- 14 Dec, 2015 1 commit
-
-
Dave Day authored
Change-Id: I0c787256e5bae7796fb4b62253a43363b32313f2 Reviewed-on: https://go-review.googlesource.com/17790Reviewed-by: David Symonds <dsymonds@golang.org>
-
- 11 Dec, 2015 5 commits
-
-
Yasuhiro Matsumoto authored
Fixes golang/go#13126 - GET/HEAD/POST method should not allowed for directory - Ignore EOFs while detecting the file. Change-Id: I5ce345408a5ea4ec0dc993631122e6c79fc64398 Reviewed-on: https://go-review.googlesource.com/16799Reviewed-by: Robert Stepanek <robert.stepanek@gmail.com> Reviewed-by: Nigel Tao <nigeltao@golang.org>
-
Dave Day authored
Change-Id: I8548d0e7a9400fd7a02c761b0b339749b69fbe31 Reviewed-on: https://go-review.googlesource.com/17665Reviewed-by: David Symonds <dsymonds@golang.org>
-
Dave Day authored
This changes the behaviour of the XSRF tokens slightly: - timestamps are rounded up to the nearest ms (not ns) - the timestamp is appended to the taken after base64 encoding Change-Id: Iaa32bd055bd76db77ef47f07b975e75c5ecaadf0 Reviewed-on: https://go-review.googlesource.com/17664Reviewed-by: David Symonds <dsymonds@golang.org>
-
Ian Lance Taylor authored
The license references puzzle programs that grep for licenses. Fixes golang/go#13573. Change-Id: I601fbc6ba2b189b476af1082c48fb02cd72f59d8 Reviewed-on: https://go-review.googlesource.com/17714Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Dave Day authored
Change-Id: Idd7bad48e585289740327b00f692fde948a2eeb0 Reviewed-on: https://go-review.googlesource.com/17662Reviewed-by: David Symonds <dsymonds@golang.org> Run-TryBot: David Symonds <dsymonds@golang.org>
-
- 09 Dec, 2015 1 commit
-
-
Brad Fitzpatrick authored
* don't send automatic Content-Length from Server on HEAD requests if response size is 0 (it's possible the handler didn't write because they looked at the method) * don't send Content-Type if handler explicitly set it to nothing. Matches the behavior of HTTP/1. Updates golang/go#13532 Updates golang/go#13495 Change-Id: If6e0898095cf88cb14efb6bbe82c88dbc2077e6b Reviewed-on: https://go-review.googlesource.com/17590Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
- 08 Dec, 2015 2 commits
-
-
Brad Fitzpatrick authored
Unlike HTTP/1, we now permit streaming the write of a request body as we read the response body, since HTTP/2's framing makes it possible. Our behavior however is based on a heuristic: we always begin writing the request body right away (like previously, and like HTTP/1), but if we're still writing the request body and the server replies with a status code over 299 (not 1xx and not 2xx), then we stop writing the request body, assuming the server doesn't care about it. There is currently no switch (and hopefully won't be) to force enable this behavior. In the case where the server replied with a 1xx/2xx and we're still writing the request body but the server doesn't want it, the server can do a RST_STREAM, which we respect as before and stop sending. Also in this CL: * adds an h2demo handler at https://http2.golang.org/ECHO to demo it * fixes a potential flow control integer truncation bug * start of clientTester type used for the tests in this CL, similar to the serverTester. It's still a bit cumbersome to write client tests, though. * fix potential deadlock where awaitFlowControl could block while waiting a stream reset arrived. fix it by moving all checks into the sync.Cond loop, rather than having a sync.Cond check followed by a select. simplifies code, too. * fix two data races in test-only code. Updates golang/go#13444 Change-Id: Idfda6833a212a89fcd65293cdeb4169d1723724f Reviewed-on: https://go-review.googlesource.com/17310Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
Nigel Tao authored
$ cd golang.org/x/net/publicsuffix/ $ go run gen.go -version "publicsuffix.org's public_suffix_list.dat, git revision 1f3ad51 (2015-12-05)" -test >table_test.go $ go run gen.go -version "publicsuffix.org's public_suffix_list.dat, git revision 1f3ad51 (2015-12-05)" >table.go Change-Id: I6f4bdd66b325b7abb097dcfe525a3cd1154b2d23 Reviewed-on: https://go-review.googlesource.com/17477Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
- 03 Dec, 2015 2 commits
-
-
Mikio Hara authored
This change replaces the existing log format separated by commas and spaces with space-separated one. Also removes unnecessary log output in test. Change-Id: I7fc5da1b58734a7ee1d06da4e15408bef0162b0b Reviewed-on: https://go-review.googlesource.com/16323Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
Mikio Hara authored
This change replaces the existing log format separated by commas and spaces with space-separated one. Also removes unnecessary log output in test. Change-Id: I1bf2b8c463db9174b2ac562533cdd89b72ed39de Reviewed-on: https://go-review.googlesource.com/16322Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
- 30 Nov, 2015 1 commit
-
-
Brad Fitzpatrick authored
