summaryrefslogtreecommitdiffstats
path: root/net
Commit message (Collapse)AuthorAgeFilesLines
* Decrease the PAC polling delay even further.eroman@chromium.org2012-01-203-72/+97
| | | | | | | | | | Reduces the poll interval for success to 12 hours, and for failures to a maximum of 4 hours. BUG=110551 Review URL: https://chromiumcodereview.appspot.com/9255033 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118551 0039d316-1c4b-4281-b951-d872f2087c98
* Pick up the fix for NSS bug 718554 (a SSL_ForceHandshake bug).wtc@chromium.org2012-01-202-4/+4
| | | | | | | | | | | | | | | | SSL_ForceHandshake returns wrong positive results after sending pending data. See https://bugzilla.mozilla.org/show_bug.cgi?id=718554. Fortunately this bug doesn't affect Chromium because Chromium compares the return value of SSL_ForceHandshake with SECSuccess (0). R=agl@chromium.org BUG=none TEST=none Review URL: http://codereview.chromium.org/9265015 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118547 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 118489 since it seems to regress linux-sizes.sadrul@chromium.org2012-01-209-133/+121
| | | | | | | | | | | | | | | | """ Adds custom ttl argument to HostCache::Set. BUG=25472,107880 TEST=net_unittests Review URL: http://codereview.chromium.org/9197009 TBR=szym@chromium.org Review URL: https://chromiumcodereview.appspot.com/9138039 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118512 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 117772 - Add instrumentation to help track down a use-after-free.eroman@chromium.org2012-01-202-58/+4
| | | | | | | | | | | | | | This instrumentation checks to see if a deleted ClientSocketPoolBaseHelper::Group is being re-used. If it catches a use-after-free of this class then it will save the callstack where the deletion occurred into the mini-dump, for manual inspection by bug investigator. BUG=109876 Review URL: http://codereview.chromium.org/9207014 TBR=eroman@chromium.org Review URL: https://chromiumcodereview.appspot.com/9226030 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118511 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 118074 - Disable FPO throughout client_socket_pool_base.cc.eroman@chromium.org2012-01-201-13/+0
| | | | | | | | | | | | | | This will help track down the origin of a use-after-free bug. BUG=109876 Review URL: http://codereview.chromium.org/9240036 TBR=eroman@chromium.org Review URL: https://chromiumcodereview.appspot.com/9271013 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118510 0039d316-1c4b-4281-b951-d872f2087c98
* Fix a use-after free in socket pool which can happen after reaching the ↵eroman@chromium.org2012-01-201-0/+5
| | | | | | | | | | | maxium number of sockets. BUG=109876 Review URL: http://codereview.chromium.org/9226011 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118506 0039d316-1c4b-4281-b951-d872f2087c98
* Adds custom ttl argument to HostCache::Set.szym@chromium.org2012-01-209-121/+133
| | | | | | | | | | BUG=25472,107880 TEST=net_unittests Review URL: http://codereview.chromium.org/9197009 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118489 0039d316-1c4b-4281-b951-d872f2087c98
* net: remove SSL blacklisted hosts that were included because of record ↵agl@chromium.org2012-01-201-28/+0
| | | | | | splitting. git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118473 0039d316-1c4b-4281-b951-d872f2087c98
* Move operator<< for StringPiece into string_piece.{h,cc}akalin@chromium.org2012-01-201-1/+2
| | | | | | | | | BUG=110651 TEST= Review URL: https://chromiumcodereview.appspot.com/9159005 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118416 0039d316-1c4b-4281-b951-d872f2087c98
* [uber page] make references to chrome_shared2.css relativeestade@chromium.org2012-01-201-1/+2
| | | | | | | | | | | otherwise open_sans.woff doesn't get loaded (not sure why) BUG=none TEST=open sans is used Review URL: https://chromiumcodereview.appspot.com/9254001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118402 0039d316-1c4b-4281-b951-d872f2087c98
* The network_moved check in DoHandshakeLoop should only apply towtc@chromium.org2012-01-191-12/+13
