We should also probably update the reference URL to https://searchfox.org/firefox-esr140/rev/8c69555d/browser/extensions/newtab/lib/ActivityStream.sys.mjs#187.

Everything else looks good, appreciate the contribution!

Thanks for the PR! It looks like [`browser.newtabpage.activity-stream.feeds.section.topstories.options` is still used in ESR](https://searchfox.org/firefox-esr140/rev/8c69555d/browser/extensions/newtab/lib/ActivityStream.sys.mjs#187), so let's keep it for now, but add a `[ESR]` tag at the end, so that we can remember to remove it for next ESR cycle. *(In general, I don't really recommend or go out of my way to support ESR, but I don't mind keeping prefs around for it, so long as they don't cause issues for release/elsewhere)*. We should also probably update the reference URL to `https://searchfox.org/firefox-esr140/rev/8c69555d/browser/extensions/newtab/lib/ActivityStream.sys.mjs#187`. Everything else looks good, appreciate the contribution!
Author
Contributor

@celenity wrote in #228 (comment):

It looks like browser.newtabpage.activity-stream.feeds.section.topstories.options is still used in ESR, so let's keep it for now, but add a [ESR] tag at the end, so that we can remember to remove it for next ESR cycle.

I can add it back, but since Pocket is dead, it should not be able to show any stories anymore.

Edit: Or did you mean the UI toggle?

@celenity wrote in https://codeberg.org/celenity/Phoenix/pulls/228#issuecomment-9999291: > It looks like [`browser.newtabpage.activity-stream.feeds.section.topstories.options` is still used in ESR](https://searchfox.org/firefox-esr140/rev/8c69555d/browser/extensions/newtab/lib/ActivityStream.sys.mjs#187), so let's keep it for now, but add a `[ESR]` tag at the end, so that we can remember to remove it for next ESR cycle. I can add it back, but since Pocket is dead, it should not be able to show any stories anymore. Edit: Or did you mean the UI toggle?
Author
Contributor

Hope you don't mind me adding a few more things while I go through Phoenix.

Hope you don't mind me adding a few more things while I go through Phoenix.
Author
Contributor

Not sure if I am reading the code wrong/missing something, but shouldn't intl.allow-insecure-text-input be OSX only?

Not sure if I am reading the code wrong/missing something, but shouldn't `intl.allow-insecure-text-input` be OSX only?
Author
Contributor

I just tested browser.preferences.experimental.hidden being set to true and it did not appear at all on first launch.

I just tested `browser.preferences.experimental.hidden` being set to `true` and it did not appear at all on first launch.
First-time contributor

there's a policy too, since v130.0.1. also the dependence on telemetry/studies (comment) no longer applies in version 146+ (ref -> should it be allowed to be enabled by the user?), but still does apply to ESR.


from the ticket:

Both Firefox Labs and one of either studies or telemetry must now be
disabled to fully disable Nimbus.
there's a [policy](https://mozilla.github.io/policy-templates/#usermessaging:~:text=FirefoxLabs) too, since v130.0.1. also the dependence on telemetry/studies (comment) no longer applies in version 146+ ([ref]( https://bugzilla.mozilla.org/show_bug.cgi?id=1972647) <del>-> should it be allowed to be enabled by the user?</del>), but _still does apply to ESR_. ___ from the ticket: ``` Both Firefox Labs and one of either studies or telemetry must now be disabled to fully disable Nimbus. ```
Author
Contributor

Since https://mozilla-ohttp.fastly-edge.com does not return anything anymore

Since https://mozilla-ohttp.fastly-edge.com does not return anything anymore
Author
Contributor

@degausser Nice to see you here too :)

@degausser Nice to see you here too :)
Owner

Thank you so much, great work here @any1here!


@any1here wrote in #228 (comment):

Not sure if I am reading the code wrong/missing something, but shouldn't intl.allow-insecure-text-input be OSX only?

