mirror of
https://github.com/fnproject/fn.git
synced 2022-10-28 21:29:17 +03:00
* add DateTime sans mgo * change all uses of strfmt.DateTime to common.DateTime, remove test strfmt usage * remove api tests, system-test dep on api test multiple reasons to remove the api tests: * awkward dependency with fn_go meant generating bindings on a branched fn to vendor those to test new stuff. this is at a minimum not at all intuitive, worth it, nor a fun way to spend the finite amount of time we have to live. * api tests only tested a subset of functionality that the server/ api tests already test, and we risk having tests where one tests some thing and the other doesn't. let's not. we have too many test suites as it is, and these pretty much only test that we updated the fn_go bindings, which is actually a hassle as noted above and the cli will pretty quickly figure out anyway. * fn_go relies on openapi, which relies on mgo, which is deprecated and we'd like to remove as a dependency. openapi is a _huge_ dep built in a NIH fashion, that cannot simply remove the mgo dep as users may be using it. we've now stolen their date time and otherwise killed usage of it in fn core, for fn_go it still exists but that's less of a problem. * update deps removals: * easyjson * mgo * go-openapi * mapstructure * fn_go * purell * go-validator also, had to lock docker. we shouldn't use docker on master anyway, they strongly advise against that. had no luck with latest version rev, so i locked it to what we were using before. until next time. the rest is just playing dep roulette, those end up removing a ton tho * fix exec test to work * account for john le cache
337 lines
10 KiB
Go
337 lines
10 KiB
Go
// Copyright 2016 The Go Authors. All rights reserved.
|
|
// Use of this source code is governed by a BSD-style
|
|
// license that can be found in the LICENSE file.
|
|
|
|
package gensupport
|
|
|
|
import (
|
|
"bytes"
|
|
"fmt"
|
|
"io"
|
|
"io/ioutil"
|
|
"mime/multipart"
|
|
"net/http"
|
|
"net/textproto"
|
|
"strings"
|
|
"sync"
|
|
|
|
"google.golang.org/api/googleapi"
|
|
)
|
|
|
|
const sniffBuffSize = 512
|
|
|
|
func newContentSniffer(r io.Reader) *contentSniffer {
|
|
return &contentSniffer{r: r}
|
|
}
|
|
|
|
// contentSniffer wraps a Reader, and reports the content type determined by sniffing up to 512 bytes from the Reader.
|
|
type contentSniffer struct {
|
|
r io.Reader
|
|
start []byte // buffer for the sniffed bytes.
|
|
err error // set to any error encountered while reading bytes to be sniffed.
|
|
|
|
ctype string // set on first sniff.
|
|
sniffed bool // set to true on first sniff.
|
|
}
|
|
|
|
func (cs *contentSniffer) Read(p []byte) (n int, err error) {
|
|
// Ensure that the content type is sniffed before any data is consumed from Reader.
|
|
_, _ = cs.ContentType()
|
|
|
|
if len(cs.start) > 0 {
|
|
n := copy(p, cs.start)
|
|
cs.start = cs.start[n:]
|
|
return n, nil
|
|
}
|
|
|
|
// We may have read some bytes into start while sniffing, even if the read ended in an error.
|
|
// We should first return those bytes, then the error.
|
|
if cs.err != nil {
|
|
return 0, cs.err
|
|
}
|
|
|
|
// Now we have handled all bytes that were buffered while sniffing. Now just delegate to the underlying reader.
|
|
return cs.r.Read(p)
|
|
}
|
|
|
|
// ContentType returns the sniffed content type, and whether the content type was succesfully sniffed.
|
|
func (cs *contentSniffer) ContentType() (string, bool) {
|
|
if cs.sniffed {
|
|
return cs.ctype, cs.ctype != ""
|
|
}
|
|
cs.sniffed = true
|
|
// If ReadAll hits EOF, it returns err==nil.
|
|
cs.start, cs.err = ioutil.ReadAll(io.LimitReader(cs.r, sniffBuffSize))
|
|
|
|
// Don't try to detect the content type based on possibly incomplete data.