| | | | | | | | | | | | | | | | | | | | | | | | | STATE_HANDSHAKE. After transport I/O completes synchronously, stay in the handshake loop only if the next state is STATE_HANDSHAKE, which is the only state that requires transport I/O to make progress. If the next state is any other state, we should ignore network_moved when determing if we should stay in the handshake loop. This changelist is the SSLServerSocketNSS version of http://codereview.chromium.org/9172005. SSLServerSocketNSS doesn't need this change now because it has no other handshake states, but this change will make it future-proof. Also remove unnecessary net:: prefixes. R=agl@chromium.org BUG=109706 TEST=existing unit tests for SSLServerSocket should continue to pass Review URL: http://codereview.chromium.org/9250031 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118356 0039d316-1c4b-4281-b951-d872f2087c98
* Disable pipelining if we encounter HTTP Authentication.simonjam@chromium.org2012-01-196-5/+56
| | | | | | | | | | BUG=102970 TEST=net_unittests Review URL: http://codereview.chromium.org/9241028 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118346 0039d316-1c4b-4281-b951-d872f2087c98
* Clean up uses of static data around singletons.rsesek@chromium.org2012-01-191-2/+7
| | | | | | | | | | | | This is in preparation for destroying singletons between each test. BUG=110594 TEST=Test suites. Review URL: http://codereview.chromium.org/9255020 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118320 0039d316-1c4b-4281-b951-d872f2087c98
* net: merge HTTP headers and body when the sum is small enough.agl@chromium.org2012-01-191-36/+46
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | (This is a reland of r118099 which was reverted in r118149 because it changed intermediate XMLHttpRequest events - breaking layout tests. We not don't free the request_body_ so that we continue to generate the same intermediate events.) When sending a small POST message we currently do two writes: one for the headers and one for the body. With 1/n-1 record splitting, this results in four TLS records when the connection uses CBC. Plesk correctly handles record splitting in HTTP headers, but incorrectly assumes that the entire POST body is in a single TLS record. Thus record splitting causes Chrome to be unable to log into Plesk (or perform any other action that requires a POST.) IE and Firefox 10 also do record splitting, but they merge the headers and body so there's only a single split after the first byte of the headers and Plesk can handle this. Our histogram data suggests that ~4% of requests can be merged in the fashion of this patch, so it seems independently worthwhile to do this. BUG=107966 TEST=none Review URL: http://codereview.chromium.org/9133007 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118298 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 118265 (added a static initializer on linux)thakis@chromium.org2012-01-196-148/+28
| | | | | | | | | | | | | | | | | | - Factor out chunk encoding logic into HttpStreamParser::EncodeChunk(). The logic is meaty enough to be factored out. Add unit tests along the way. BUG=72001 TEST=add unit tests. Review URL: http://codereview.chromium.org/9242018 TBR=satorux@chromium.org Review URL: https://chromiumcodereview.appspot.com/9264010 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118295 0039d316-1c4b-4281-b951-d872f2087c98
* Factor out chunk encoding logic into HttpStreamParser::EncodeChunk().satorux@chromium.org2012-01-196-28/+148
| | | | | | | | | | | | | The logic is meaty enough to be factored out. Add unit tests along the way. BUG=72001 TEST=add unit tests. Review URL: http://codereview.chromium.org/9242018 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118265 0039d316-1c4b-4281-b951-d872f2087c98
* Minor clean-up for HttpStreamParser::DoSendBody()satorux@chromium.org2012-01-191-4/+3
| | | | | | | | | | | | | This shouldn't change the behavior, as chunk_lengths_ is guaranteed to be zero at the two places, as wtc pointed out. BUG=72001 TEST=run tests. Review URL: http://codereview.chromium.org/9232032 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118227 0039d316-1c4b-4281-b951-d872f2087c98
* Fix a bug where an alternate protocol request to an unsafe portmmenke@chromium.org2012-01-193-26/+94
| | | | | | | | | | | | | | would hang indefinitely, and a logging bug from the same original CL, add unit test. Buggy commit: http://codereview.chromium.org/8898008/ R=wtc@chromium.org BUG=93326 Review URL: http://codereview.chromium.org/9240012 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118210 0039d316-1c4b-4281-b951-d872f2087c98
* Add comments to HttpStreamParser::DoSendBody().satorux@chromium.org2012-01-191-2/+10