You're right, good catch. I see you added the [OSX-ONLY] tag to the end of the intl.allow-insecure-text-input pref - also make sure to add it to the end of the title (/// Crash on insecure password input) - otherwise it'll just result in an empty /// Crash on insecure password input section appearing in the prefs for releases outside of OS X.

Thank you so much, great work here @any1here! ___ @any1here wrote in https://codeberg.org/celenity/Phoenix/pulls/228#issuecomment-10007358: > Not sure if I am reading the code wrong/missing something, but shouldn't `intl.allow-insecure-text-input` be OSX only? You're right, good catch. I see you added the `[OSX-ONLY]` tag to the end of the `intl.allow-insecure-text-input` pref - also make sure to add it to the end of the title *(`/// Crash on insecure password input`)* - otherwise it'll just result in an empty `/// Crash on insecure password input` section appearing in the prefs for releases outside of OS X.
Author
Contributor

Not sure about removing browser.search.widget.inNavBar. I am not able to find anything setting this. If I missed something, I will revert that commit.

Not sure about removing `browser.search.widget.inNavBar`. I am not able to find anything setting this. If I missed something, I will revert that commit.
First-time contributor

fyi, network.proxy.type != 5 prevents BITS usage for background updates (= via the scheduled task)

user_pref("app.update.background.previous.reasons", '["on Windows but cannot usually use BITS"]')

https://searchfox.org/firefox-main/source/toolkit/mozapps/update/BackgroundUpdate.sys.mjs#213-219
https://searchfox.org/firefox-main/source/toolkit/mozapps/update/UpdateService.sys.mjs#864-872

fyi, `network.proxy.type != 5` prevents [BITS](https://en.wikipedia.org/wiki/Background_Intelligent_Transfer_Service) usage for background updates (= via the scheduled task) `user_pref("app.update.background.previous.reasons", '["on Windows but cannot usually use BITS"]')` https://searchfox.org/firefox-main/source/toolkit/mozapps/update/BackgroundUpdate.sys.mjs#213-219 https://searchfox.org/firefox-main/source/toolkit/mozapps/update/UpdateService.sys.mjs#864-872
Author
Contributor

I made accessibility.uia.enable windows only, since it only has any meaningful impact there. No idea why they set it for other platforms.

I made accessibility.uia.enable windows only, since it only has any meaningful impact there. No idea why they set it for other platforms.
Author
Contributor

I added privacy.restrict3rdpartystorage.heuristic.recently_visited_time with value 0 for defense in depth. Normally it does nothing since privacy.restrict3rdpartystorage.heuristic.recently_visited is set to false

I added `privacy.restrict3rdpartystorage.heuristic.recently_visited_time `with value `0` for defense in depth. Normally it does nothing since `privacy.restrict3rdpartystorage.heuristic.recently_visited` is set to `false`
Author
Contributor

@degausser
Do you know if setting security.sandbox.content.shadow-stack.enabled to true causes any problems? I added it in the default false state.

@degausser Do you know if setting `security.sandbox.content.shadow-stack.enabled` to `true` causes any problems? I added it in the default false state.
First-time contributor

@any1here #110


also as said in the librewolf pull comments, network.http.referer.defaultPolicy.trackers*->1 breaks youtube embeds with error 153, e.g. itch.io homepage

https://developers.google.com/youtube/terms/required-minimum-functionality#embedded-player-api-client-identity:~:text=strict%2Dorigin%2Dwhen%2Dcross%2Dorigin

@any1here https://codeberg.org/celenity/Phoenix/issues/110 ___ also as said in the librewolf pull comments, `network.http.referer.defaultPolicy.trackers*->1` breaks youtube embeds with error 153, e.g. itch.io homepage https://developers.google.com/youtube/terms/required-minimum-functionality#embedded-player-api-client-identity:~:text=strict%2Dorigin%2Dwhen%2Dcross%2Dorigin
Author
Contributor

@degausser
I was planning on adding your recommendations after I had finished my initial scroll through of Phoenix. Currently I am just seeing what catches my eye.

