curl/lib/request.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

219 lines
8.5 KiB
C
Raw Normal View History

#ifndef HEADER_CURL_REQUEST_H
#define HEADER_CURL_REQUEST_H
/***************************************************************************
* _ _ ____ _
* Project ___| | | | _ \| |
* / __| | | | |_) | |
* | (__| |_| | _ <| |___
* \___|\___/|_| \_\_____|
*
* Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al.
*
* This software is licensed as described in the file COPYING, which
* you should have received as part of this distribution. The terms
* are also available at https://curl.se/docs/copyright.html.
*
* You may opt to use, copy, modify, merge, publish, distribute and/or sell
* copies of the Software, and permit persons to whom the Software is
* furnished to do so, under the terms of the COPYING file.
*
* This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
* KIND, either express or implied.
*
* SPDX-License-Identifier: curl
*
***************************************************************************/
/* This file is for lib internal stuff */
#include "curl_setup.h"
#include "bufq.h"
/* forward declarations */
struct UserDefined;
enum expect100 {
EXP100_SEND_DATA, /* enough waiting, just send the body now */
EXP100_AWAITING_CONTINUE, /* waiting for the 100 Continue header */
EXP100_SENDING_REQUEST, /* still sending the request but will wait for
the 100 header once done with the request */
EXP100_FAILED /* used on 417 Expectation Failed */
};
enum upgrade101 {
UPGR101_INIT, /* default state */
UPGR101_WS, /* upgrade to WebSockets requested */
UPGR101_H2, /* upgrade to HTTP/2 requested */
UPGR101_RECEIVED, /* 101 response received */
UPGR101_WORKING /* talking upgraded protocol */
};
/*
* Request specific data in the easy handle (Curl_easy). Previously,
* these members were on the connectdata struct but since a conn struct may
* now be shared between different Curl_easys, we store connection-specific
* data here. This struct only keeps stuff that's interesting for *this*
* request, as it will be cleared between multiple ones
*/
struct SingleRequest {
curl_off_t size; /* -1 if unknown at this point */
curl_off_t maxdownload; /* in bytes, the maximum amount of data to fetch,
-1 means unlimited */
curl_off_t bytecount; /* total number of bytes read */
curl_off_t writebytecount; /* number of bytes written */
struct curltime start; /* transfer started at this time */
unsigned int headerbytecount; /* received server headers (not CONNECT
headers) */
unsigned int allheadercount; /* all received headers (server + CONNECT) */
unsigned int deductheadercount; /* this amount of bytes doesn't count when
we check if anything has been transferred
at the end of a connection. We use this
counter to make only a 100 reply (without
a following second response code) result
in a CURLE_GOT_NOTHING error code */
int headerline; /* counts header lines to better track the
first one */
curl_off_t offset; /* possible resume offset read from the
Content-Range: header */
int httpcode; /* error code from the 'HTTP/1.? XXX' or
'RTSP/1.? XXX' line */
int keepon;
struct curltime start100; /* time stamp to wait for the 100 code from */
enum expect100 exp100; /* expect 100 continue state */
enum upgrade101 upgr101; /* 101 upgrade state */
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk (to be made into a sperate PR, also) Added as documented [in CLIENT-READER.md](https://github.com/curl/curl/blob/5b1f31dfbab8aef467c419c68aa06dc738cb75d4/docs/CLIENT-READERS.md). - old `Curl_buffer_send()` completely replaced by new `Curl_req_send()` - old `Curl_fillreadbuffer()` replaced with `Curl_client_read()` - HTTP chunked uploads are now formatted in a client reader added when needed. - FTP line-end conversions are done in a client reader added when needed. - when sending requests headers, remaining buffer space is filled with body data for sending in "one go". This is independent of the request body size. Resolves #12938 as now small and large requests have the same code path. Changes done to test cases: - test513: now fails before sending request headers as this initial "client read" triggers the setup fault. Behaves now the same as in hyper build - test547, test555, test1620: fix the length check in the lib code to only fail for reads *smaller* than expected. This was a bug in the test code that never triggered in the old implementation. Closes #12969
2024-02-15 23:22:53 +08:00
/* Client Writer stack, handles transfer- and content-encodings, protocol
* checks, pausing by client callbacks. */
struct Curl_cwriter *writer_stack;