| | | | | | | | | | | | | This function is relatively complex. Add comments to clarify what this function does. Will clean up the code in a separate patch. BUG=72001 TEST=chrome and tests build as before. Review URL: http://codereview.chromium.org/9213018 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118193 0039d316-1c4b-4281-b951-d872f2087c98
* Change the polling policy for PAC scripts to be much less frequent (new: 2 ↵eroman@chromium.org2012-01-182-48/+65
| | | | | | | | | | | | | | | | | | | | | | | | | | hours, old: 2 minutes). The poll intervals are now as follows: In response to a failure fetching the PAC script: 8 seconds 32 seconds 2 minutes 2 hours 2 hours ... (2 hours) In response to a success fetching the PAC script: 2 minutes 2 hours 2 hours ... (2 hours) BUG=110551 Review URL: http://codereview.chromium.org/9255023 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118180 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "net: merge HTTP headers and body when the sum is small enough."agl@chromium.org2012-01-181-45/+36
| | | | | | | This reverts commit 118099 becauase it may have broken some layout tests. git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118149 0039d316-1c4b-4281-b951-d872f2087c98
* SPDY - FlipServer - changes to use BufferedSpdyFramer.rtenneti@chromium.org2012-01-184-32/+70
| | | | | | | | | TEST=network unittests R=wtc Review URL: http://codereview.chromium.org/9174026 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118139 0039d316-1c4b-4281-b951-d872f2087c98
* Do not clobber dynamic public key pins when handling the ↵palmer@chromium.org2012-01-181-20/+35
| | | | | | | | | | | | | | | Strict-Transport-Security header. Rather than unilaterally calling EnableHost, do like ProcessPublicKeyPinsHeader and first look up any existing state for the given host. BUG=109691 TEST=Set a pin for a domain that serves HSTS, visit that domain, and then make sure the pin is still present Review URL: http://codereview.chromium.org/9231029 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118126 0039d316-1c4b-4281-b951-d872f2087c98
* net: merge HTTP headers and body when the sum is small enough.agl@chromium.org2012-01-181-36/+45
| | | | | | | | | | | | | | | | | | | | | | | | | | When sending a small POST message we currently do two writes: one for the headers and one for the body. With 1/n-1 record splitting, this results in four TLS records when the connection uses CBC. Plesk correctly handles record splitting in HTTP headers, but incorrectly assumes that the entire POST body is in a single TLS record. Thus record splitting causes Chrome to be unable to log into Plesk (or perform any other action that requires a POST.) IE and Firefox 10 also do record splitting, but they merge the headers and body so there's only a single split after the first byte of the headers and Plesk can handle this. Our histogram data suggests that ~4% of requests can be merged in the fashion of this patch, so it seems independently worthwhile to do this. BUG=107966 TEST=none Review URL: http://codereview.chromium.org/9133007 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118099 0039d316-1c4b-4281-b951-d872f2087c98
* Disable FPO throughout client_socket_pool_base.cc.eroman@chromium.org2012-01-181-0/+13
| | | | | | | | | | | This will help track down the origin of a use-after-free bug. BUG=109876 Review URL: http://codereview.chromium.org/9240036 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118074 0039d316-1c4b-4281-b951-d872f2087c98
* SPDY - bug fix for WINDOW_UPDATE loggingrtenneti@chromium.org2012-01-181-3/+3
| | | | | | | | | | BUG=108356 R=wtc TEST=network unit tests Review URL: http://codereview.chromium.org/9242021 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118036 0039d316-1c4b-4281-b951-d872f2087c98
* net/base: remove some wstringsevan@chromium.org2012-01-182-5/+5
| | | | | | Review URL: http://codereview.chromium.org/9221030 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118024 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 118001 - Revert 117968 - Rename is_hsts_host to fatal.apatrick@chromium.org2012-01-189-26/+29