@degausser I was planning on adding your recommendations after I had finished my initial scroll through of Phoenix. Currently I am just seeing what catches my eye.
Author
Contributor

I have finished my scroll through. Everything below this comment (except 1 2) come from suggestions @degausser made here librewolf/settings#107 (comment)

Or should this be a new PR?

I have finished my scroll through. Everything below this comment (except [1](https://codeberg.org/celenity/Phoenix/commit/547a4f6a83837e382b24185dae2aac334bcc54e6) [2](https://codeberg.org/celenity/Phoenix/commit/3baeec1ab995204e9fe5b43f262a4a87df1cce5b)) come from suggestions @degausser made here https://codeberg.org/librewolf/settings/pulls/107#issuecomment-9991563 Or should this be a new PR?
First-time contributor

@any1here native messaging is covered in the WebCompat wiki entry and was fairly recently closed as wontfix. i stand by what i wrote about it in the comment (disabled signons + keepassxc not working ootb is bad), but i don't think it's a wanted change in phoenix. (likely the same with jit, fpp, rs intervals...). for example disabling push is covered in Network Connections as "not recommended".

if desired, a new pr in general seems cleaner, to facilitate merging this one without hassle.

@celenity for phoenix, at least these could be considered:

// cloudflare turnstile gets intermittently stuck with this
--accessibility.blockautorefresh
--network.http.prompt-temp-redirect

// breaks YouTube embeds with error 153 (see this pr also)
--network.http.referer.defaultPolicy.trackers*

// this draws a hovering hand cursor over clickable links only after the whole page has loaded (there is no "hand with spinner" cursor, at least on Windows); bad concession
--browser.spin_cursor_while_busy

// no reason to disable extended taskbar shortcut menu on Windows
--browser.taskbar.lists*

// reenable alt key to show menu
--ui.key.menuAccessKeyFocuses

// reenable animated webpage loading thumbnail, instead of an hourglass
--ui.prefersReducedMotion

permissions.default.localhost->2 could at least get a wiki entry about this; one would already need a BlockLAN (on by default in assets.json) uBO list @@ scoped exception to get it working, and then allow the permission prompt.

btw, dom.text_fragments.create_text_fragment.enabled no longer exists in firefox-main, and the entry is in the context menu by default; it exists in ESR as default false.

@any1here native messaging is covered in the [WebCompat](https://codeberg.org/celenity/Phoenix/wiki/Web-Compat) wiki entry and was fairly recently closed as [wontfix](https://codeberg.org/celenity/Phoenix/issues/219). i stand by what i wrote about it in the comment (disabled signons + keepassxc not working ootb is bad), but i don't think it's a wanted change in phoenix. (likely the same with jit, fpp, rs intervals...). for example disabling push is covered in [Network Connections](https://codeberg.org/celenity/Phoenix/wiki/Network-Connections#push-service-https-mozilla-push-service-readthedocs-io-en-latest) as "not recommended". if desired, a new pr in general seems cleaner, to facilitate merging this one without hassle. @celenity for phoenix, at least these could be considered: ``` // cloudflare turnstile gets intermittently stuck with this --accessibility.blockautorefresh --network.http.prompt-temp-redirect // breaks YouTube embeds with error 153 (see this pr also) --network.http.referer.defaultPolicy.trackers* // this draws a hovering hand cursor over clickable links only after the whole page has loaded (there is no "hand with spinner" cursor, at least on Windows); bad concession --browser.spin_cursor_while_busy // no reason to disable extended taskbar shortcut menu on Windows --browser.taskbar.lists* // reenable alt key to show menu --ui.key.menuAccessKeyFocuses // reenable animated webpage loading thumbnail, instead of an hourglass --ui.prefersReducedMotion ``` `permissions.default.localhost->2` could at least get a wiki entry about [this](https://jdownloader.org/knowledge/wiki/glossary/cnl2); one would already need a BlockLAN (on by default in assets.json) uBO list @@ scoped exception to get it working, _and_ then allow the permission prompt. btw, `dom.text_fragments.create_text_fragment.enabled` no longer exists in firefox-main, and the entry is in the context menu by default; it exists in ESR as default false.
Author
Contributor

I will leave this PR as-is until @celenity can give input on what can be changed in Phoenix.