Fixes golang/go#13397 Updates golang/go#6891 Change-Id: I1e4c7bfe60c6abf9a03f2888aa6abc3891c309e7 Reviewed-on: https://go-review.googlesource.com/17134Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
- 28 Nov, 2015 1 commit
-
-
Mikio Hara authored
Also adds test cases for the Len method of MessageBody interface. Fixes golang/go#13141. Change-Id: I8ab9e38727104ca094dfdb3020e8d42e611911e0 Reviewed-on: https://go-review.googlesource.com/16616Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
- 27 Nov, 2015 1 commit
-
-
Yasuhiro Matsumoto authored
Fixes golang/go#13286 Change-Id: If1e727bc18c64232b82484d9e82063cc59bcc826 Reviewed-on: https://go-review.googlesource.com/16859Reviewed-by: Robert Stepanek <robert.stepanek@gmail.com> Reviewed-by: Nigel Tao <nigeltao@golang.org>
-
- 26 Nov, 2015 1 commit
-
-
Brad Fitzpatrick authored
Tests are in net/http (clientserver_test.go, TestH12_AutoGzip) from https://golang.org/cl/17241 Fixes golang/go#13298 Change-Id: I3f0b237ffdf6d547d57f29383e1a78c4f272fc44 Reviewed-on: https://go-review.googlesource.com/17242Reviewed-by: Andrew Gerrand <adg@golang.org>
-
- 25 Nov, 2015 1 commit
-
-
Brad Fitzpatrick authored
Fixes found in the process of adding more A/B tests to net/http, comparing HTTP/1 and HTTP/2 behaviors. Most of the new tests are in Gerrit change Id9c45fad44cdf70ac9 in the "go" repo. Fixes golang/go#13315 Fixes golang/go#13316 Fixes golang/go#13317 Fixes other stuff found in the process too Updates golang/go#6891 (http2 support in general) Change-Id: I83b5bfb471047312c0dcb0a0b21d709008f34136 Reviewed-on: https://go-review.googlesource.com/17204Reviewed-by: Andrew Gerrand <adg@golang.org>
-
- 21 Nov, 2015 2 commits
-
-
Mikio Hara authored
Change-Id: Iac5121529bb1044e64724937f5d5e2bff8a336a6 Reviewed-on: https://go-review.googlesource.com/16321Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
Mikio Hara authored
Change-Id: I86f7d9e18f9d51f0a92ce32690a9c4aab1c79ab4 Reviewed-on: https://go-review.googlesource.com/16320Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
- 20 Nov, 2015 1 commit
-
-
Brad Fitzpatrick authored
Change-Id: I138f9941726767eb48f8459fb6af417a2c22c569 Reviewed-on: https://go-review.googlesource.com/16955Reviewed-by: Andrew Gerrand <adg@golang.org>
-
- 14 Nov, 2015 1 commit
-
-
Aiden Scandella authored
This fixes an incorrect contraction, and makes it easier to read. Change-Id: I8b88b6f54df25c4c36e00e6fbe77b3c32e518e12 Reviewed-on: https://go-review.googlesource.com/16895Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
- 11 Nov, 2015 2 commits
-
-
Blake Mizerany authored
Add a settings flag for users to send an initial settings frame immediatly after connection to the server. By default an empty settings frame is sent unless otherwise specified by the user. Change-Id: Ic3b5ce638b6c8371557b6489fa7539050fcb630f Reviewed-on: https://go-review.googlesource.com/16530Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Brad Fitzpatrick authored
Change-Id: Ieb02b751f3a0c1100e5201e7617ad9f055baf28e Reviewed-on: https://go-review.googlesource.com/16820Reviewed-by: Andrew Gerrand <adg@golang.org>
-
- 08 Nov, 2015 1 commit
-
-
Brad Fitzpatrick authored
To make an net/http.Transport (http1) be HTTP2-ified. Move this code out of net/http internals, so others can use it. And add tests. Some reflection and +build tag work is required so it is a no-op and returns an error for Go 1.5 and below. Change-Id: I539f233509602009e6c1179b810ff509a1f83ae3 Reviewed-on: https://go-review.googlesource.com/16734Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
- 07 Nov, 2015 3 commits
-
-
Brad Fitzpatrick authored
Change-Id: I50f7084bca2e5f7adb4cf278a32251f5bb239dbe Reviewed-on: https://go-review.googlesource.com/16731Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
Brad Fitzpatrick authored
Change-Id: Icbf40b26a25c7084efd062a0a66385450ec537aa Reviewed-on: https://go-review.googlesource.com/16699Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
Brad Fitzpatrick authored
For integration with the net/http.Transport. Updates golang/go#6891 Change-Id: I7a44e4c4259aa57cec1b54666c333b5fc519caf8 Reviewed-on: https://go-review.googlesource.com/16692Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
- 02 Nov, 2015 1 commit
-
-
Volker Dobler authored
Change-Id: I4e4eb9c710b0ee21a2ade97fd7d38cbd984a6b7f Reviewed-on: https://go-review.googlesource.com/16424Reviewed-by: Nigel Tao <nigeltao@golang.org>
-
- 30 Oct, 2015 1 commit
-
-
Brad Fitzpatrick authored
Kill off that field and simplify the code. Also update some docs. Thanks to Totoro W for the diagnosis in https://golang.org/cl/16465 Sadly this race was never detected due to golang/go#13097 Fixes golang/go#13089 Change-Id: Ia5400ea6733c7824e872affc3c6720bd436911fe Reviewed-on: https://go-review.googlesource.com/16490Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