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Per wtc: the old name was too specific, and became confusing because we now also have certificate errors that users cannot override, but for reasons other than (although still closely related to) HSTS. Failures were: http://chromegw.corp.google.com/i/chromium/builders/XP%20Tests%20%28dbg%29%286%29/builds/7935 http://chromegw.corp.google.com/i/chromium/builders/XP%20Tests%20%28dbg%29%282%29/builds/13488 (view as text) WorkerFileSystemTest.FileWriterSyncTruncateExtend: [1132:3724:0117/155729:2434562:INFO:ui_layout_test.cc(57)] Assume LayoutTests in E:\b\build\slave\XP_Tests__dbg__6_\build\src\chrome\test\data\layout_tests\LayoutTests [1132:3724:0117/155729:2434562:INFO:ui_layout_test.cc(143)] Assume LayoutTests in fast\filesystem [1132:3724:0117/155730:2434765:INFO:ui_layout_test.cc(143)] Assume LayoutTests in fast\js .\test\ui\ui_test.cc(720): error: Failed Timeout reached in WaitUntilCookieNonEmpty Google Test trace: .\test\ui\ui_layout_test.cc(174): file-writer-sync-truncate-extend.html .\test\ui\ui_layout_test.cc(254): error: Value of: inner_text.c_str() Actual: "[Worker] Test using FileWriterSync.truncate to extend a file. On success, you will see a series of "PASS" messages, followed by "TEST COMPLETE". Starting worker: ../resources/file-writer-sync-truncate-extend.js " Expected: expected_result_value.c_str() Which is: "[Worker] Test using FileWriterSync.truncate to extend a file. On success, you will see a series of "PASS" messages, followed by "TEST COMPLETE". Starting worker: ../resources/file-writer-sync-truncate-extend.js PASS [Worker] Truncate extension verified. PASS successfullyParsed is true TEST COMPLETE " Google Test trace: .\test\ui\ui_layout_test.cc(174): file-writer-sync-truncate-extend.html BUG=109941 Review URL: https://chromiumcodereview.appspot.com/9188019 TBR=palmer@chromium.org TBR=apatrick@chromium.org git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118011 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 117968 - Rename is_hsts_host to fatal.apatrick@chromium.org2012-01-189-29/+26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Per wtc: the old name was too specific, and became confusing because we now also have certificate errors that users cannot override, but for reasons other than (although still closely related to) HSTS. Failures were: http://chromegw.corp.google.com/i/chromium/builders/XP%20Tests%20%28dbg%29%286%29/builds/7935 http://chromegw.corp.google.com/i/chromium/builders/XP%20Tests%20%28dbg%29%282%29/builds/13488 (view as text) WorkerFileSystemTest.FileWriterSyncTruncateExtend: [1132:3724:0117/155729:2434562:INFO:ui_layout_test.cc(57)] Assume LayoutTests in E:\b\build\slave\XP_Tests__dbg__6_\build\src\chrome\test\data\layout_tests\LayoutTests [1132:3724:0117/155729:2434562:INFO:ui_layout_test.cc(143)] Assume LayoutTests in fast\filesystem [1132:3724:0117/155730:2434765:INFO:ui_layout_test.cc(143)] Assume LayoutTests in fast\js .\test\ui\ui_test.cc(720): error: Failed Timeout reached in WaitUntilCookieNonEmpty Google Test trace: .\test\ui\ui_layout_test.cc(174): file-writer-sync-truncate-extend.html .\test\ui\ui_layout_test.cc(254): error: Value of: inner_text.c_str() Actual: "[Worker] Test using FileWriterSync.truncate to extend a file. On success, you will see a series of "PASS" messages, followed by "TEST COMPLETE". Starting worker: ../resources/file-writer-sync-truncate-extend.js " Expected: expected_result_value.c_str() Which is: "[Worker] Test using FileWriterSync.truncate to extend a file. On success, you will see a series of "PASS" messages, followed by "TEST COMPLETE". Starting worker: ../resources/file-writer-sync-truncate-extend.js PASS [Worker] Truncate extension verified. PASS successfullyParsed is true TEST COMPLETE " Google Test trace: .\test\ui\ui_layout_test.cc(174): file-writer-sync-truncate-extend.html BUG=109941 Review URL: https://chromiumcodereview.appspot.com/9188019 TBR=palmer@chromium.org git-svn-id: svn://svn.chromium.org/chrome/trunk/src@118001 0039d316-1c4b-4281-b951-d872f2087c98
* net: randomly remove 10% of the SSL blacklist.agl@chromium.org2012-01-171-233/+0
| | | | git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117969 0039d316-1c4b-4281-b951-d872f2087c98
* Rename is_hsts_host to fatal.palmer@chromium.org2012-01-179-26/+29
| | | | | | | | | | | Per wtc: the old name was too specific, and became confusing because we now also have certificate errors that users cannot override, but for reasons other than (although still closely related to) HSTS. BUG=109941 Review URL: https://chromiumcodereview.appspot.com/9188019 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117968 0039d316-1c4b-4281-b951-d872f2087c98
* Fix compile error for ↵Guanqun.Lu@gmail.com2012-01-141-6/+15
