| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
thus giving the user control over it.
This involved adding in a "creation" date to the metadata so we can respect the
user's choice of how far back to go when deleting browsing data. Care is taken
to handle older metadata without the creation date set.
Also fix a bug whereby we weren't making sure to persist the removed metadata
when it expires.
BUG=33445
TEST=TransportSecurityStateTest.DeleteSince, TransportSecurityStateTest.SerializeOld
Review URL: http://codereview.chromium.org/652035
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39684 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
| |
DoInitConnectionComplete and Read.
R=willchan
BUG=none
TEST=none
Review URL: http://codereview.chromium.org/652011
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39673 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
| |
cache histograms.
BUG=14193
TEST=none
Review URL: http://codereview.chromium.org/650068
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39654 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When SSPI is used (for Windows builds), the NTLM and Negotiate handler
factories determine the maximum token length the first time it is used.
The SSPI call to determinine the maximum length also returns an error code
if the scheme is unsupported. The factories remember if the scheme is
unsupported and will not attempt to create any handlers. If the token length
is found, it is remembered. If a different error occurs, don't create a
handler this round, but try again in the future.
BUG=None
TEST=Manually used an incorrect auth scheme and validated that it worked. Working on a mock SSPI Library I can use for unit testing.
Review URL: http://codereview.chromium.org/600129
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39600 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
| |
Also create necessary infrastructure to know the address
a client socket is connected to.
TEST=Covered by net_unittests.
BUG=35670
Review URL: http://codereview.chromium.org/598071
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39559 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
| |
previous checkin. It needs more work.
TBR=willchan
BUG=35108
TEST=none
Review URL: http://codereview.chromium.org/650109
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39541 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
| |
an SSLClientSocket by renaming the methods and adding a DCHECK.
R=willchan
BUG=none
TEST=No compilation errors.
Review URL: http://codereview.chromium.org/650085
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39528 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
| |
Allocate the -6xx block to FTP errors.
R=phajdan.jr
BUG=32945
TEST=None
Review URL: http://codereview.chromium.org/650006
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39510 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
errors for SPDY, we need to set |result| to OK.
Add rudimentary support for restarting a SPDY network transaction after
the user ignores a certificate error. This requires turning |use_spdy|
into a class member.
R=willchan
BUG=35108
TEST=Testing SPDY with NPN against a server with a broken certificate
should not result in an SSL certificate error page.
Review URL: http://codereview.chromium.org/651054
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39508 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
| |
page (so use of the --host-resolver-rules flag doesn't limit features), and run through the Shutdown() method on exit.
This reworks the "IsHostResolverImpl()" hack to be a bit more general.
BUG=36053
Review URL: http://codereview.chromium.org/646003
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39488 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
are some pretty complicated rules, and this helps isolate that code and better test it.
This also lays a framework for addressing bug 9835 (IP/CIDR matching)
Lastly, adds support for the exclusion format ".domain" on all platforms, which is interpreted as "*.domain".
BUG=28112
TEST=ProxyBypassRulesTest.*
Review URL: http://codereview.chromium.org/601070
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39486 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
| |
It's not often that you can make something both dynamic and static.
BUG=16831
TEST=none
Review URL: http://codereview.chromium.org/651055
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39485 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
| |
Review URL: http://codereview.chromium.org/650057
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39478 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
using SpdySession::InitializeWithSocket. The is_secure_ member needs
to be set to true.
R=willchan
BUG=none
TEST=Use SPDY with Next Protocol Negotiation (NPN). Chrome should show
a lock icon and should be able to display the server certificate.
Review URL: http://codereview.chromium.org/646074
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39474 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
| |
Getting the name right is important if this code is to compile with both the
10.5 SDK (where we define the type) and the 10.6 SDK (where the system defines
it). The error was introduced in r39389.
BUG=16831
TEST=10.6 SDK build
Review URL: http://codereview.chromium.org/651044
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39467 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
| |
The change broke certain linux buildbots due to netlink.h forward declaring struct net, which conflicts with the net namespace. The fix is to move the netlink code into a separate file which is not in the net namespace.
Review URL: http://codereview.chromium.org/650009
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39465 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A series of exchanges such as
C: GET / HTTP/1.1
S: 401, WWW-Authenticate: Basic realm=foo
C: GET / HTTP/1.1, Authenticate: Basic <token>
S: 401, WWWW-Authenticate: Basic realm=bar
Should be treated as a rejection for the credentials presented for the foo
realm.
BUG=None
TEST=net_unittests.exe --gtest_filter="*ChangeAuthTest*"
Review URL: http://codereview.chromium.org/647034
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39449 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
| |
file."
This somehow broke net_unittests on the Mac.
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39448 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If one tries to upload a file that one doesn't have read access to,
the browser livelocks. It tries to read from the file, gets nothing
but spins forever because it knows that it hasn't finished reading.