Edit: I reverted the last 2 commits regarding native-messaging

I will leave this PR as-is until @celenity can give input on what can be changed in Phoenix. Edit: I reverted the last 2 commits regarding native-messaging
celenity requested changes 2026-01-22 23:52:45 +01:00
Dismissed
celenity left a comment

Few minor changes, but overall looks good! Thank you very much again for your time and work here @any1here, very much appreciated.

Few minor changes, but overall looks good! Thank you very much again for your time and work here @any1here, very much appreciated.
@ -423,10 +423,6 @@ pref("doh-rollout.trrRace.randomSubdomainCount", 0); // [HIDDEN]
// https://firefox.settings.services.mozilla.com/v1/buckets/main/collections/moz-essential-domain-fallbacks/changeset?_expected=0
pref("network.essential_domains_fallback", false); // [DEFAULT]
/// Disable Fakespot
Owner

Despite Fakespot no longer operating, it looks like these prefs are still used for "private attribution", so I think we should keep them.

Despite Fakespot no longer operating, it looks like [these prefs are still used](https://searchfox.org/firefox-main/source/dom/privateattribution/PrivateAttributionService.sys.mjs) for "private attribution", so I think we should keep them.
Author
Contributor
I can add it back, but like I said before, since https://mozilla-ohttp.fastly-edge.com does not serve anything anymore, it does nothing https://searchfox.org/firefox-main/rev/3eaf7e2acf8186eb7aa579561eaa1312cb89132b/toolkit/components/dap/DAPSender.sys.mjs#292
Author
Contributor
https://codeberg.org/any1here/Phoenix/commit/6d06e21e4f12dec8b7a1c8b50ee7b873cb973c7f
@ -2945,0 +2930,4 @@
// 0: Never. [WINDOWS-ONLY]
// 1: Always. [WINDOWS-ONLY]
// 2: Enable unless incompatible accessibility clients are detected. (default) [WINDOWS-ONLY]
pref("accessibility.uia.enable", 0); // [WINDOWS-ONLY]
Owner

I agree with your conclusion that this pref only appears to have an impact on Windows, but since it is also defined for other platforms, I think we should still set it there to be safe. I'd replace [WINDOWS-ONLY] with [Windows] - that'll indicate the fact that it only takes effect on Windows, but will ensure it's still set for other platforms.

I agree with your conclusion that this pref only appears to have an impact on Windows, but since it is also defined for other platforms, I think we should still set it there to be safe. I'd replace `[WINDOWS-ONLY]` with `[Windows]` - that'll indicate the fact that it only takes effect on Windows, but will ensure it's still set for other platforms.
Author
Contributor

any1here/Phoenix@46f3067a5e I only added it after the pref. Hope that is fine

https://codeberg.org/any1here/Phoenix/commit/46f3067a5e7b91d901a1caa8f62b6aa8addbb079 I only added it after the pref. Hope that is fine
celenity marked this conversation as resolved
@ -3129,0 +3117,4 @@
// 0=no-referrer, 1=same-origin, 2=strict-origin-when-cross-origin (default),
// 3=no-referrer-when-downgrade.
// Setting to 1 currently breaks various functionality https://codeberg.org/celenity/Phoenix/pulls/228#issuecomment-10051167
pref("network.http.referer.defaultPolicy.trackers", 2); // [DEFAULT]
Owner

I don't love relaxing these, but the breakage it causes does seem problematic.