| | | | | | | | | | | | | | | net/third_party/mozilla_security_manager/nsNSSCertTrust.cpp Because in the new nss3 package, CERTDB_VALID_PEER is deprecated and in the chromium build, deprecated is marked as errors. BUG=108656 TEST= compiled under nss-devel-3.12.10-6.fc15 and nss-devel-3.13.1-9.fc16 Review URL: http://codereview.chromium.org/9041002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117791 0039d316-1c4b-4281-b951-d872f2087c98
* Handle socket errors better while pipelining:simonjam@chromium.org2012-01-143-4/+57
| | | | | | | | | | | | | | | | - Treat empty responses the same as a closed socket (recover by retrying without pipelining). - If we get a socket error on the first request and haven't pipelined any other requests, don't count it as a pipelining failure. Notably, the first point fixes www.pogo.com, which closes the connection after sending back the root document. Oddly, it's actually able to pipeline everything else. BUG=None TEST=net_unittests and visit www.pogo.com with --enable-http-pipelining Review URL: http://codereview.chromium.org/9188066 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117776 0039d316-1c4b-4281-b951-d872f2087c98
* Add instrumentation to help track down a use-after-free.eroman@chromium.org2012-01-142-4/+58
| | | | | | | | | | | This instrumentation checks to see if a deleted ClientSocketPoolBaseHelper::Group is being re-used. If it catches a use-after-free of this class then it will save the callstack where the deletion occurred into the mini-dump, for manual inspection by bug investigator. BUG=109876 Review URL: http://codereview.chromium.org/9207014 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117772 0039d316-1c4b-4281-b951-d872f2087c98
* Don't poll the PAC script during periods of network inactivity.eroman@chromium.org2012-01-133-69/+373
| | | | | | | | | BUG=109310 Review URL: http://codereview.chromium.org/9139070 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117709 0039d316-1c4b-4281-b951-d872f2087c98
* Prevent callbacks after URLRequest/URLRequestJob have been ↵battre@chromium.org2012-01-138-29/+304
| | | | | | | | | | | | | | | | | | | | canceled/completed/killed. With this CL, URLRequestJobs notify their NetworkDelegate when they are being destroyed to insure that the NetworkDelegate does not access any of their variables or call callbacks. Also the WebRequest API does not execute any callbacks if they URLRequest has informed the API that the Request has been completed or canceled. Both cases could happen in rare race conditions. BUG=108968 TEST=no Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=117342 Review URL: http://codereview.chromium.org/9147018 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117685 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "Debug checks that URLRequest::set_method arguments match the contract."apavlov@chromium.org2012-01-131-4/+1
| | | | | | | | | | | | This change broke WebKit test http/tests/xmlhttprequest/methods-lower-case.html This reverts commit a964df83606a31b3674e20989e7dc6db9f3d4e86. TBR=cbentzel@chromium.org Review URL: http://codereview.chromium.org/9207003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117641 0039d316-1c4b-4281-b951-d872f2087c98
* chrome.experimental.downloads.download works for non-GET methods.cbentzel@chromium.org2012-01-131-4/+13
| | | | | | | | | | | | This wasn't working because ResourceDispatcherHost::BeginDownload was overriding the method. I traced through all callers of BeginDownload and confirmed that the method did not need to be set. BUG=110044 TEST=DownloadsApiTest.Downloads [the downloadPost test case] Review URL: http://codereview.chromium.org/9148070 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117637 0039d316-1c4b-4281-b951-d872f2087c98
* Debug checks that URLRequest::set_method arguments match the contract.cbentzel@chromium.org2012-01-131-1/+4
| | | | | | Review URL: http://codereview.chromium.org/9187060 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117635 0039d316-1c4b-4281-b951-d872f2087c98
* Fix incorrect constant used in status bitmaskjoth@chromium.org2012-01-131-2/+2
| | | | | | | | | | | | Turns out the test passes OK (given access to java side pieces, failure of which this bug masks) BUG=109783 TEST= Review URL: http://codereview.chromium.org/9187007 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117632 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 117342 - Prevent callbacks after URLRequest/URLRequestJob have been ↵battre@chromium.org2012-01-134-290/+17