To address this, firstly we add a check at stat() time to make sure
that we can read the file. However, this doesn't take care of the case
where the access() call was incorrect, or the permissions have changed
under us. In this case, we replace the missing file with NULs.
http://codereview.chromium.org/541022
BUG=30850
TEST=Try to upload a file that isn't readable (i.e. /etc/shadow). The resulting upload should be a 0 byte file.
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39446 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
towards redoing the Callback interfaces.
Added and removed includes as needed.
BUG=35223
TEST=trybots
Review URL: http://codereview.chromium.org/646061
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39419 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
| |
It broke the chromium arm build.
BUG=26156
Review URL: http://codereview.chromium.org/651006
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39418 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
| |
BUG=26156
TEST=Browse a bit. Observe about:net-internals/hostresolver.hostcache to note that the cache is populated with data. sudo /sbin/ifconfig lo down; sudo /sbin/ifconfig lo up. Observe about:net-internals/hostresolver.hostcache to note that the cache should now be empty (or almost empty, since other host resolutions may have happened in the meanwhile).
Review URL: http://codereview.chromium.org/645001
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39417 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
| |
specified in RFC 2428.
R=eroman,phajdan.jr
BUG=32945
TEST=ftp://ftp.netbsd.org/ should work on a machine with dual IPv4/IPv6
connectivity.
Review URL: http://codereview.chromium.org/592001
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39416 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
than the desired max size, and adjust eviction to avoid
having long periods without it.
There's a few users with caches that are too big, and it
looks like the problem is that eviction could be stopped
forever, while still adding new items to the cache.
Now we only allow one minute (at a time) with eviction suspended.
BUG=none
TEST=none
Review URL: http://codereview.chromium.org/647054
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39407 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
| |
This includes sending an existing identity cert, and asking the user which cert to use. Doesn't yet handle SSL renegotiation, or key-gen.
BUG=16831
TEST=none
Review URL: http://codereview.chromium.org/604067
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39389 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
| |
This is a generalization of the --testing-fixed-server flag.
BUG=36053
TEST=MappedHostResolverTest.*
Review URL: http://codereview.chromium.org/647001
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39342 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For benchmarking we would like to pretend that we're in a world with
OCSP stapling and OCSP disk caches etc. Since we currently don't check
certificates with SPDY anyway, it's no loss if we don't check OCSP
either.
This change needs to be reverted when we start checking certificates.
Hopefully by then we'll have a better OCSP world to live in.
BUG=32020
TEST=none
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39314 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
| |
This will prevent us from attempting NPN when --use_spdy is true, but not the "npn" option is not specified.
Review URL: http://codereview.chromium.org/633006
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39312 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
| |
BUG=none
TEST=none
http://codereview.chromium.org/593089
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39123 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
| |
BUG=35869
TEST=gclient runhooks --force should work for arm configurations of linux.
Review URL: http://codereview.chromium.org/600144
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39110 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
| |
TEST=This is to reduce false positive "Unable to terminate process" log entries in unit tests.
BUG=none
Review URL: http://codereview.chromium.org/600108
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39097 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
GetFileExtensionFromPath function.
BUG=None
TEST=compiles
Patch from Thiago Farina <thiago.farina@gmail.com>.
Review URL: http://codereview.chromium.org/604049
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39070 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The driving rationale for this change was to prevent choosing an AuthHandler when it
is not supported on the system due to a missing runtime component (such as not being
able to locate a gssapi shared library when seeing a Negotiate scheme).
It also has the advantage (currently unused) of determining some per-auth-scheme properties
only the first time that a challenge for that scheme is seen (such as maximum token length for
the SSPI implementation of NTLM).
Finally, it may make unit tests easier to generate since the factory can be easily mocked.
BUG=34795
TEST=New unit test for HttpAuthHandlerDispatchFactory.
Review URL: http://codereview.chromium.org/582007
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39065 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
SSLClientSocketMac reports certificate error before SSL handshake is completed,
so just returning OK for LOAD_IGNORE_CERT_* won't work (completed_handshake_ is false yet, so we can't Read()/Write() on the socket).
Add the cert in allowed_bad_certs, and reconnect again.
BUG=35108
TEST=none
Review URL: http://codereview.chromium.org/593013
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39037 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
| |
to higher parts of the network stack.
BUG=23794
TEST=Covered by net_unittest, also see the bug.
Review URL: http://codereview.chromium.org/601027
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39017 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
explicit path set. This CL also ensures that we set the expiry info properly.
This means using CanonPath and CanonExpiration from cookie_monster.cc. Instead
of exposing those methods, I just expose a CanonicalCookie constructor that
takes a ParsedCookie.
R=pkasting
BUG=none
TEST=none
Review URL: http://codereview.chromium.org/597031
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38984 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Mozilla.