We could technically set urlclassifier.trackingAnnotationSkipURLs to https://youtube.com/* instead for the instance of YouTube breaking - but I don't think that's a good solution since I believe it would also result in other tracking protections being relaxed for YouTube.

So, until we can find a better solution, I agree with your approach - it seems like setting these to 2 is the best move.

I don't love relaxing these, but the breakage it causes does seem problematic. We could technically set `urlclassifier.trackingAnnotationSkipURLs` to `https://youtube.com/*` instead for the instance of YouTube breaking - but I don't think that's a good solution since I believe it would also result in other tracking protections being relaxed for YouTube. So, until we can find a better solution, I agree with your approach - it seems like setting these to `2` is the best move.
celenity marked this conversation as resolved
@ -3333,6 +3322,7 @@ pref("security.sandbox.gmp.shadow-stack.enabled", true); // [WINDOWS-ONLY] [DEFA
pref("security.sandbox.gpu.shadow-stack.enabled", true); // [WINDOWS-ONLY] [DEFAULT]
pref("security.sandbox.rdd.shadow-stack.enabled", true); // [WINDOWS-ONLY] [DEFAULT]
pref("security.sandbox.socket.shadow-stack.enabled", true); // [WINDOWS-ONLY] [DEFAULT]
pref("security.sandbox.content.shadow-stack.enabled", false); // [WINDOWS-ONLY] [DEFAULT]
Owner

I don't think we should explicitly set this to false, so that once it's ready and Mozilla decides to enable it by default upstream, it'll also be enabled for our users.

I think it's worth keeping this pref for reference though, but let's just comment it out, and maybe add a link to #110.

I don't think we should explicitly set this to false, so that once it's ready and Mozilla decides to enable it by default upstream, it'll also be enabled for our users. I think it's worth keeping this pref for reference though, but let's just comment it out, and maybe add a link to https://codeberg.org/celenity/Phoenix/issues/110.
Author
Contributor
https://codeberg.org/any1here/Phoenix/commit/af92ef45e90cf10fac127e30fcf72d4351964eb4
celenity marked this conversation as resolved
@ -3417,3 +3407,3 @@
/// Protect against MIME Exploits
// https://www.pcmag.com/encyclopedia/term/mime-exploit
pref("dom.workers.importScripts.enforceStrictMimeType", true); // [DEFAULT]
pref("dom.workers.importScripts.enforceStrictMimeType", true); // [DEFAULT] [EST]
Owner

Minor typo: I assume you meant [ESR].

Minor typo: I assume you meant `[ESR]`.
Author
Contributor

any1here/Phoenix@3eb28cb5a9 Nah, this is a timezone specific setting now

https://codeberg.org/any1here/Phoenix/commit/3eb28cb5a978fadc314a9fdc194b5ae6550cb253 Nah, this is a timezone specific setting now
@ -3819,3 +3800,2 @@
pref("network.dnsCacheExpiration", 3600); // (Default = 60)
pref("network.dnsCacheExpirationGracePeriod", 120); // (Default = 60)
pref("network.dnsCacheEntries", 10000); // (Default = 800)
pref("network.dnsCacheExpirationGracePeriod", 600); // (Default = 600)
Owner

TBH, I think we should just remove this pref (network.dnsCacheExpirationGracePeriod) - I don't see a point in explicitly setting the default there.

TBH, I think we should just remove this pref *(`network.dnsCacheExpirationGracePeriod`)* - I don't see a point in explicitly setting the default there.
Author
Contributor
https://codeberg.org/any1here/Phoenix/commit/4231c4cb68d85ee0795dc7490d0d443776b47e91
Owner

Thank you for your feedback @degausser! To address your points:

// cloudflare turnstile gets intermittently stuck with this
--accessibility.blockautorefresh
--network.http.prompt-temp-redirect

Do you mind sharing more details here - ex. can you provide an example of an impacted website (would help for testing)? I'm not saying that you're wrong, this may very well be an issue (and if so, those prefs can indeed be reconsidered. I'm just surprised to hear this as it's not a problem I've encountered or heard from other users).

