perl github-issue-fetcher.pl \
--target_repo=golang/go \
--state=closed \
--assignee=bradfitz
上記の実行結果を貼っておきます。bradfitz さんのアクティビティ、すごいですね。
もっと当たり障りのない例があれば、差し替えます（インデックスされてここに到達したら無益なので）。
自分が開発しているリポジトリで使うと、振り返りができるかもしれません。
1 年間の活動実績をまとめるときには、多少使えそうです。
コードは https://github.com/yoheimuta/github-issue-fetcher.pl を参照ください。
Activities in Jan 2016
- net/http: go1.6beta1, nil pointer in http.Client
- net/http: net/http.(*http2pipe).Write: invalid memory address or nil pointer dereference
- net/http: TestServerValidatesHostHeader timeout
- net/http: Transport.CloseIdleConnections doesn't close idle http2 connections
- net/http: http2 negative content length
- net/http: http2 transport doesn't respect all http1 Transport fields
- x/net/http2: validate received header values
- build: darwin-386-10_10 builder not using cgo?
- net/http/httputil: DumpResponse could corrupt Response object
- x/net/context/ctxhttp: race condition preventing response from being closed
- net/http: check that http2 logs no more than http1
- doc: Windows default build from source behavior is CGO enabled but docs don't mention C compiler dependency
Total number of label
- The number of Builders is 1
- The number of Documentation is 1
Activities in Feb 2016
- net/http: http2 Transport retains Request.Body after request is complete, not GCed
- x/net/http2: Exporting h2 connection setup
- x/net/http2: should interpret Connection: close request header as Request.Close = true
- runtime: typo in comment about timer type's fields
- net/http: Server.ListenAndServeTLS not compatible with tls.Config.GetCertificate
- os/exec: understand 387 fd leak
- net/http: allow control of case of header key
- net/http: rejects requests where headers have whitespace
- mime/multipart: sort written headers
- net/http/httputil: NewSingleHostReverseProxy doesn't pass http authentication
- x/net/http2: Transport.TLSConfig.ServerName is ignored
- net/http/httputil: permit ReverseProxy to change Host header
- net/http/httptest: Server.CloseClientConnections does not force a connection error in go1.6rc2
Total number of label
- The number of Documentation is 1
Activities in Mar 2016
- net/url: Trailing question marks in query string appears to be misparsed
- os/user: implement basic cgo-free Current user function
- Failed building 1.6 from source on docker
- crypto/tls: timeout tests
- cmd/compile: oversized arrays
- net/http: ParseMultipartForm should populate PostForm with field values
- x/sys/unix: TimevalToNsec missing on linux arm 32bit
- net/http/httptest: make easier way to create Requests for testing
- net/http/httptest: add helper to create a server requests
- build: include C:\mingw\bin in Windows PATH by default
- net/http: remove support for continuation lines in header parsing
- net/http: don't require Host header for "PRI * HTTP/2.0" requests
- crypto/x509: export systemRootsPool or equivilant
- net/http/httptest: investigate removing the net.Conn.Close goroutine for Go 1.7
Total number of label
- The number of OS-Windows is 1
Activities in Apr 2016
- net/http: client's response header size not limited
- net/http/httputil: ReverseProxy should use StatusBadGateway if backend can't be connected
- net/http: documentation of "func (*Client) Do" doesn't state clearly when to close response.Body
- net: don't attempt DNS on .onion addresses
- os: missing ENOTEMPTY check in IsExist function
- x/net/http2: client should drop connection specific headers
- net/http/httptest: something about the ResponseRecorder change
- x/net/http2: make Server reject conn-specific request headers
- x/tools/imports: add FormatOnly option
- doc: be consistent in Deprecated sentence case
- net/http: don't read Request headers after ServeHTTP starts
- x/net/http2: validate connection headers more
- x/build/coordinator: re-use git clones in watcher process
- x/build/buildlet: trim go1.4 tarballs
- x/build: add more arm64 builders
- html: UnescapeString documentation error
- net/http: make it possible to access http.Server via http.ResponseWriter
- bytes: Reader should use new io.SeekStart, io.SeekCurrent, and io.SeekEnd constants
- net/http: empty port for ":port" causes it to default to 0
- net/http: document explicitly that http2 doesn't support Hijacker
- crypto/x509: Wrong error text when failing to unmarshal key-id
- net/http: Go 1.6 http.Client doesn't support http2 by default
- mime/multipart/partReader ignores first EOF it reads and fails if next Read call returns another error
- net/http: backport TimeoutHandler fix to Go 1.6.2
- cmd/api: recognize vendor packages as not contributing to public API
- context: typo in documentation
- net: clarify DialContext's lifetime
- net/http: close of closed channel
Total number of label
- The number of Documentation is 6
- The number of Builders is 3
Activities in May 2016
- net/http: add Transport.MaxIdleConns limit
- net/http: ErrWriteAfterFlush unused
- net/http: race using same long/infinite Request.Body after first is cut off by server
- x/build: tracking bug for Kubernetes buildlet pool support
- x/build: run sharded nacl builds
- net: clean up port parsing
- net/http: client crash with custom RoundTrip
- x/build/cmd/coordinator: hang if writing source tarball fails midway
- net/http: Response.Write has bogus Connection header after auto-ungzipped Content-Length response
- builders: ppc64le GNU/Linux builder DNS lookup is flaky
- net/http: make Transport.CancelRequest somehow work against http2
- net/http: custom net.Conn impls break http.Request.TLS
- net/http: http.MaxBytesReader has a bug
- x/sys: remove dragonfly/386 support
- net/http: StatusNoContent closes HTTP/1.0 keep-alive connections
- syscall: guard against Windows DLL preloading attacks
- src: move things from internal to vendor for Go 1.7
- net/http: with Transfer-Encoding: chunked, Transport doesn't reuse the tcp connection
- net/http: Transport analytics
- net/http: redirection response has closed body
- x/net/http2: use Context
- net/http: HTTP/2 cipher suites validation
- x/net/http2: validate transmitted header field names & values
- net/http: further restrict which trailers may be sent
- make.bash: GOBIN should be unset
- net/http/httptest: ResponseRecorder does not mimic real behavior of WriteHeader()
- net/http/httptest: ResponseRecorder doesn't match reality
- x/net/http2: delay sending request body in Transport if 100-continue is set
- net/http: TestTransportEventTrace_h2 fails with "expected substring "WroteRequest: {Err:}" in output."