This corresponds to their patch ec019978e8dc, from 2010-02-11 13:39 +0000.
Also update instructions for updating this file.
BUG=27523
TEST=none
Review URL: http://codereview.chromium.org/600077
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38946 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In particular, if the database has multiple cookies defined for the same (host, cookie name, cookie path), we will delete all but the most recently created cookie.
This step happens right after loading the cookie database.
I don't expect this to have any real impact on startup performance, since the cookie database is small, and compared to the cost of reading the DB from disk, should be cheap.
This CL also includes two other related changes:
(1) Added a histogram "Net.NumDuplicateCookiesInDb" that measures how common the need to remove duplicates is.
(2) If the in-memory store is ever found to contain duplicates after initializion, fail hard (crash).
The effect of this change will be users that had a bad database will get it fixed, and will no longer be sending duplicate cookies to servers.
-----------------------
Background:
-----------------------
Ok, so why does the corruption happen in the first place?
From what I can tell, the problem stems from how the interface between CookieMonster and the PersistentCookieStore is defined. Notably:
* It implements overwrites as "delete" + "insert" operations.
* It doesn't define the behavior when a "delete" or "insert" operation fails.
The steps that CookieMonster() runs through to overwrite a cookie is:
(a) Find the existing cookie, |old_cookie|, in the in-memory store.
(b) Delete |old_cookie| from the in-memory store.
(c) Ask the persistent cookie backing store to delete |old_cookie| (keyed by its creation time)
(d) Insert |new_cookie| into the in-memory store.
(e) Ask the persistent backing store to insert |new_cookie|.
This ordering assumes that no failures can happen during steps (c) and (e).
However in actuality, SQLitePersistentCookieStore swallows any errors it encounters writing to the DB, and does not attempt retries.
Here is one sequence of steps that could lead to your database getting hosed:
(1) Your cookie database becomes temporarily unwritable (maybe your home directory is network mounted and your kerberose ticket just expired).
(2) The browser gets a set-cookie to overwrite an existing cookie (perhaps a ping-back from gmail, which happen often).
Clearly steps (c) and (e) will now fail, since the database is offline. So the in-memory store will get changed, but the on-disk one won't.
(3) Now your cookie database becomes writable again (maybe you renewed the ticket).
(4) Another cookie update is received. This time, the update will cause us to insert a duplicate into the cookie database.
This happens because in step (c) the on-disk database is asked to delete the previous (at least according to the in-memory store) cookie. Well, the on-disk DB never wrote that value, so this has no effect, and the actual previous value remains. Next we insert the new value, kaboom.
A next step will be to re-work the PersistentCookieStore interface so SQLitePersistentCookieStore isn't as fragile to errors while overwriting.
BUG=17855
TEST=CookieMonsterTest.DontImportDuplicateCookies
Review URL: http://codereview.chromium.org/602029
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38880 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
| |
BUG=27870
TEST=none
Review URL: http://codereview.chromium.org/604022
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38857 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
| |
BUG=35288
TEST=see bug
Review URL: http://codereview.chromium.org/596054
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38854 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
| |
cache grab it when needed.
BUG=26729
TEST=none.
Review URL: http://codereview.chromium.org/594041
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38848 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
| |
BUG=none
TEST=none
Review URL: http://codereview.chromium.org/598074
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38818 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
callbacks to notify the transaction about the completion
of AddTransactionToEntry.
BUG=26729
TEST=current tests.
Review URL: http://codereview.chromium.org/593058
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38808 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
| |
http://codereview.chromium.org/596013
BUG=none
TEST=none
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38773 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The chrome frame failure appears to be unrelated.
Add some tests to CookieMonster for overwriting persistent cookies, and checking that the PersistentCookieStore interface is exercised correctly.
Review URL: http://codereview.chromium.org/600040
TBR=eroman@chromium.org
Review URL: http://codereview.chromium.org/596048
TBR=eroman@chromium.org
Review URL: http://codereview.chromium.org/604012
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38728 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
| |
BUG=26729
TEST=unittest
Review URL: http://codereview.chromium.org/603011
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38713 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
even though this should be a strictly no-op change...
Add some tests to CookieMonster for overwriting persistent cookies, and checking that the PersistentCookieStore interface is exercised correctly.
Review URL: http://codereview.chromium.org/600040
TBR=eroman@chromium.org
Review URL: http://codereview.chromium.org/596048
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38707 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
| |
checking that the PersistentCookieStore interface is exercised correctly.
Review URL: http://codereview.chromium.org/600040
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38694 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
|
|
|
| |
This should prevent slowing down Mac startup.
BUG=34926
Review URL: http://codereview.chromium.org/593039
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38679 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
|
|
|
| |
Review URL: http://codereview.chromium.org/578026
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38659 0039d316-1c4b-4281-b951-d872f2087c98
|