// breaks YouTube embeds with error 153 (see this pr also)
--network.http.referer.defaultPolicy.trackers*

I don't like having to relax those prefs, but I agree that it unfortunately seems necessary, so I'm fine with setting those back to the default values (2 ) until we can find a better solution (This should be covered by @any1here's PR).

// this draws a hovering hand cursor over clickable links only after the whole page has loaded (there is no "hand with spinner" cursor, at least on Windows); bad concession
--browser.spin_cursor_while_busy

I see what you're saying - it appears to do this on at least macOS as well (I just tested it). Will have to think about it - I think it's overall a nice feature from a UX perspective, but I agree it's not ideal that it prevents the hand cursor from drawing unless the whole page has loaded.

// no reason to disable extended taskbar shortcut menu on Windows
--browser.taskbar.lists*

Fair enough, good catch. I'll enable browser.taskbar.lists.enabled and browser.taskbar.lists.tasks.enabled for next release - those seem both harmless and legitimately useful. I think we should probably keep browser.taskbar.lists.frequent.enabled and browser.taskbar.lists.recent.enabled though.

// reenable alt key to show menu
--ui.key.menuAccessKeyFocuses

You're right, will enable for next release. No reason to disable IMO.

// reenable animated webpage loading thumbnail, instead of an hourglass
--ui.prefersReducedMotion

Not sure how to feel about this one, curious how others feel.

permissions.default.localhost->2 could at least get a wiki entry about this; one would already need a BlockLAN (on by default in assets.json) uBO list @@ scoped exception to get it working, and then allow the permission prompt.

Great point, thanks for bringing this up. Will do some testing. Honestly, instead of removing permissions.default.localhost, I think I'd rather remove/disable the Block LAN uBo list - I agree we shouldn't set both, and I think it's better to rely on native browser features for functionality like this over extensions where possible.

Thank you for your feedback @degausser! To address your points: > // cloudflare turnstile gets intermittently stuck with this > --accessibility.blockautorefresh > --network.http.prompt-temp-redirect Do you mind sharing more details here - ex. can you provide an example of an impacted website *(would help for testing)*? I'm not saying that you're wrong, this may very well be an issue *(and if so, those prefs can indeed be reconsidered. I'm just surprised to hear this as it's not a problem I've encountered or heard from other users)*. > // breaks YouTube embeds with error 153 (see this pr also) > --network.http.referer.defaultPolicy.trackers* I don't like having to relax those prefs, but I agree that it unfortunately seems necessary, so I'm fine with setting those back to the default values *(`2` )* until we can find a better solution *(This should be covered by @any1here's PR)*. > // this draws a hovering hand cursor over clickable links only after the whole page has loaded (there is no "hand with spinner" cursor, at least on Windows); bad concession > --browser.spin_cursor_while_busy I see what you're saying - it appears to do this on at least macOS as well *(I just tested it)*. Will have to think about it - I think it's overall a nice feature from a UX perspective, but I agree it's not ideal that it prevents the hand cursor from drawing unless the whole page has loaded. > // no reason to disable extended taskbar shortcut menu on Windows > --browser.taskbar.lists* Fair enough, good catch. I'll enable `browser.taskbar.lists.enabled` and `browser.taskbar.lists.tasks.enabled` for next release - those seem both harmless and legitimately useful. I think we should probably keep `browser.taskbar.lists.frequent.enabled` and `browser.taskbar.lists.recent.enabled` though. > // reenable alt key to show menu > --ui.key.menuAccessKeyFocuses You're right, will enable for next release. No reason to disable IMO. > // reenable animated webpage loading thumbnail, instead of an hourglass > --ui.prefersReducedMotion Not sure how to feel about this one, curious how others feel. > `permissions.default.localhost->2` could at least get a wiki entry about [this](https://jdownloader.org/knowledge/wiki/glossary/cnl2); one would already need a BlockLAN (on by default in assets.json) uBO list @@ scoped exception to get it working, and *then* allow the permission prompt. Great point, thanks for bringing this up. Will do some testing. Honestly, instead of removing `permissions.default.localhost`, I think I'd rather remove/disable the Block LAN uBo list - I agree we shouldn't set both, and I think it's better to rely on native browser features for functionality like this over extensions where possible.
First-time contributor