- x/net/http2: transport.CloseIdleConnections() does not have any effect
- net/http: TestTransportPersistConnLeak test failure/flake on windows 2003
- net/http: TestTransportAndServerSharedBodyRace_h2 flake
- os/exec: delete RunContext, WaitContext
- net/http: TestTransportEventTrace_h2 fails with "expected substring "WroteRequest: {Err:}" in output." after the fix of #15744
- x/net/http2: Server can write invalid HTTP/2 if handler writes before reading 100-continue request body
Total number of label
- The number of Builders is 4
- The number of NeedsInvestigation is 2
- The number of Testing is 2
- The number of NeedsFix is 1
- The number of Vendoring is 1
- The number of OS-Dragonfly is 1
- The number of Documentation is 1
- The number of Security is 1
Activities in Jun 2016
- net/http: Server sends duplicate
Transfer-Encoding: chunkedwhen set explicitly - x/net/http2: receive buffer memory usage increases by 400x with HTTP2
- net/http: cannot register ServeMux handler for all CONNECT requests regardless of host or path
- net/http: http.ServeMux can't find registered handler for CONNECT requests
- net/http: Server removes all explicitly set
Transfer-Encodingheaders - document http client/transport thread safety
- http.TestClient and TestRedirect FAIL: Get http://www.google.com/robots.txt: unexpected EOF
- http client interface needs rewrite
- HTTP Conn needs a timeout
- Http server doesn't support writing multiple headers with the same name
- os.Readdirnames should return 0, os.EOF at eof
- http.Request.Write ignores Proto{Major,Minor}
- HTTP header injection bug
- mime.ParseMediaType can't handle non-ASCII filename
- crypto/tls, net/http: make PeerCertificates and other information availabe from within an http Handler
- Unable to set the Transfer-Encoding header using the SetHeader method on the http ResponseWriter
- net: dns crashes on bad responses
- html.Tokenizer.TagAttr shouldn't require quoted attr values
- mime/multipart: memory exhaustion due to no limit on line length
- CGI Location
- http.DefaultTransport.Do crashes if req.URL is nil
- HTTP Cookies Incorrectly Escape Paths
- /opt/go/src/pkg/http/triv.go can not compile
- http race?
- (*http.Response)Body.Close() does not work with gzip compression
- panic: runtime error: slice bounds out of range in bufio -> net/textproto
- Shouldn't the content type in http/fs.go be text/plain instead of text-plain ?
- Typo in http docs
- http.Request with multi Cookie error.
- HTTP client panics issuing HEAD request when response reports gzip content encoding
- add a mime/multipart.Writer
- regression: unable to listen on all interfaces
- Expires-field in set cookie not parsed properly (=at all) by http package
- Proxy-Connection: keep-alive
- http/spdy: incorrect frame type and flate: corrupt input
- About request Host field and copy Cookie in reverseproxy.go
- http.ListenAndServe denial of service
- http.ServeFile fails with compressed output streams.
- "gofmt -d" is broken. (and "gofix -diff" too)
- exec: Add WaitTimeout
- http ResponseWriter shouldn't allow writes on HEAD requests
- net/http: HTTP + TLS Handshaking documentation seems inaccurate
- Missing API http.Server.ListenAndServeTLS()
- Reading a multipart email returns only 1 Part object with all the parts inside as content
- net/http: connection never closes
- New http.Request.Cookie(name string) fails to find named cookie
- http.Redirect should not clean the query params of the redirect URL
- multipart.Writer.CreateFormFile should accept content type as a parameter
- http client doesn't buffer requests
- http: ReverseProxy adds "Transfer-Encoding: chunked" to GET requests
- assume HTTP/1.1 for the zero value of Request ProtoMajor/ProtoMinor?
- Unable to disable keep-alive in http server
- http.ReverseProxy should sometimes flush eagerly
- godoc should have a -markdown flag for github README.md
- json struct tag to omit zero value from objects
- http.ReverseProxy sends Transfer-Encoding: chunked in GET requests
- net/http: add ServeContent
- http.ReadRequest will not parse a responce
- http: unsolicited idle response
- blocked CGI children
- speed up http considerably via fix for io.Discard
- net/http: Denial of Service Protection in Go HTTP Servers
- http client does not support file: scheme
- http.DefaultTransport mutates the http.Request
- http: RoundTripper mutating *Request
- net/http: no_proxy does not handle numeric IP addr/range
- http: Server sends no error for malformed requests.
- http: http.Request ParseForm does not parse form data for PUT requests
- net/http: work around spurious 100 CONTINUE
- archive/zip: Header<->os.FileInfo functions
- document whether http.Response.Body is ever nil
- textproto reader multiline does not deal with rfc959 correctly
- [http] Content-Length header should be included for 0 length POST and PUT
- host header not suppressed for proxy connections
- crypto/tls: TLS client handshake is blocking incoming connections
- Reading from exec.Cmd.StdoutPipe() occasionally produces bad file descriptor error
- http: false User-Agent in DumpRequest
- Given an http.Response, get the resolved Location
- http server sends 400 after FIN
- database/sql: doesn't support output bool or float types
- exec package doesn't allow extra files beyond fds 0, 1 and 2
- json: confusing error message with unnecessary ,string modifier
- http: ReverseProxy should remove Connection: close headers
- http: URLs with RawPath domain.com?query
- net/http: buffer before chunking
- godoc http should have a package comment
- net/http: Transport VerifyHostname w/ TLS currently non optional
- mime API is inconsistent
- net/http/cgi spurious failure
- database/sql: does not allow "exporting" row data
- net.Write doesn't handle timeouts in partial writes
- fcgi.writePairs() writes key length twice rather than key length then value length
- panic can be lost
- openpgp: set clock during signing
- net/http: add CloseNotifier
- net/http: some test times out? or something fails? or some noise?