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk (to be made into a sperate PR, also) Added as documented [in CLIENT-READER.md](https://github.com/curl/curl/blob/5b1f31dfbab8aef467c419c68aa06dc738cb75d4/docs/CLIENT-READERS.md). - old `Curl_buffer_send()` completely replaced by new `Curl_req_send()` - old `Curl_fillreadbuffer()` replaced with `Curl_client_read()` - HTTP chunked uploads are now formatted in a client reader added when needed. - FTP line-end conversions are done in a client reader added when needed. - when sending requests headers, remaining buffer space is filled with body data for sending in "one go". This is independent of the request body size. Resolves #12938 as now small and large requests have the same code path. Changes done to test cases: - test513: now fails before sending request headers as this initial "client read" triggers the setup fault. Behaves now the same as in hyper build - test547, test555, test1620: fix the length check in the lib code to only fail for reads *smaller* than expected. This was a bug in the test code that never triggered in the old implementation. Closes #12969
2024-02-15 23:22:53 +08:00
/* Client Reader stack, handles transfer- and content-encodings, protocol
* checks, pausing by client callbacks. */
struct Curl_creader *reader_stack;
struct bufq sendbuf; /* data which needs to be send to the server */
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk The method for sending not just raw bytes, but bytes that are either "headers" or "body". The send abstraction stack, to to bottom, now is: * `Curl_req_send()`: has parameter to indicate amount of header bytes, buffers all data. * `Curl_xfer_send()`: knows on which socket index to send, returns amount of bytes sent. * `Curl_conn_send()`: called with socket index, returns amount of bytes sent. In addition there is `Curl_req_flush()` for writing out all buffered bytes. `Curl_req_send()` is active for requests without body, `Curl_buffer_send()` still being used for others. This is because the special quirks need to be addressed in future parts: * `expect-100` handling * `Curl_fillreadbuffer()` needs to add directly to the new `data->req.sendbuf` * special body handlings, like `chunked` encodings and line end conversions will be moved into something like a Client Reader. In functions of the pattern `CURLcode xxx_send(..., ssize_t *written)`, replace the `ssize_t` with a `size_t`. It makes no sense to allow for negative values as the returned `CURLcode` already specifies error conditions. This allows easier handling of lengths without casting. Closes #12964
2024-02-15 23:22:53 +08:00
size_t sendbuf_hds_len; /* amount of header bytes in sendbuf */
time_t timeofdoc;
long bodywrites;
char *location; /* This points to an allocated version of the Location:
header data */
char *newurl; /* Set to the new URL to use when a redirect or a retry is
wanted */
/* Allocated protocol-specific data. Each protocol handler makes sure this
points to data it needs. */
union {
struct FILEPROTO *file;
struct FTP *ftp;
struct HTTP *http;
struct IMAP *imap;
struct ldapreqinfo *ldap;
struct MQTT *mqtt;
struct POP3 *pop3;
struct RTSP *rtsp;
struct smb_request *smb;
struct SMTP *smtp;
struct SSHPROTO *ssh;
struct TELNET *telnet;
} p;
#ifndef CURL_DISABLE_DOH
struct dohdata *doh; /* DoH specific data for this request */
#endif
#ifndef CURL_DISABLE_COOKIES
unsigned char setcookies;
#endif
BIT(header); /* incoming data has HTTP header */
BIT(content_range); /* set TRUE if Content-Range: was found */
BIT(download_done); /* set to TRUE when download is complete */
BIT(eos_written); /* iff EOS has been written to client */
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk (to be made into a sperate PR, also) Added as documented [in CLIENT-READER.md](https://github.com/curl/curl/blob/5b1f31dfbab8aef467c419c68aa06dc738cb75d4/docs/CLIENT-READERS.md). - old `Curl_buffer_send()` completely replaced by new `Curl_req_send()` - old `Curl_fillreadbuffer()` replaced with `Curl_client_read()` - HTTP chunked uploads are now formatted in a client reader added when needed. - FTP line-end conversions are done in a client reader added when needed. - when sending requests headers, remaining buffer space is filled with body data for sending in "one go". This is independent of the request body size. Resolves #12938 as now small and large requests have the same code path. Changes done to test cases: - test513: now fails before sending request headers as this initial "client read" triggers the setup fault. Behaves now the same as in hyper build - test547, test555, test1620: fix the length check in the lib code to only fail for reads *smaller* than expected. This was a bug in the test code that never triggered in the old implementation. Closes #12969
2024-02-15 23:22:53 +08:00
BIT(eos_read); /* iff EOS has been read from the client */
BIT(rewind_read); /* iff reader needs rewind at next start */
BIT(upload_done); /* set to TRUE when all request data has been sent */
BIT(upload_aborted); /* set to TRUE when upload was aborted. Will also
* show `upload_done` as TRUE. */
BIT(ignorebody); /* we read a response-body but we ignore it! */
BIT(http_bodyless); /* HTTP response status code is between 100 and 199,
204 or 304 */
BIT(chunk); /* if set, this is a chunked transfer-encoding */
BIT(ignore_cl); /* ignore content-length */
