Works in Tor Browser and Firefox 67 (with a different dtd file as in the bugzilla PoC), probably also next ESR.
Did not do a PoC but it would be easy to get a specific string in all locales, and just compare with value obtained via a hidden iframe that loads an xml with the translated string.
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information
Child items ...
Show closed items
Linked items 0
Link issues together to show that they're related.
Learn more.
I was going to bring this up (GMTA).. seriously in the last few days this has been on my mind, I am not super savy in JS etc, and would appreciate a PoC for TZP ( if its possible via a github.io)
I'm making a couple of assumptions I would need to check. First, I'm assuming the only way to load locale DTDs is via chrome://*/locale/*. If there is a way to load them via re[/*](/*) or some chromeURIs that do not follow that structure, I think the patch would not work. Second, there is no way to load a resource from web content asabout:*`. If this was possible we could bypass the protection, because we are adding an exception for about pages.
Then, there is breakage again. Since Firefox relies on several re[/](/) and chrome://` URIs to load in web content for some features, these changes might break something. I think #8725 (moved) and corresponding https://bugzilla.mozilla.org/show_bug.cgi?id=863246 are relevant here. Arthur mentions in bugzilla several features that use these (video controls, image display, text-box resizing, and directory listing). I checked this and they work with this patch in ESR60 (except text-box resizing, not sure which one is that).
So I could not find this breaking in current ESR60. BUT, the same patch breaks several things in Firefox Nightly. First, the browser UI itself (for some reason it's loading chrome://... dtds from moz-nullprincipal origins). Video controls are also broken, displaying xml with errors (because the locale DTD could not be loaded).
I think I'll update the bugzilla ticket with a patch for central (even if it's broken), and ni ckerschb as Tom suggested.
Still, in central applying this will require fixing some things breaking due to some internal UI changes, but in ESR60 I did not see anything break so far. I think it would be ok to apply, but not sure if we want to wait for some progress on the bugzilla before.
I think we should not just copy the method but just have one in the tree and make all caller using that one (as you did in your patch for mozilla-central). Apart from that this looks fine to me (although I have not tested the fix yet)
Still, in central applying this will require fixing some things breaking due to some internal UI changes, but in ESR60 I did not see anything break so far. I think it would be ok to apply, but not sure if we want to wait for some progress on the bugzilla before.
So, this fix won't very likely make it into Tor Browser 8.5 and thus in no ESR 60 based stable release. Therefore, I am inclined to just fix it on ESR 68 and get it early on into our Tor Browser alpha builds based on that series (to iron out possible breakage). I am marking this ticket with ff68-esr to get it onto our radar.
Looks good to me. I guess we want to have a fix for https://bugzilla.mozilla.org/show_bug.cgi?id=1573276, too? Either way, I cherry-picked what we have at the moment and applied it to tor-browser-68.1.0esr-9.0-2 (commit 93c30885b92760fcdf6bd06b235ddd3c87b09b97, 2fe57c62af9706be45fc2085796a9398c3b10763, and 7baed647f56e5d3e7c92eaffadb1c18c07aabe7f).
Trac: Keywords: TorBrowserTeam201909R deleted, TorBrowserTeam201909 added Status: needs_review to needs_revision