- database/sql: conn() can cause a deadlock
- os: should not have auto-generated signal lists
- net: OS X allows duplicate listen in single process?
- darwin build fails with export CGO_ENABLED=0
- net/http: Client / Transport deadlock
- database/sql: sqlite corruption
- database/sql: with gopgsqldriver Keep-Alive DoS: *db.Query Rows.Close() is not automatic
- net: DialTimeout does not clean up fd immediately
- http.Request.write doesn't handle bw.Flush()'s error
- go build ... includes the _obj directory?
- database/sql: should support time.Time
- database/sql: should copy into []byte by default; add RawBytes type for speed
- database/sql: add NullInt64, NullBool, etc
- net/http/httputil: TestDumpRequest failure
- Valid keys for json.Mashall
- net SetTimeout changes: time.Duration, total operation time
- os.Chmod should accept os.FileMode
- net/http/httputil: remove hop-by-hop headers in ReverseProxy?
- flag: cannot suppress output to os.Stderr
- net/http: handle CONNECT in ReadRequest/WriteRequest
- net: write good package comment
- net: SetDeadline docs aren't clear
- net/http: add Request.RequestURI
- database/sql: convert null values to []byte as nil
- cmd/goapi: embedded interfaces don't work
- openpgp: crash in signatureWriter.Write
- database/sql: convert NULL to empty string
- test/run: replace shell script with go program
- database/sql/driver: API
- net/http/httputil: NewSingleHostReverseProxy doesn't preserve query parameters
- cmd/go: 'go run' doesn't provide os.Stdin to the program it's running.
- database/sql: Sqlite database locks up when using database/sql across multiple go routines.
- cmd/api: cannot analyse exported constants referring to unexported ones.
- cmd/api: does not handle shift by a types constant.
- net/http: not using proxy for https get
- mime: API
- os: API
- os/exec: API
- cmd/go: go get mydomain.com
- net/http/httputil: DumpRequest() / DumpRequestOut() non-functional for SSL URLs
- database/sql: document stateful / pooling rules better
- archive/zip: Writer produces zip files unreadable by the OS X Finder's extractor
- transport.CloseIdleConnections() nils map, putIdleConn() doesn't check for nil and panics
- archive/tar: add FileInfoHeader function for symmetry with archive/zip
- misc/dist: use tar.FileInfoHeader once it exists
- misc/dist: archives should not exp/, old/, cmd/cov, or cmd/prof
- net/http: Documentation of implementation details for clients
- net/http: Default{Client,Transport} need better timeout support when used with connection pooling/redirects
- database/sql: Exec doesn't propagate error from subsetTypeArgs
- regexp: simple Ruby regular expression doesn’t work
- database/sql: document args
- os/exec: Failure in Cmd.Start() can orphan descriptors
- net/http: unit test failure
- net/http: Client doesn't correctly set cookies for all requests - patch attached
- os/user: tests fail if there is no home directory for current user
- net/http: handle headers with tokens better
- CONNECT method invalid in net/http
- mime: be stricter parsing media types
- net/http: write error from swiftly dropped connection
- net/http: Response.Write repeats status code
- database/sql: inconsistent parameter handling
- net/http: http Client regression
- encoding/json: Documentation incorrectly refers to InvalidTypeError
- cmd/api: add go1.1.txt file
- net/http: client connection close request not echoed in response headers
- net/http: Client can't close HTTP stream.
- database/sql: use Stmt.NumInput in DB.Exec
- net/http: accept * as URI in request parser
- io: document ReaderFrom
- net/http/client.go - Client.CheckRedirect does not have the claimed error behavior
- database/sql: document that it is valid to have multiple Rows for the same Stmt
- net/http: support multiple byte ranges
- net/http has broken compatibility on redirection handler
- net/http: parent directory listing vulnerability
- database/sql: deadlock with simultaneous queries
- database/sql: dangling statement(s)
- net/http: data race on Request.Method
- net/http: HTTP/1.0 keep-alive request to non-writing handler should set Content-Length 0
- net/http: connection leak on panic
- net/http: HEAD request doesn't populate Response.ContentLength
- net/http: Client doesn't support 302/303 redirects from POST->GET
- net/http: Request.Host field's definition is ambiguous
- net/http/fcgi: resets connection with GOMAXPROCS > 1
- build.golang.org: not in UTF-8
- net/http/httptest: NewRecorder does not initialize Code to 200
- net/http: Unrecoverable client panic crashes my program
- net/http: docs imply DefaultTransport doesn't use keep-alive connections, but it does
- api: handle the case where contexts re-converge
- net: io.Copy to timed-out net.Conn hangs?