BIT(upload_chunky); /* set TRUE if we are doing chunked transfer-encoding
on upload */
BIT(getheader); /* TRUE if header parsing is wanted */
BIT(no_body); /* the response has no body */
BIT(authneg); /* TRUE when the auth phase has started, which means
that we are creating a request with an auth header,
but it is not the final request in the auth
negotiation. */
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk The method for sending not just raw bytes, but bytes that are either "headers" or "body". The send abstraction stack, to to bottom, now is: * `Curl_req_send()`: has parameter to indicate amount of header bytes, buffers all data. * `Curl_xfer_send()`: knows on which socket index to send, returns amount of bytes sent. * `Curl_conn_send()`: called with socket index, returns amount of bytes sent. In addition there is `Curl_req_flush()` for writing out all buffered bytes. `Curl_req_send()` is active for requests without body, `Curl_buffer_send()` still being used for others. This is because the special quirks need to be addressed in future parts: * `expect-100` handling * `Curl_fillreadbuffer()` needs to add directly to the new `data->req.sendbuf` * special body handlings, like `chunked` encodings and line end conversions will be moved into something like a Client Reader. In functions of the pattern `CURLcode xxx_send(..., ssize_t *written)`, replace the `ssize_t` with a `size_t`. It makes no sense to allow for negative values as the returned `CURLcode` already specifies error conditions. This allows easier handling of lengths without casting. Closes #12964
2024-02-15 23:22:53 +08:00
BIT(sendbuf_init); /* sendbuf is initialized */
};
/**
* Initialize the state of the request for first use.
*/
CURLcode Curl_req_init(struct SingleRequest *req);
/**
* The request is about to start.
*/
CURLcode Curl_req_start(struct SingleRequest *req,
struct Curl_easy *data);
/**
* The request is done. If not aborted, make sure that buffers are
* flushed to the client.
* @param req the request
* @param data the transfer
* @param aborted TRUE iff the request was aborted/errored
*/
CURLcode Curl_req_done(struct SingleRequest *req,
struct Curl_easy *data, bool aborted);
/**
* Free the state of the request, not usable afterwards.
*/
void Curl_req_free(struct SingleRequest *req, struct Curl_easy *data);
/**
* Reset the state of the request for new use, given the
* settings.
*/
void Curl_req_reset(struct SingleRequest *req, struct Curl_easy *data);
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk (to be made into a sperate PR, also) Added as documented [in CLIENT-READER.md](https://github.com/curl/curl/blob/5b1f31dfbab8aef467c419c68aa06dc738cb75d4/docs/CLIENT-READERS.md). - old `Curl_buffer_send()` completely replaced by new `Curl_req_send()` - old `Curl_fillreadbuffer()` replaced with `Curl_client_read()` - HTTP chunked uploads are now formatted in a client reader added when needed. - FTP line-end conversions are done in a client reader added when needed. - when sending requests headers, remaining buffer space is filled with body data for sending in "one go". This is independent of the request body size. Resolves #12938 as now small and large requests have the same code path. Changes done to test cases: - test513: now fails before sending request headers as this initial "client read" triggers the setup fault. Behaves now the same as in hyper build - test547, test555, test1620: fix the length check in the lib code to only fail for reads *smaller* than expected. This was a bug in the test code that never triggered in the old implementation. Closes #12969
2024-02-15 23:22:53 +08:00
#ifndef USE_HYPER
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk The method for sending not just raw bytes, but bytes that are either "headers" or "body". The send abstraction stack, to to bottom, now is: * `Curl_req_send()`: has parameter to indicate amount of header bytes, buffers all data. * `Curl_xfer_send()`: knows on which socket index to send, returns amount of bytes sent. * `Curl_conn_send()`: called with socket index, returns amount of bytes sent. In addition there is `Curl_req_flush()` for writing out all buffered bytes. `Curl_req_send()` is active for requests without body, `Curl_buffer_send()` still being used for others. This is because the special quirks need to be addressed in future parts: * `expect-100` handling * `Curl_fillreadbuffer()` needs to add directly to the new `data->req.sendbuf` * special body handlings, like `chunked` encodings and line end conversions will be moved into something like a Client Reader. In functions of the pattern `CURLcode xxx_send(..., ssize_t *written)`, replace the `ssize_t` with a `size_t`. It makes no sense to allow for negative values as the returned `CURLcode` already specifies error conditions. This allows easier handling of lengths without casting. Closes #12964
2024-02-15 23:22:53 +08:00
/**