| | | | | | | | | | | | | | | | | | | | | canceled/completed/killed. With this CL, URLRequestJobs notify their NetworkDelegate when they are being destroyed to insure that the NetworkDelegate does not access any of their variables or call callbacks. Also the WebRequest API does not execute any callbacks if they URLRequest has informed the API that the Request has been completed or canceled. Both cases could happen in rare race conditions. BUG=108968 TEST=no Review URL: http://codereview.chromium.org/9147018 TBR=battre@chromium.org Review URL: http://codereview.chromium.org/9173005 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117620 0039d316-1c4b-4281-b951-d872f2087c98
* Minor cleanup for UploadDataStream's constructor.satorux@chromium.org2012-01-132-2/+5
| | | | | | | | | | | | | This is in preparation for reworking UploadDataStream. We'll make more changes in the upcoming patches. BUG=72001 TEST=try bots Review URL: http://codereview.chromium.org/9187021 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117588 0039d316-1c4b-4281-b951-d872f2087c98
* Initialized variable to fix -Wmaybe-uninitialized warning.asharif@chromium.org2012-01-131-1/+1
| | | | | | | | | | BUG=none TEST=try bots. Review URL: http://codereview.chromium.org/9139041 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117584 0039d316-1c4b-4281-b951-d872f2087c98
* The network_moved check in DoHandshakeLoop should only apply towtc@chromium.org2012-01-131-7/+11
| | | | | | | | | | | | | | | | | | | | | | | | STATE_HANDSHAKE. After transport I/O completes synchronously, stay in the handshake loop only if the next state is STATE_HANDSHAKE, which is the only state that requires transport I/O to make progress. If the next state is any other state, we should ignore network_moved when determing if we should stay in the handshake loop. This changelist is the synchronous completion version of http://codereview.chromium.org/660131, which dealt with this bug when transport I/O completes asynchronously. R=agl@chromium.org BUG=109706 TEST=Run a Chrome debug build with --enable-origin-bound-certs and connect to a Google server such as gmail over HTTPS. Should not hit a DCHECK failure in net_log.cc. Review URL: http://codereview.chromium.org/9172005 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117578 0039d316-1c4b-4281-b951-d872f2087c98
* Handle Z_BUF_ERROR as a non-fatal error while incrementally decompressingrtenneti@google.com2012-01-132-4/+2
| | | | | | | | | | | | | | SYN_STREAM/SYN_REPLY headers. This also fixes bug when server sends SYN_REPLY in separate chunks (i.e. header + zlib chunks ) chrome fails to recognize it. BUG=108596 R=jar TEST=ASAN tests and network unit tests Review URL: http://codereview.chromium.org/9185036 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117563 0039d316-1c4b-4281-b951-d872f2087c98
* Cleanup after debugging with the XP builder.agl@chromium.org2012-01-122-23/+3
| | | | git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117437 0039d316-1c4b-4281-b951-d872f2087c98
* Remove unused code from UploadData and URLRequest.satorux@chromium.org2012-01-125-37/+11
| | | | | | | | | | | | | URLRequest::AppendFileToUpload() is no longer used as now file uploads are done with UploadData. TEST=chrome and tests build as before. BUG=72001 Review URL: http://codereview.chromium.org/9148016 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117357 0039d316-1c4b-4281-b951-d872f2087c98
* Prevent callbacks after URLRequest/URLRequestJob have been ↵battre@chromium.org2012-01-124-17/+290
| | | | | | | | | | | | | | | | | | canceled/completed/killed. With this CL, URLRequestJobs notify their NetworkDelegate when they are being destroyed to insure that the NetworkDelegate does not access any of their variables or call callbacks. Also the WebRequest API does not execute any callbacks if they URLRequest has informed the API that the Request has been completed or canceled. Both cases could happen in rare race conditions. BUG=108968 TEST=no Review URL: http://codereview.chromium.org/9147018 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117342 0039d316-1c4b-4281-b951-d872f2087c98
* Continuing to debug via the builders since there's no XP trybot.agl@chromium.org2012-01-111-1/+5
| | | | git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117310 0039d316-1c4b-4281-b951-d872f2087c98
* Continuing to debug with the XP builders.agl@chromium.org2012-01-111-0/+10
| | | | | | (There aren't any XP/Vista trybots any more.) git-svn-id: svn://svn.chromium.org/chrome/trunk/src@117298 0039d316-1c4b-4281-b951-d872f2087c98
generated by cgit v1.1 at 2025-03-01 20:44:14 +0000