- mime: we don't decode quoted-printable
- bytes: Reader.WriteTo returns EOF incorrectly
- database/sql: driver.ErrBadConn masked in *sql.DB.Begin
- net/http/httptest: Server.Close should also client connections
- net/http: client fails to handle requests with no body and chunked encoding
- encoding/pem: server fails with "failed to parse key PEM data"
- mime/multipart: Writer option to control boundary string
- net/http: handle cookie redirect TODO
- net/http: Transport leaks goroutines when request.ContentLength is explicitly short
- io/ioutil: Discard's ReadFrom method's intentional raciness is actually harmful
- time: race in Now
- net/http/cgi: TestDirWindows should be case insensitive
- net/http: server keep-alive timeouts cannot be implemented
- cmd/api: func params with names
- time: TestReset is unreliable
- net/http: data race in tests
- net/http: setting only WriteTimeout doesn't work
- net/http: client support for multipart/byteranges
- database/sql: Open() doesn't actually open a connection
- syscall: implement Getwd on OS X
- databaes/sql: don't panic on nil scan pointers
- net/http: fix, test, document requesting and serving %2F in paths
- database/sql: Stmt.Exec is wasting database resources
- net: flaky test TestConnAndListener
- net/http: test failures on linux/386
- database/sql: must call (*driver.Stmt).Close before correspondent (*driver.Conn).Close
- os/exec: exec pipes not closed when LookupPath fails
- bufio: no way to give/take underlying buffers
- database/sql: Result interface's methods should be documented
- bytes: bytes.Buffer uses an unbounded amount of memory even if its length stays bounded
- net/http: don't create io.LimitedReaders for 0 byte GET requests
- database/sql: assumes that driver.Stmt.Close does not need the connection
- mime/multipart: automatic quoted-printable decoding is not documented
- Clarify archive/zip docs abour path handling inside archives.
- database/sql: prepared statements leak connections with high concurrency
- net/http: server ResponseWriter should implement WriteString(string) error
- cmd/godoc: add link checker
- net/http: assume POST has Content-Type: application/x-www-form-urlencoded if missing (IE)
- image/png: wrong error message in png.Encode
- go/token: AddFile API is racy
- cmd/godoc: logs noise on HEAD requests
- net/http: treat HEAD methods less specially
- database/sql: extra RemoveDep
- net/http: malformed HTTP request when GET follows HEAD request
- net/http: ReadResponse needs a Request
- net/http: ServeFile doesn't work for streaming videos anymore
- website: tip.golang.org "playground not defined"
- net/http: close of Body closes keep-alive connection
- net/http: inconsistent results using os/exec in handler
- database/sql: prepared statements are bound to connections, autoreconnection doesn't work
- database/sql: mysql TLS support for Open
- net/http: Transport should err on EOF before declared Content-Length
- net/http: memory leak in http.Transport
- cmd/go: get fails for when document has encoding other than utf-8
- database/sql: Change error message for closed connection
- net/http: tls.Config.ServerName ignored on line 501 of net/http/transport.go
- godoc: makes too many system calls
- net/http: TestChunkReaderAllocs fails
- database/sql: statements become slower to invoke over time as new connections are opened
- bufio: revert central buffer pool? figure out a plan for garbage.
- cmd/api: run api check on builders
- compress/flate: add Reset method to let callers avoid garbage
- HTTP server shouldn't set Content-Length header for 304 Not Modified response
- net/http Request.Body.Close Documentation
- godoc indexing fails
- misc/pprof: don't rely on either just LWP::UserAgent or just curl
- http.ServeContent reads too much before calling DetectContentType
- os: add Unsetenv
- net/http/httputil: DumpRequestOut(req, false) fails if req.Body != nil
- net/http: ReadResponse should return ErrUnexpectedEOF on truncated response
- net/url: '?' in userinfo is not escaped properly
- net/url: regression from issue 11698045
- database/sql: redundant check
- crypto/rc4: rc4_ref.go does not compile
- path/filepath: Walk is not walking over all files
- strings: Replacer fails at beginning of string
- os: Readdir returns error if file disappears during call
- net/http: 204 and 100 response handling problems
- net/textproto: CanonicalMIMEHeaderKey panic on strings with spaces
- cmd/godoc: turn off log message about 'skipping example'
- net/http: Response.Write should write Content-Length: 0 header if present and not chunked
- net/http: add GetBasicAuth method to *http.Request
- go.tools/astutil: AddImport() works improperly when zero imports, comment present.
- net/http: Transport.CancelRequest doesn't work on connections blocked in dial
- net/http: test failure
- net/http: Request.Body should specify body is not closed on errors
- net/http: data race in BenchmarkServerHijack
- net/http/httputil: HTTP protocol errors using ReverseProxy
- syscall: linux F_SETLK constant is wrong
- archive/zip: error reading large files from zip64 files
- database/sql: Lock contention in func (rs *Rows) Scan
- net/http: segfault in bufio
- syscall: linux/386 socketcall uintptr
- net/http/cgi: CGI requests leak processes on output failure
- net/http/cgi: CGI returns 200 OK for invalid results (including patch)
- database/sql: last connection in pool not handed out correctly
- os/exec: Command does a LookPath in the wrong place.
- net/http: data race in TestTransportResponseHeaderTimeout
- net/http: server TLS handshake failures are silent
- go.tools/import: goimports does not correctly handle "." includes.
- os: System crash on Mac OS X after double kill
- net/http: TestClientWithIncorrectTLSServerName is broken
- crypto/cipher: incorrect return value documentation for AEAD.Open interface method
- net/http: Request Body error should not be ignored
- database/sql: TestMaxOpenConns failing
- net/http: Transport should always verify we wrote the request
- net/http: Response.Body should be closed while RoundTripper returned a response & error
- net/http: Requests to the server with Expect: 100-continue fail if Content-Length: 0
- http/httptest: waitgroup race detection seems to give false positives relating to httptest.Server.Close
- net/http: Setting custom "Host" request header doesn't have effect
- crypto/tls: TestConnReadNonzeroAndEOF fails or hangs
- build: add linux/387 builder
- code.google.com/p/go.tools/imports: too much concurrency
- encoding/base64: DecodeString accepts incorrectly padded input and may produce incorrect output.