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk (to be made into a sperate PR, also) Added as documented [in CLIENT-READER.md](https://github.com/curl/curl/blob/5b1f31dfbab8aef467c419c68aa06dc738cb75d4/docs/CLIENT-READERS.md). - old `Curl_buffer_send()` completely replaced by new `Curl_req_send()` - old `Curl_fillreadbuffer()` replaced with `Curl_client_read()` - HTTP chunked uploads are now formatted in a client reader added when needed. - FTP line-end conversions are done in a client reader added when needed. - when sending requests headers, remaining buffer space is filled with body data for sending in "one go". This is independent of the request body size. Resolves #12938 as now small and large requests have the same code path. Changes done to test cases: - test513: now fails before sending request headers as this initial "client read" triggers the setup fault. Behaves now the same as in hyper build - test547, test555, test1620: fix the length check in the lib code to only fail for reads *smaller* than expected. This was a bug in the test code that never triggered in the old implementation. Closes #12969
2024-02-15 23:22:53 +08:00
* Send request headers. If not all could be sent
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk The method for sending not just raw bytes, but bytes that are either "headers" or "body". The send abstraction stack, to to bottom, now is: * `Curl_req_send()`: has parameter to indicate amount of header bytes, buffers all data. * `Curl_xfer_send()`: knows on which socket index to send, returns amount of bytes sent. * `Curl_conn_send()`: called with socket index, returns amount of bytes sent. In addition there is `Curl_req_flush()` for writing out all buffered bytes. `Curl_req_send()` is active for requests without body, `Curl_buffer_send()` still being used for others. This is because the special quirks need to be addressed in future parts: * `expect-100` handling * `Curl_fillreadbuffer()` needs to add directly to the new `data->req.sendbuf` * special body handlings, like `chunked` encodings and line end conversions will be moved into something like a Client Reader. In functions of the pattern `CURLcode xxx_send(..., ssize_t *written)`, replace the `ssize_t` with a `size_t`. It makes no sense to allow for negative values as the returned `CURLcode` already specifies error conditions. This allows easier handling of lengths without casting. Closes #12964
2024-02-15 23:22:53 +08:00
* they will be buffered. Use `Curl_req_flush()` to make sure
* bytes are really send.
* @param data the transfer making the request
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk (to be made into a sperate PR, also) Added as documented [in CLIENT-READER.md](https://github.com/curl/curl/blob/5b1f31dfbab8aef467c419c68aa06dc738cb75d4/docs/CLIENT-READERS.md). - old `Curl_buffer_send()` completely replaced by new `Curl_req_send()` - old `Curl_fillreadbuffer()` replaced with `Curl_client_read()` - HTTP chunked uploads are now formatted in a client reader added when needed. - FTP line-end conversions are done in a client reader added when needed. - when sending requests headers, remaining buffer space is filled with body data for sending in "one go". This is independent of the request body size. Resolves #12938 as now small and large requests have the same code path. Changes done to test cases: - test513: now fails before sending request headers as this initial "client read" triggers the setup fault. Behaves now the same as in hyper build - test547, test555, test1620: fix the length check in the lib code to only fail for reads *smaller* than expected. This was a bug in the test code that never triggered in the old implementation. Closes #12969
2024-02-15 23:22:53 +08:00
* @param buf the complete header bytes, no body
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk The method for sending not just raw bytes, but bytes that are either "headers" or "body". The send abstraction stack, to to bottom, now is: * `Curl_req_send()`: has parameter to indicate amount of header bytes, buffers all data. * `Curl_xfer_send()`: knows on which socket index to send, returns amount of bytes sent. * `Curl_conn_send()`: called with socket index, returns amount of bytes sent. In addition there is `Curl_req_flush()` for writing out all buffered bytes. `Curl_req_send()` is active for requests without body, `Curl_buffer_send()` still being used for others. This is because the special quirks need to be addressed in future parts: * `expect-100` handling * `Curl_fillreadbuffer()` needs to add directly to the new `data->req.sendbuf` * special body handlings, like `chunked` encodings and line end conversions will be moved into something like a Client Reader. In functions of the pattern `CURLcode xxx_send(..., ssize_t *written)`, replace the `ssize_t` with a `size_t`. It makes no sense to allow for negative values as the returned `CURLcode` already specifies error conditions. This allows easier handling of lengths without casting. Closes #12964
2024-02-15 23:22:53 +08:00
* @return CURLE_OK (on blocking with *pnwritten == 0) or error.