@celenity Thank you for your consideration.

cloudflare turnstile

can be tested e.g. here ->https://www.techspot.com/. I recall testing it in late December - it manifested as a intermittently stuck redirect, after automatically or manually ticking the captcha box. However I cannot reliably reproduce it now - the prompt informing of the redirect flashing multiple (2-3) times isn't ideal (I guess it does its job), but doesn't prevent the turnstile to eventually suceed. However, I tested with JIT disabled back then, but with stock Firefox settings now, so maybe the js check just sometimes timed out, on slow(er) hardware. Still not great, but reload (CTRL+R) eventually coerced it through (turnstile again->website).

permissions.default.localhost

As far as I can tell, the ui in about:preferences#privacy doesn't allow manually entering exceptions, and with 2 it doesn't show any indicator in urlbar, to add an 'allow' exception for current webpage. The BlockLAN list are mostly exceptions; all of them would break. IMHO 0 (default prompt) and keeping the list is a better choice, at least for now.

btw, click'n'load-relevant entry

@celenity Thank you for your consideration. >cloudflare turnstile can be tested e.g. here ->`https://www.techspot.com/`. I recall testing it in late December - it manifested as a intermittently stuck redirect, _after_ automatically or manually ticking the captcha box. However I cannot reliably reproduce it now - the prompt informing of the redirect flashing multiple (2-3) times isn't ideal (I guess it does its job), but doesn't prevent the turnstile to eventually suceed. However, I tested with JIT disabled back then, but with stock Firefox settings now, so maybe the js check just sometimes timed out, on slow(er) hardware. Still not great, but reload (CTRL+R) eventually coerced it through (turnstile again->website). >permissions.default.localhost As far as I can tell, the ui in `about:preferences#privacy` doesn't allow manually entering exceptions, and with `2` it doesn't show any indicator in urlbar, to add an 'allow' exception for current webpage. The BlockLAN list are [_mostly exceptions_](https://github.com/uBlockOrigin/uAssets/blob/master/filters/lan-block.txt); all of them would break. IMHO `0` (default prompt) and keeping the list is a better choice, at least for now. btw, [click'n'load-relevant entry](https://github.com/uBlockOrigin/uAssets/issues/28218)
celenity approved these changes 2026-01-24 02:18:49 +01:00
celenity left a comment

LGTM! Thanks again for all your time and work here @any1here, very much appreciated ❤️.

LGTM! Thanks again for all your time and work here @any1here, very much appreciated ❤️.
Owner

@degausser

As far as I can tell, the ui in about:preferences#privacy doesn't allow manually entering exceptions, and with 2 it doesn't show any indicator in urlbar, to add an 'allow' exception for current webpage. The BlockLAN list are mostly exceptions; all of them would break. IMHO 0 (default prompt) and keeping the list is a better choice, at least for now.

Yeah, I did some testing, and I'm inclined to agree with you. Fixed for next release: 62fb7dd81e.

@degausser > As far as I can tell, the ui in `about:preferences#privacy` doesn't allow manually entering exceptions, and with `2` it doesn't show any indicator in urlbar, to add an 'allow' exception for current webpage. The BlockLAN list are *mostly exceptions*; all of them would break. IMHO `0` (default prompt) and keeping the list is a better choice, at least for now. Yeah, I did some testing, and I'm inclined to agree with you. Fixed for next release: https://codeberg.org/celenity/Phoenix/commit/62fb7dd81e0918c03e06104269151ada787fbaa0.
Sign in to join this conversation.
No description provided.