Not sure how strict your use case is. A local proxy might fit, but not
if you're trying to prevent even socket creation; for the latter
something like SELinux or systrace might work, but harder to set up and
manage.
--
You received this bug notification because you are a member of Ubuntu
Bugs, w
(In reply to Gervase Markham [:gerv] from comment #54)
> If I switch Firefox to offline mode in current nightly and type in
> "http://localhost/";, it tells me Firefox is offline and can't browse the
> web. Surely I should expect it to work?
The localhost name is just a name. It follows the norma
Green try at:
https://tbpl.mozilla.org/?tree=Try&rev=9147e7e13b02
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/367531
Title:
firefox should be smarter about offline access
To manage notifications
Created attachment 662281
Prepare for checkin
Thanks, Patrick!
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/367531
Title:
firefox should be smarter about offline access
To manage notifications ab
Created attachment 661458
Initialize STS when offline
Sorry for the review churn here. Previous patch wasn't passing try for
Android, which I had thought was due to an intermittent problem but
proved to be patch's fault.
Change here is in STS::Init, allowing initialization when offline.
Greenis
Created attachment 660621
Prepare for checkin
Pushed to try:
https://tbpl.mozilla.org/?tree=Try&rev=2b84bacf7276
> From the user's point of view, I wouldn't call it "the right
> thing". We're not showing her a page even though we have the
> data.
and
> anyway, I'm ok with this removal. just s
Created attachment 658634
Save and restore the network proxy preference value in tests for robustness
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/367531
Title:
firefox should be smarter about offl
Created attachment 658606
Address feedback
(In reply to Patrick McManus [:mcmanus] from comment #40)
> >
> > I'm also seeing warnings about potential deadlock again. Is this just a
> > conservative warning, or something to be addressed?
> >
>
> That's quite real.. I think the reset() logic sho
Created attachment 657902
Unbitrot
The last patch required slight fuzzing to apply (from the
|nsAutoCString| change).
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/367531
Title:
firefox should be s
Doing a full try push this time, to try to sniff out any remaining
problems before I ask for review.
https://tbpl.mozilla.org/?tree=Try&rev=40afcaef2976
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/3
Created attachment 656706
Correct order of nsSocketTransportService default constructor arguments
Not sure why this didn't hit before; maybe try didn't have -Werror or
had an exception for it before?
Try:
https://tbpl.mozilla.org/?tree=Try&rev=2766d56f0ff9
--
You received this bug notification
Created attachment 656680
Fix broken tests, move ensureSocketThreadTargetIfOnline to
ensureSocketThreadTarget
The |test_error_codes.js| test failure was getting
|NS_ERROR_NOT_INITIALIZED| from |nsHTTPConnectionMgr::PostEvent|;
|EnsureSocketThreadTargetIfOnline| becomes |EnsureSocketThreadTarget|
Created attachment 655837
Unbitrot and prepare for checkin
FWIW I tried the 'expires' test on http://www.procata.com/cachetest/
using the patch.
STR:
1. Go to http://www.procata.com/cachetest/
2. Click on the Expires "Perform Test" link
3. Click on the "Click here to go to step 2" link
4. Go Offl
Created attachment 644067
Address feedback
Thanks, Christian.
> > +++ b/netwerk/base/public/nsASocketHandler.h
> > +// for server socket, whether only accepts local connections.
>
> whether -> which, I think
It was supposed to be whether, but I cleaned up the comment a bit to be
clearer. Fo
(In reply to Chris Lee (:cleer) from comment #58)
> As discussed by the UX team in
> https://bugzilla.mozilla.org/show_bug.cgi?id=565552 , we'd like the find
> bar's visibility, search term, and probably toggle options to be specific to
> each tab rather than global.
The bug for the Find Bar is bu
My assumption was that LOAD_LINK/LOAD_FLAGS_IS_LINK would be strictly
for links and form submissions; the new flag gets used by some JS and
chrome loads.
The problem being solved is that each load requires either:
1. persisting the search if it's part of an existing browsing activity
2. resetting
Created attachment 630282
WIP v7: frontend: Unbitrot
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Firefox search box should be tab specific
To manage notifications about this bug go
Created attachment 626243
WIP v7: tests: Don't change other tests, don't change or test for change of
engine.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Firefox search box should b
Created attachment 626241
WIP v7: frontend: Don't make engine tab-specific, new tabs get reset search.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Firefox search box should be tab sp
Created attachment 626240
WIP v7: backend: Don't remove the global currentEngine.
>From reading all the dupes, the selected engine shouldn't be tab-
specific. If that's desirable, the previous work here can be
resurrected.