- net/http: An empty reply causes false race conditions (go1.3beta1+)
- mime/multipart: add Example for NewReader
- os/exec: stop stdin copying goroutine after process exits
- mime: FormatMediaType inconsistent results
- net/http: Request.FormValue document return value if missing key
- net/http: double Content-Length in Response.Write on POST/PUT Request
- archive/zip: can't handle certain Java-generated zip (apk) files
- net/http: TimeoutHandler panics with concurrent requests
- net/http: ReadRequest doesn't populate req.Close
- net/http: document that conn.Hijack might have left-over read/write timeouts on net.Conn
- net/http/httputil: ReverseProxy doesn't forward Close notifications
- http.TimeoutHandler breaks with GOMAXPROCS n > 1
- net/http: document the behavior of %2F in url strings
- net/http: data race in test
- crypto/rand: use new getrandom system call on Linux
- net/http: Allow verification of certificates beyond just hostname
- mime: ParseMedia fails to parse googleapis batch boundary
- net/http: ResponseWriter leaves connection open after WriteTimeout
- encoding/json: inconsistent behavior in *(numeric type) and string tag option
- net/http: ListenAndServeTLS() can't be started with empty arguments
- net: wrong sample code in package document
- x/build/builders: subrepos shouldn't cause full rebuild of Go
- net/http: FileServer calls FileSystem.Open with unclean path for //index.html
- builder: add gcflags="-N -l" variant
- net/http: get for literal ipv6 address url doesn't follow redirection on ipv6-only transport environment
- net/http: Content-Range not honored by http.Client
- net/http: test fails with "Too many goroutines running" after benchmarks
- net/http: ResponseHeaderTimeout can cause the Heap & GC to grow/increase rapidly
- cmd/go: cache meta lookups during go get
- net/http: No easy way to get a server-side Request for testing.
- go.tools/imports or cmd/gofmt: Mode to reorganize imports to separate stdlib and 3rd party, but do not add/remove entries
- runtime: malloc2.go not gofmt'd
- tipgodoc breaks tool installation
- website: tip.golang.org fails with 502 Bad Gateway
- builders: windows version assumptions for builder should be documented
- builder: fatal: destination path 'go.tools.TMP' already exists and is not an empty directory.
- go/build: MultiplePackageError values swapped.
- net/smtp: STARTTLS with server certificate hash pinning
- builders: dashboard should return all work items for a builder
- builders: let coordinator give dashboard a build status URL
- net/http: CancelRequest break another irrelevant request
- net/http: TestResponseWriterWriteStringAllocs fails on windows-amd64
- database/sql: data race
- net/http: something about 100-continue
- net/smtp: package document does not explain how to add subject and body in msg parameter
- io/ioutil: ReadDir: no error if
dirnameis a file - net/http: ServeFile should recognise unix epoch as "zero time" ?
- net/http: data race in closing http.Response.Body when still reading it
- net/http: Server sets both Content-Length & Transfer-Encoding: chunked
- net/http: bug in late binding
- net/http: ServeContent for a directory unexpectedly closes the socket
- x/sys: apply errno amortization CL
- net/http: serveContent should 500 instead of 404 on error
- net/http: PostForm/Post/Get option to overwrite header for request
- net/http: CancelRequest twice on the same request panics
- net/http: flaky TestIdleConnChannelLeak
- net/http: some left out logging in net/http/cookie.go
- net/http: possible readloop goroutine leak
- mime/quotedprintable: return a *Reader, not *io.Reader
- net: DNS requests do not work on Android
- net/http: HTTP Server messed up pipelined HTTP requests after POST request
- archive/zip: cap out of range
- x/build: all-compile builder only do weird things
- os/exec: add boolean to tell whether exec is implemented on this platform
- archive/zip: checksum error
- archive/zip: file with wrong checksum is successfully decompressed
- x/build: talks repo fails due to size
- net/http: allows empty header names
- net/http: http.ServeFile and http.ServeContent Explicitly Set http.StatusOK
- net: LookupIP("127.0.0.1") does not work when CGO_ENABLED=1
- net/http: ServeMux uses un-escaped paths in the implicit /tree → /tree/ redirect
- net/http: Calling request.Body.Close() triggers DOS attack vector
- net/http: data race in test
- net/http: can't send star request
- net/http: various HTTP/0.9 bugs
- net/http: Serve violates tls.Config immutability promise
- crypto/tls: message too long for RSA public key size
- net/http: test timeouts on dashboard
- x/net/http2: EOF when Transport received GOAWAY (http2 client)
Total number of label
- The number of FrozenDueToAge is 369
- The number of Documentation is 17
- The number of Suggested is 11
- The number of RaceReport is 8
- The number of Testing is 4
- The number of GarbageCollector is 3
- The number of Builders is 2
- The number of NeedsFix is 2
- The number of Performance is 1
- The number of Security is 1
- The number of ExpertNeeded is 1
- The number of HelpWanted is 1
- The number of Unfortunate is 1
- The number of NeedsInvestigation is 1
Activities in Jul 2016
- internal/singleflight: flaky TestDoDupSuppress
- net/http: miss host header when serve as forward proxy.
- net/http: TestClientRedirectContext is flaky
- net/http: Server.ReadTimeout closes connection before response is finished
- net/http: document and protect against unsafe use of ServeFile
- net/http: MaxBytesReader's behavior does not match the comments
- net/http: flaky TestTransportHeadChunkedResponse
- net: add mechanism to dump DNS strategy
- cmd/go: return an error for unsupported GOOS/GOARCH pair
- net/http: something about ACME letsencrypt
- proposal: context: new package for standard library
- net/http: HTTP/1.1 requests without Host header should be rejected
- x/net/http2: flaky TestTransportReqBodyAfterResponse_403
- x/net/http2: Client POST request fails against golang http/2 server if client request.Body isn't completely consumed.