- 29 Oct, 2015 2 commits
-
-
Brad Fitzpatrick authored
Change-Id: I700832c477a38ab11da39a382186bdc7d3d3186e Reviewed-on: https://go-review.googlesource.com/16445Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
Brad Fitzpatrick authored
Adapation of Blake's proposed fix in https://golang.org/cl/16463 with a few changes: -- bug fix (advance the buffer after writing) -- don't reacquire/release the buffer in the loop. it was done like that in case the max frame size changed while writing. Instead, push that down into awaitFlowControl since it has to acquire that lock anyway. Now it returns between 1 and the lower of how much we read and how much we're allowed to write. This does mean that if we start a request with a max frame size of 32KB, we'll never write larger than 32KB frames until the the next request (because our scratch buffer we read into is only 32KB), but we will start writing smaller DATA frames immediately once we see the peer's SETTINGS frame. Change-Id: I47fc503062f9602fe448cf7a36fc500e5d6b8ef9 Reviewed-on: https://go-review.googlesource.com/16443Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
- 28 Oct, 2015 2 commits
-
-
Blake Mizerany authored
This commit fixes two bugs. The first bug returned io.EOF when a zero bytes were read from the request body. The second bug was a hang where the Transport waited for more flow tokens than initialWindowSize BEFORE sending the first data frame which never gave the server a chance to send flow tokens, so the client never got enough to unblock awaitFlowControl. This commit changes awaitFlowControl to wait for for [1,max] tokens, where max is the length of the scratch buffer. Change-Id: Ibbac0a38cd672535917a38330998d3b48d46f5f1 Reviewed-on: https://go-review.googlesource.com/16411Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Blake Mizerany authored
Change-Id: Id10c0a9a5652bc38d6ae324f160b41af95bbee7f Reviewed-on: https://go-review.googlesource.com/16412Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
- 27 Oct, 2015 7 commits
-
-
Blake Mizerany authored
This commit allows a client of Transport to supply their own Dial function that assumes all TLS checks have been performed and the returned net.Conn is an h2 ready client connection. Change-Id: If35b5c47c3bd6912a990d6cd89feefa3303bb42b Reviewed-on: https://go-review.googlesource.com/16289Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Brad Fitzpatrick authored
Change-Id: Iaed80981b985e6e02ede5afc4a224490c4ceac01 Reviewed-on: https://go-review.googlesource.com/16380Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
Brad Fitzpatrick authored
Noticed during a crash in earlier testing. This now matches all the other channel sends, selecting on the serverConn's terminiation channel. Change-Id: I27ad3f9fd2d61154a5b95da82735d34fc2cbbc0b Reviewed-on: https://go-review.googlesource.com/16381Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
Brad Fitzpatrick authored
Change-Id: I3cbdf6107dfe13a0fb57d677059ca2106b1afb25 Reviewed-on: https://go-review.googlesource.com/16334Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
Brad Fitzpatrick authored
Change-Id: I795d230b78b43aea4c2088b1d04c927b2418a7a3 Reviewed-on: https://go-review.googlesource.com/16333Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
Brad Fitzpatrick authored
Some polish and testing remains, but this should be the bulk of the Transport work that was remaining. There's one notable (but easy) TODO about sending WINDOW_UPDATE frames every few hundred MB or so, but this is a checkpoint. Updates golang/go#6891 Change-Id: Iced9850804bf2c069c75118895ee7c3750ba31b5 Reviewed-on: https://go-review.googlesource.com/16310Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
Brad Fitzpatrick authored
Make the pipe code take an interface as the backing store. Now a pipe is something that's goroutine-safe and does the Cond waits but its underlying data is now an interface: anything that's a ReaderWriter with a Len method (such as a *bytes.Buffer), or a fixedBuffer (renamed in this CL from 'buffer'). This opens the ground to having a non-fixed buffer used with pipe. This also moves the CloseWithError code up into the pipe code, out of fixedBuffer. Change-Id: Ia3b853e8aa8920807b705ff4e41bed934a8c67b7 Reviewed-on: https://go-review.googlesource.com/16312Reviewed-by: Blake Mizerany <blake.mizerany@gmail.com>
-
- 22 Oct, 2015 1 commit
-
-
Brad Fitzpatrick authored
We decided to glue together the HTTP/1 and HTTP/2 Transports in the other direction, having the HTTP/1 code (net/http.Transport) start the flow. Remove the http2 Transport.Fallback for now, rather than leaving it half implemented. For background, see https://golang.org/cl/16090 Updates golang/go#6891 Change-Id: I511bc6d35a1a9a8e20010bd95ff694a894f42aa4 Reviewed-on: https://go-review.googlesource.com/16181Reviewed-by: Andrew Gerrand <adg@golang.org>
-