The other stand-out from the dupes was that most people want new tabs to
Created attachment 625811
Avoid deadlocks from events.
This only makes the final pass through the event loop when shutting
down.
Calling |Reset| in that locked block has a potential deadlock while
running the tests:
> ###!!! ASSERTION: Potential deadlock detected:
> Cyclical dependency starts at
Created attachment 625248
WIP v6: tests: Rework search tests now that the "engine-current" topic is gone
from search service notifications.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Created attachment 625247
WIP v6: backend: Removes current engine from nsSearchService.js and
nsIBrowserSearchService.idl
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Firefox search
Created attachment 625243
WIP v6: frontend: Rework the window.open() case so it won't recur.
The window.open() case could recur in the previous iteration, so this
prevents that by ensuring it only happens if the data hasn't been set
yet.
The backend patch removes the notion of a current engine fr
Created attachment 624246
WIP v5: tests: Update to work with other current changes.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Firefox search box should be tab specific
To manage n
Created attachment 624245
WIP v5: backend: Set browsed flag for window open calls from JS.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Firefox search box should be tab specific
To m
Created attachment 624244
WIP v5: frontend: Enable getting search data for new window via its opener.
The opener-based searchData code is specifically for sites that use
|window.open()| when the user does not have new windows redirected to
tabs. The updated backend patch is to support that by set
Created attachment 622898
backend: Treat meta refresh/similar as browsed to.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Firefox search box should be tab specific
To manage notifica
Created attachment 622897
tests: Add a meta refresh test
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Firefox search box should be tab specific
To manage notifications about this bug
Created attachment 620559
WIP v3: Flag loads as "browsed to" for better control/coverage of
persisting/resetting search
The goal for this version was to approximate the behavior described in
Bug 565740, Comment 4.
This patch has the various URI loaders that should persist the search
use |nsIWebN
Created attachment 620909
WIP v4: tests: Add tests.
The tests so far are:
1. New tab with persisting search
2. New tab with reset search
3. Changing tabs with different search values for each
4. Page load with reset search
--
You received this bug notification because you are a member of Ubuntu
Created attachment 620907
WIP v4: backend: split out as separate patch
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/306017
Title:
Firefox search box should be tab specific
To manage notifications
Created attachment 620906
WIP v4: frontend: Rework search.xml changes to not use indices.
In building some tests for this feature I realized that tabs keeping the
engine index wouldn't work if the engines are reordered, removed, added.
Now this uses the engine name, which is unique.
--
You recei
Created attachment 619220
netwerk: Add offline mode to DNS service and STS. Add RESOLVE_OFFLINE flag for
nsHostResolver.
When the STS is offline, the socket thread detects that and calls
|Reset()|, which detaches the sockets. If not shutting down, it only
detaches non-local sockets.
When the D
Created attachment 605120
WIP v2: Implements sessionstore, restore-to-default on locationbar edit
This iteration saves/restores the search data per tab across sessions.
That data is the engine, query, and whether the engine/query has been
modified by the user. That third piece of data is used for
Created attachment 603512
WIP v1: Initial tab-based searchbar.
The basics of it work. Known concerns:
Default engine: Should the first-in-list engine be restored on a new
browsing session/new window? (Currently the selected engine upon
startup will be the engine of the tab that was selected whe
Created attachment 575570
NSPR: Add a pair of tests (one IPv4, one IPv6 with IPv4 mapped address)
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/367531
Title:
firefox should be smarter about offline
Created attachment 57
NSPR: PR_IsNetAddrType should check for IPv4 loopback as the entire 127/8
instead of just 127.0.0.1
I split this out into a separate patch, as NSPR is a separate product.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed
Created attachment 575552
netwerk: single check for observer service before notifying
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/367531
Title:
firefox should be smarter about offline access
To m
Created attachment 573692
Implements offline-mode with respect for loopback.
This doesn't give network zones, but it does rework offline mode to let
loopback stay live.
It's not completely network-silent, as DNS requests are still attempted
(and succeed if the network is available). Suggestions
Both Nautilus and Epiphany behave this way for their Bookmarks menus.
They also respond to ctrl+click, which the awesome bar already supports.
I'm not aware of this behavior on other platforms or browsers, however.
The feature fits well with the semantics of using middle-mouse to create
new tabs,
Created attachment 524711
Proposed, exceedingly simple patch
All this does is adds a duplicate XBL:handler that specifies button="1".
This allows the history popup to be displayed while holding, and on
release over an item it opens in a new tab.
--
You received this bug notification because you
42 matches
Mail list logo