- net/url: scheme gets lost for opaque paths
- net/http: server-side expect 100-continue behavior seem broken on persistent connections
- x/build/cmd/coordinator: building subrepos at run repo head
- x/net/http2: flaky TestTransportReusesConn_RequestClose and TestTransportReusesConn_ConnClose
- x/tools/imports: upper-cased import paths not being selected as candidates
- x/tools/cmd/goimports: add configuration mechanism for ignore list
- x/build/cmd/gomote: put14 reports "unknown builder"
- api, cmd/api: exit status 1 during API check for Go1.5 beta 2
- x/build/watcher: crashing after finding commit in two branches
- x/build/cmd/coordinator: watcher's git mirroring is slow/flaky
- x/tools/cmd/goimports: too slow
- x/build: add a fake Windows XP-ish builder
- net/http: Post http://example.com: net/http: HTTP/1 transport connection broken: readLoopPeekFailLocked: EOF
- net/http: timeout error no longer returned from Transport.RoundTrip
- net/http: undocumented race condition on simultaneous calls to (*http.Server).Serve
Total number of label
- The number of FrozenDueToAge is 9
- The number of Builders is 4
- The number of Testing is 3
- The number of WaitingForInfo is 1
- The number of Proposal is 1
- The number of HelpWanted is 1
- The number of OS-Windows is 1
Activities in Aug 2016
- x/net/http2: return flow control when stream closes with buffered data
- x/net/http2: flow control over payload length, not data length
- net: race where Dialer.DialContext returns connections with altered write deadlines
- x/net/http2: Transport hangs on malformed response headers
- net: timeout support for Dial
- net/http: proxy code should know not to proxy 127.0.0.1
- encoding/base64: SetReadDeadline fails to set deadline on socket when piped thru base64 decoder
- os/exec: close read end of stdout/stderr pipe when copy loop stops
- all: refresh the contributors file for the release
- net/http: wrong error on canceled http request
- net/url: parseHost doesn't valid ports or contents after IPv6 literals
- internal/singleflight: test flake
- x/build: reverse roundtripper hang
- net/http: request body errors don't close a connection
- cmd/release: must run api checker
- x/net/http2: hung request to vision.googleapis.com (bad flow control?)
- net/http: imperva HTTP/2 attack vectors report
- dashboard: top-level view of subrepos doesn't match subrepo view
- net/http/httputil: request may be killed by mistake by former cancel request in ReverseProxy.ServeHTTP
- cmd/yacc: move to subrepo
- os: Getenv documentation update to point user to os.LookupEnv
- log/syslog: Dial function needs examples
- doc: go1.7 release notes include typo for TLSConfig.NextProtos
- x/net/http2: too much coupling between connection and stream window sizes
- net/http: must not copy tls.Config
- dashboard: builder with GOEXPERIMENTs?
- doc: deprecation message for Transport.CancelRequest is not correct
- io: endless loop in MultiReader in Go 1.7
- cmd/internal/obj: use keyed ProgInfo literals
- context: decide on naming convention for context key variables
- os/dir: File.Readdir claims it will return an error on empty slice, but doesn't
- cmd/dist: should C warnings cause make.bash to fail?
- net/http: TestTransportEventTrace_h2 flake
- net/http: set Dialer.DualStack on DefaultTransport?
Total number of label
- The number of FrozenDueToAge is 15
- The number of Documentation is 4
- The number of Builders is 3
- The number of NeedsInvestigation is 2
- The number of Suggested is 1
- The number of WaitingForInfo is 1
Activities in Sep 2016
- net/http: flaky TestH12_RequestContentLength_Known_NonZero
- net/http: Response.Write includes Content-Length header on 204s
- net/http/httputil: ReverseProxy incompatible with http.Transport retries
- net/http: no way for Server to set Trailer headers
- io: add WriteString support to io.MultiWriter
- net/http: consider reusing buffers replacing io.Copy calls with io.CopyBuffer
- io/ioutil: TempDir fails if "/tmp" doesn't exist and first argument is empty
- AUTHORS: missing Oracle
- net/url: getting the hostname of a URL
- net: flesh out LookupPort static services for nacl, etc
- net/http: unable to post body to an HTTP/2 server.
- net/http: TestTransportEventTraceRealDNS fails on OS X 10.11.6
- website: inconsistency in supported OS X versions
- net/http, net/http/httptest: locks being copied
- x/build: plan9 is holding up trybots again
- doc: misaligned text for covariant result types section of FAQ
- net: no way to set timeout for LookupAddr
- x/build/builders: move 'commit' polling from the builder code to the coordinator
- x/build: shard the shootout on builders
- x/build: coordinator updates for coreos changes
- x/build: reverse mac builder GOROOT issue
- x/build: coordinator tracks dead reverse buildlets for a long time
- x/build/coordinator: retrying test on same buildlet after timeout
- net/http: Transport should de-UTF-8 hostname in Host header
- x/net/http2: support CONNECT method
- net/http: 307 and 308 status codes should not change method on redirect
- x/net/http2: Transport returning "http2: no cached connection was available"
- net: TestResolveIPAddr and TestLookupPort_Minimal fails
- x/net/http2: no public method to send PING frames
- x/build: cross-compile ARM for speed?