|
|
if cs.err != nil {
|
|
return "", false
|
|
}
|
|
|
|
cs.ctype = http.DetectContentType(cs.start)
|
|
return cs.ctype, true
|
|
}
|
|
|
|
// DetermineContentType determines the content type of the supplied reader.
|
|
// If the content type is already known, it can be specified via ctype.
|
|
// Otherwise, the content of media will be sniffed to determine the content type.
|
|
// If media implements googleapi.ContentTyper (deprecated), this will be used
|
|
// instead of sniffing the content.
|
|
// After calling DetectContentType the caller must not perform further reads on
|
|
// media, but rather read from the Reader that is returned.
|
|
func DetermineContentType(media io.Reader, ctype string) (io.Reader, string) {
|
|
// Note: callers could avoid calling DetectContentType if ctype != "",
|
|
// but doing the check inside this function reduces the amount of
|
|
// generated code.
|
|
if ctype != "" {
|
|
return media, ctype
|
|
}
|
|
|
|
// For backwards compatability, allow clients to set content
|
|
// type by providing a ContentTyper for media.
|
|
if typer, ok := media.(googleapi.ContentTyper); ok {
|
|
return media, typer.ContentType()
|
|
}
|
|
|
|
sniffer := newContentSniffer(media)
|
|
if ctype, ok := sniffer.ContentType(); ok {
|
|
return sniffer, ctype
|
|
}
|
|
// If content type could not be sniffed, reads from sniffer will eventually fail with an error.
|
|
return sniffer, ""
|
|
}
|
|
|
|
type typeReader struct {
|
|
io.Reader
|
|
typ string
|
|
}
|
|
|
|
// multipartReader combines the contents of multiple readers to create a multipart/related HTTP body.
|
|
// Close must be called if reads from the multipartReader are abandoned before reaching EOF.
|
|
type multipartReader struct {
|
|
pr *io.PipeReader
|
|
ctype string
|
|
mu sync.Mutex
|
|
pipeOpen bool
|
|
}
|
|
|
|
func newMultipartReader(parts []typeReader) *multipartReader {
|
|
mp := &multipartReader{pipeOpen: true}
|
|
var pw *io.PipeWriter
|
|
mp.pr, pw = io.Pipe()
|
|
mpw := multipart.NewWriter(pw)
|
|
mp.ctype = "multipart/related; boundary=" + mpw.Boundary()
|
|
go func() {
|
|
for _, part := range parts {
|
|
w, err := mpw.CreatePart(typeHeader(part.typ))
|
|
if err != nil {
|
|
mpw.Close()
|
|
pw.CloseWithError(fmt.Errorf("googleapi: CreatePart failed: %v", err))
|
|
return
|
|
}
|
|
_, err = io.Copy(w, part.Reader)
|
|
if err != nil {
|
|
mpw.Close()
|
|
pw.CloseWithError(fmt.Errorf("googleapi: Copy failed: %v", err))
|
|
return
|
|
}
|
|
}
|
|
|
|
mpw.Close()
|
|
pw.Close()
|
|
}()
|
|
return mp
|
|
}
|
|
|
|
func (mp *multipartReader) Read(data []byte) (n int, err error) {
|
|
return mp.pr.Read(data)
|
|
}
|
|
|
|
func (mp *multipartReader) Close() error {
|
|
mp.mu.Lock()
|
|
if !mp.pipeOpen {
|
|
mp.mu.Unlock()
|
|
return nil
|
|
}
|
|
mp.pipeOpen = false
|
|
mp.mu.Unlock()
|
|
return mp.pr.Close()
|
|
}
|
|
|
|
// CombineBodyMedia combines a json body with media content to create a multipart/related HTTP body.
|
|
// It returns a ReadCloser containing the combined body, and the overall "multipart/related" content type, with random boundary.
|
|
//
|
|
// The caller must call Close on the returned ReadCloser if reads are abandoned before reaching EOF.