- x/net/http2: finish Transport.IdleConnTimeout
- Proposal: Allow to change net/http.Server base context
Total number of label
- The number of FrozenDueToAge is 9
- The number of Builders is 8
- The number of Testing is 2
- The number of Documentation is 2
- The number of Proposal is 1
- The number of HelpWanted is 1
- The number of WaitingForInfo is 1
- The number of Proposal-Accepted is 1
- The number of FeatureRequest is 1
Activities in Oct 2016
- net/http: client
Expect: 100-continuereads first byte of input on failure - proposal: file system notifications library
- sort: make sorting easier, add Slice, SliceStable, SliceIsSorted, reflect.Swapper
- x/build/coordinator: negative waitgroup counter panic
- context: DeadlineExceeded doesn't implement net.Error interface
- net/http: improve performance for parsePostForm
- x/sys/unix appears to be broken
- net/http/httputil: NewSingleHostReverseProxy fails to transparently send requests
- x/build/revdial: hung revdial buildlets
- net/textproto: fails to trim header value
- x/net/http2: serverConn.readFrames goroutine leak
- x/net/http2: limits on header block size
- net/http: impossible to use both CloseNotify and Hijack in the same Server
- database/sql: add support for named input parameters
- crypto/rand: Panic fetching random boundary string
- database/sql: Exec() never returns when IP is not reachable and maxOpenConns is reached
- x/net/http2: limit on sum of header block fragments
- net/http: support multiple identical Content-Length headers
- database/sql: panic inserting a nil pointer when underlying type implements driver.Valuer
- database/sql: set maximum lifetime of connection
- net/http: make http1 Server always in a Read, like http2
- net/http: cancel context when underlying connection closes
- database/sql: expose database original column type
- Proposal: x/net/socket, vendor/golang.org/x/net/socket: new package to register socket types with package net
- net/http: ChunkedReader blocks Read until the next chunk arrives
- net/http: copy headers when following redirects
- x/net/http2: data race in request body buffer sharing
- x/crypto/chacha20poly1305: needs +build go1.7 tags
- net/http/httputil: New method ReadBody(*http.Response) ([]byte, error)
- net/http/httptrace: function called after RoundTrip returns an error
- net: new Resolver type does not respect PreferGo
- net/http: Redirect unresponding for URL containing Unicode
- net: cgoLookupAddrPTR returns single PTR regardless of number of PTR records
- net/http: client
Expect: 100-continuedocs not clear - x/net/http2: Requesting data via http2 and closing the response body early uses 3-5x more memory than via http1
- net/http: *http.Transport is never collected
- x/net/http2: The server should not select a banned cipher if it selects h2
- net/http: finishRequest() aborts ongoing background reads
- encoding/json: panic if marshaling pointer and non-pointer of same base type
- x/net/http2: does not conform to RFC strictly
- misc/cgo/fortran: test fails on linux-386
- proposal: freeze more packages
- net/http: Transport's automatic http2 too aggressive?
- net/http/httptest: Better document usage of ResponseRecorder.HeaderMap vs. ResponseRecorder.Result
- net/http/httptest: test Server blocked after 5 seconds with StateNew conns
- http2: attempt to send frame on half-closed-local stream
- net/http: http.Server.ReadTimeout never reset in http2
- net/http: add http1 and http2 Server IdleTimeout
- proposal: net: TCPConn supports Writev
Total number of label
- The number of FrozenDueToAge is 12
- The number of NeedsFix is 12
- The number of Proposal is 6
- The number of Documentation is 5
- The number of HelpWanted is 4
- The number of NeedsInvestigation is 3
- The number of Builders is 2
- The number of NeedsDecision is 2
- The number of Proposal-Accepted is 2
- The number of Testing is 2
- The number of Suggested is 2
- The number of Performance is 1
- The number of FeatureRequest is 1
- The number of Proposal-Declined is 1
- The number of Blocked is 1
Activities in Nov 2016
- net/http: add built-in graceful shutdown support to Server
- net/http: send User-Agent in proxy CONNECT
- net/http: arbitrary string between port and path in url is silently accepted
- net/http: receiving DATA after sending (RST_STREAM) kills connection.
- net/http: build test sometimes fails
- net/http: Redirect conditions are incorrect, support 307/308
- database/sql: add Context support
- net/http: http tries to reuse dead TLS connections
- database/sql: Control over connection pool
- net/http/httputil: io.Copy in function of ReverseProxy is not high efficiency
- os/exec: include stderr in ExitError
- x/build: add builders for recently added envvars
- net/http: TestLinuxSendfile is flaky
- syscall: incorrect arguments of prlimit64 syscall on linux
- net/http: test race on linux/amd64 trybot
- syscall: exec_linux.go not using cross architecture safe SYS_SETGROUPS
- net/http: panic: Fail in goroutine after TestIdleConnH2Crash has completed
- net/http: flaky "connection reset by peer" on macOS Sierra
- net/http: Server.Close does defer Lock
- time: update tzdata to IANA release 2016i (November 2016)
- net/http: Client returns "request canceled" on timeout
- net/http: canceled requests with Timeout set always return a Timeout error
- net/http: Cannot set default 'h2' and other NextProto handler concurrently
- net/http: document, audit ProtocolError
- net/http: document Server.Shutdown ignores hijacked conns
- net/http: document more how to abort a response from Handlers
- net/http: documentation for interaction between http2 imported and http pkgs.
- net/http: TestServerSetKeepAlivesEnabledClosesConns flaking with "idle count after SetKeepAlivesEnabled called = 1; want 0"
- net: TestTCPSupriousConnSetupCompletion is flaky
- net: figure out IDNA Punycode story
- x/net/http2: if http1 Server's keep-alives are disabled, only do one request
- net: Conn.Close doesn't unblock Conn.Read on nacl
- x/net/http2: RSTStream stream error write scheduling might panic with priority scheduler
- net/http: invalid method strings
- x/build/cmd/release: do not ship cmd/api binary
- x/net/ipv6: OS X 10.8 kernel panic during tests
- x/build/coordinator: re-enable plan9 support
- net/http: invalid method ""
- net/http: TestTimeoutHandlerRaceHeader flaky?