|
|
func CombineBodyMedia(body io.Reader, bodyContentType string, media io.Reader, mediaContentType string) (io.ReadCloser, string) {
|
|
mp := newMultipartReader([]typeReader{
|
|
{body, bodyContentType},
|
|
{media, mediaContentType},
|
|
})
|
|
return mp, mp.ctype
|
|
}
|
|
|
|
func typeHeader(contentType string) textproto.MIMEHeader {
|
|
h := make(textproto.MIMEHeader)
|
|
if contentType != "" {
|
|
h.Set("Content-Type", contentType)
|
|
}
|
|
return h
|
|
}
|
|
|
|
// PrepareUpload determines whether the data in the supplied reader should be
|
|
// uploaded in a single request, or in sequential chunks.
|
|
// chunkSize is the size of the chunk that media should be split into.
|
|
//
|
|
// If chunkSize is zero, media is returned as the first value, and the other
|
|
// two return values are nil, true.
|
|
//
|
|
// Otherwise, a MediaBuffer is returned, along with a bool indicating whether the
|
|
// contents of media fit in a single chunk.
|
|
//
|
|
// After PrepareUpload has been called, media should no longer be used: the
|
|
// media content should be accessed via one of the return values.
|
|
func PrepareUpload(media io.Reader, chunkSize int) (r io.Reader, mb *MediaBuffer, singleChunk bool) {
|
|
if chunkSize == 0 { // do not chunk
|
|
return media, nil, true
|
|
}
|
|
mb = NewMediaBuffer(media, chunkSize)
|
|
_, _, _, err := mb.Chunk()
|
|
// If err is io.EOF, we can upload this in a single request. Otherwise, err is
|
|
// either nil or a non-EOF error. If it is the latter, then the next call to
|
|
// mb.Chunk will return the same error. Returning a MediaBuffer ensures that this
|
|
// error will be handled at some point.
|
|
return nil, mb, err == io.EOF
|
|
}
|
|
|
|
// MediaInfo holds information for media uploads. It is intended for use by generated
|
|
// code only.
|
|
type MediaInfo struct {
|
|
// At most one of Media and MediaBuffer will be set.
|
|
media io.Reader
|
|
buffer *MediaBuffer
|
|
singleChunk bool
|
|
mType string
|
|
size int64 // mediaSize, if known. Used only for calls to progressUpdater_.
|
|
progressUpdater googleapi.ProgressUpdater
|
|
}
|
|
|
|
// NewInfoFromMedia should be invoked from the Media method of a call. It returns a
|
|
// MediaInfo populated with chunk size and content type, and a reader or MediaBuffer
|
|
// if needed.
|
|
func NewInfoFromMedia(r io.Reader, options []googleapi.MediaOption) *MediaInfo {
|
|
mi := &MediaInfo{}
|
|
opts := googleapi.ProcessMediaOptions(options)
|
|
if !opts.ForceEmptyContentType {
|
|
r, mi.mType = DetermineContentType(r, opts.ContentType)
|
|
}
|
|
mi.media, mi.buffer, mi.singleChunk = PrepareUpload(r, opts.ChunkSize)
|
|
return mi
|
|
}
|
|
|
|
// NewInfoFromResumableMedia should be invoked from the ResumableMedia method of a
|
|
// call. It returns a MediaInfo using the given reader, size and media type.
|
|
func NewInfoFromResumableMedia(r io.ReaderAt, size int64, mediaType string) *MediaInfo {
|
|
rdr := ReaderAtToReader(r, size)
|
|
rdr, mType := DetermineContentType(rdr, mediaType)
|
|
return &MediaInfo{
|
|
size: size,
|
|
mType: mType,
|
|
buffer: NewMediaBuffer(rdr, googleapi.DefaultUploadChunkSize),
|
|
media: nil,
|
|
singleChunk: false,
|
|
}
|
|
}
|
|
|
|
func (mi *MediaInfo) SetProgressUpdater(pu googleapi.ProgressUpdater) {
|
|
if mi != nil {
|
|
mi.progressUpdater = pu
|
|
}
|
|
}
|
|
|
|
// UploadType determines the type of upload: a single request, or a resumable
|
|
// series of requests.
|
|
func (mi *MediaInfo) UploadType() string {
|
|
if mi.singleChunk {
|
|
return "multipart"
|
|
}
|
|
return "resumable"
|
|
}
|
|
|
|
// UploadRequest sets up an HTTP request for media upload. It adds headers
|
|
// as necessary, and returns a replacement for the body and a function for http.Request.GetBody.
|
|
func (mi *MediaInfo) UploadRequest(reqHeaders http.Header, body io.Reader) (newBody io.Reader, getBody func() (io.ReadCloser, error), cleanup func()) {
|
|
cleanup = func() {}
|
|
if mi == nil {
|
|
return body, nil, cleanup
|
|
}
|
|
var media io.Reader
|
|
if mi.media != nil {
|
|
// This only happens when the caller has turned off chunking. In that
|
|
// case, we write all of media in a single non-retryable request.
|
|
media = mi.media
|
|
} else if mi.singleChunk {
|
|
// The data fits in a single chunk, which has now been read into the MediaBuffer.
|
|
// We obtain that chunk so we can write it in a single request. The request can
|
|
// be retried because the data is stored in the MediaBuffer.
|
|
media, _, _, _ = mi.buffer.Chunk()
|
|
}
|
|
if media != nil {
|
|
fb := readerFunc(body)
|
|
fm := readerFunc(media)
|
|
combined, ctype := CombineBodyMedia(body, "application/json", media, mi.mType)
|
|
if fb != nil && fm != nil {
|
|
getBody = func() (io.ReadCloser, error) {
|
|
rb := ioutil.NopCloser(fb())
|
|
rm := ioutil.NopCloser(fm())
|
|
r, _ := CombineBodyMedia(rb, "application/json", rm, mi.mType)
|
|
return r, nil
|
|
}
|
|
}
|
|
cleanup = func() { combined.Close() }
|
|
reqHeaders.Set("Content-Type", ctype)
|
|
body = combined
|
|
}
|
|
if mi.buffer != nil && mi.mType != "" && !mi.singleChunk {
|
|
reqHeaders.Set("X-Upload-Content-Type", mi.mType)
|
|
}
|
|
return body, getBody, cleanup
|
|
}
|
|
|
|
// readerFunc returns a function that always returns an io.Reader that has the same
|
|
// contents as r, provided that can be done without consuming r. Otherwise, it
|
|
// returns nil.
|
|
// See http.NewRequest (in net/http/request.go).
|
|
func readerFunc(r io.Reader) func() io.Reader {
|
|
switch r := r.(type) {
|
|
case *bytes.Buffer:
|
|
buf := r.Bytes()
|
|
return func() io.Reader { return bytes.NewReader(buf) }
|
|
case *bytes.Reader:
|
|
snapshot := *r
|
|
return func() io.Reader { r := snapshot; return &r }
|
|
case *strings.Reader:
|
|
snapshot := *r
|
|
return func() io.Reader { r := snapshot; return &r }
|
|
default:
|
|
return nil
|
|
}
|
|
}
|
|
|
|
// ResumableUpload returns an appropriately configured ResumableUpload value if the
|
|
// upload is resumable, or nil otherwise.
|
|
func (mi *MediaInfo) ResumableUpload(locURI string) *ResumableUpload {
|
|
if mi == nil || mi.singleChunk {
|
|
return nil
|
|
}
|
|
return &ResumableUpload{
|
|
URI: locURI,
|
|
Media: mi.buffer,
|
|
MediaType: mi.mType,
|
|
Callback: func(curr int64) {
|
|
if mi.progressUpdater != nil {
|
|
mi.progressUpdater(curr, mi.size)
|
|
}
|
|
},
|
|
}
|
|
}
|