- time: Parse does not detect bad days ("00" -> returns wrong date && err=nil)
- database/sql: TestPendingConnsAfterErr flaky on darwin/arm64
- net/http, crypto/tls: cloneTLSConfig is out of sync with tls.Config
- x/tools/cmd/goimports: mishandles named imports used under original name
- x/net/http2: Transport should do transparent gzip like HTTP/1
- x/net/http2: Transport doesn't set ContentLength etc the same as HTTP/1
- x/net/http2: TestChunkedResponseHeaders fails for http2
- x/net/http2: Test304Responses_h2 fails
- database/sql: ? vs $ inconsistency
Total number of label
- The number of NeedsFix is 16
- The number of FrozenDueToAge is 14
- The number of Testing is 7
- The number of Documentation is 5
- The number of NeedsInvestigation is 5
- The number of Builders is 3
- The number of NeedsDecision is 1
- The number of WaitingForInfo is 1
- The number of HelpWanted is 1
- The number of Suggested is 1
- The number of RaceReport is 1
Activities in Dec 2016
- x/tools/cmd/godoc: enable links to fields of structs
- database/sql: NamedParam and Param are wrong names
- net/http: document Request.Context's lifetime again
- net/http: backport "multipart ReadForm close file after copy" to 1.7
- net/http: support HTTP/2
- mime: ParseMediaType parses invalid media types
- mime: Parse/Format corrupt parameters (2)
- net/http: merge http2 Transport dials
- net/http: nil pointer dereference in closeConnIfStillIdle
- x/crypto/autocert: should trim trailing dots
- net/http: NewRequest started setting ContentLength to -1 and breaking people?
- net/http: TestServerKeepAlivesEnabled_h2 test flake
- doc: installation docs don't list all OS architectures
- cmd/go: when testing, pass Stdout to test child process directly
- net: TestLookupPort_Minimal and TestResolveTCPAddr fail on android/386
- x/net/ipv4: flaky on Windows
- net/http: also make Transport retry POST request if no bytes were written
- net/http: Transport.RoundTrip regression when retrying requests on Request.Body read error
- doc: write Go 1.8 release notes
- all: inconsistent spelling: canceling or cancelling
- database/sql: Context cancellation doesn't
- x/tools/playground/socket: TestBuffer has bug and is and flaky
- net/http: continue to reduce server CPU and allocations
- net/http: setting Content-Type header to []string{""} in http handler causes two Content-Type headers to be emitted
- mime: Parse/Format corrupt parameters
- net/http: verify and test that Request.Cancel interrupts bodies too
- net/http: CloseNotifier fails to fire when underlying connection is gone
- net/http: discrepancy between HeadContentLength of "" in http2 vs http1
- net/http: discrepancy in contentLength given invalid "Content-Encoding":"foo" between http1 and http2
- net/http: TestHandlerPanicNil_h2 failing
- net/http: http2 client return EOF
- database/sql: don't encumber Context with options
- net/http: TestClientTimeout_Headers_h2 fails
- builders: upgrade OpenBSD builders to 6.0
- x/net/http2: many tests are flaky: Error while expecting an RSTStream frame: timeout waiting for frame
- x/net/http2: error dialing IPv6 literal hostnames without ports
- net/http: TestServerTimeouts flake
- crypto/x509: TestSystemRoots flaky & slow in non-cgo case
- testing: add T.Context() accessor
- net/http: TestInterruptWithPanic_h2 flakes on ppc64le builder
- website: golang.org/doc/install#help should mention Slack
- net/http: first byte of outgoing request body is sometimes lost
- x/net/http2: Transport hangs forever if using a pipe that has not been written to yet as req.Body
- net/http: document NoBody change in release notes
- net/http: does not verify validity of Host header
- net/http: data race reading a "Expect: 100-Continue" request while writing response
- net/http: TestClientTrailers_h2 is failing
- net/http: more Method=="" vs Method=="GET" inconsistency
- net/http: flake in TestHijackAfterCloseNotifier?
- net/http: allows read from Response.Body after Close
Total number of label
- The number of FrozenDueToAge is 19
- The number of NeedsFix is 10
- The number of Documentation is 7
- The number of Testing is 5
- The number of NeedsDecision is 3
- The number of NeedsInvestigation is 3
- The number of Security is 1
- The number of Proposal-Accepted is 1
- The number of Performance is 1
- The number of Builders is 1
- The number of GarbageCollector is 1
- The number of Proposal is 1
- The number of HelpWanted is 1
- The number of OS-OpenBSD is 1
- The number of Suggested is 1
Activities in Jan 2017
- net/http: Requests with already canceled context are passed to Handlers
- x/net/http2: decide what to do about Transport.CancelRequest
- net/http/httputil: ReverseProxy uses CloseNotifier before body consumption
- net/http: TestServerEmptyBodyRace_h1 flake
- net/http: Request.Context() canceled on Server.ReadTimeout
- net/http: HTTP request hangs when dumping non-nil body using httputil.DumpRequestOut
- time: update tzdata to IANA release 2016j (November 2016)
- net/http: ErrorLog not generic enough to get at the net.Conn info for errors
- net/http: allows invalid characters in header values
- net/http: TestTransportCancelBeforeResponseHeaders is flaky
- net/http: TestTransportAndServerSharedBodyRace_h2 is failing
- runtime: document missing GODEBUG settings
- net/http: HTTP/2 POST via Akamai (maybe) corrupted
- net/http: http2 transport doesn't send a default User-Agent
- net/http: call out CloseNotifier behavior change in 1.6 release notes
- net/http: background read can lose a byte
- net/http: panic: persistConn was already in LRU
- net/http: TestRetryIdempotentRequestsOnError flake
- net: ability to replace Resolver impl
Total number of label
- The number of FrozenDueToAge is 10
- The number of NeedsFix is 2
- The number of WaitingForInfo is 2
- The number of Documentation is 2
- The number of HelpWanted is 1
- The number of NeedsInvestigation is 1
- The number of NeedsDecision is 1
- The number of Testing is 1