*/
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk (to be made into a sperate PR, also) Added as documented [in CLIENT-READER.md](https://github.com/curl/curl/blob/5b1f31dfbab8aef467c419c68aa06dc738cb75d4/docs/CLIENT-READERS.md). - old `Curl_buffer_send()` completely replaced by new `Curl_req_send()` - old `Curl_fillreadbuffer()` replaced with `Curl_client_read()` - HTTP chunked uploads are now formatted in a client reader added when needed. - FTP line-end conversions are done in a client reader added when needed. - when sending requests headers, remaining buffer space is filled with body data for sending in "one go". This is independent of the request body size. Resolves #12938 as now small and large requests have the same code path. Changes done to test cases: - test513: now fails before sending request headers as this initial "client read" triggers the setup fault. Behaves now the same as in hyper build - test547, test555, test1620: fix the length check in the lib code to only fail for reads *smaller* than expected. This was a bug in the test code that never triggered in the old implementation. Closes #12969
2024-02-15 23:22:53 +08:00
CURLcode Curl_req_send(struct Curl_easy *data, struct dynbuf *buf);
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk The method for sending not just raw bytes, but bytes that are either "headers" or "body". The send abstraction stack, to to bottom, now is: * `Curl_req_send()`: has parameter to indicate amount of header bytes, buffers all data. * `Curl_xfer_send()`: knows on which socket index to send, returns amount of bytes sent. * `Curl_conn_send()`: called with socket index, returns amount of bytes sent. In addition there is `Curl_req_flush()` for writing out all buffered bytes. `Curl_req_send()` is active for requests without body, `Curl_buffer_send()` still being used for others. This is because the special quirks need to be addressed in future parts: * `expect-100` handling * `Curl_fillreadbuffer()` needs to add directly to the new `data->req.sendbuf` * special body handlings, like `chunked` encodings and line end conversions will be moved into something like a Client Reader. In functions of the pattern `CURLcode xxx_send(..., ssize_t *written)`, replace the `ssize_t` with a `size_t`. It makes no sense to allow for negative values as the returned `CURLcode` already specifies error conditions. This allows easier handling of lengths without casting. Closes #12964
2024-02-15 23:22:53 +08:00
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk (to be made into a sperate PR, also) Added as documented [in CLIENT-READER.md](https://github.com/curl/curl/blob/5b1f31dfbab8aef467c419c68aa06dc738cb75d4/docs/CLIENT-READERS.md). - old `Curl_buffer_send()` completely replaced by new `Curl_req_send()` - old `Curl_fillreadbuffer()` replaced with `Curl_client_read()` - HTTP chunked uploads are now formatted in a client reader added when needed. - FTP line-end conversions are done in a client reader added when needed. - when sending requests headers, remaining buffer space is filled with body data for sending in "one go". This is independent of the request body size. Resolves #12938 as now small and large requests have the same code path. Changes done to test cases: - test513: now fails before sending request headers as this initial "client read" triggers the setup fault. Behaves now the same as in hyper build - test547, test555, test1620: fix the length check in the lib code to only fail for reads *smaller* than expected. This was a bug in the test code that never triggered in the old implementation. Closes #12969
2024-02-15 23:22:53 +08:00
#endif /* !USE_HYPER */
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk The method for sending not just raw bytes, but bytes that are either "headers" or "body". The send abstraction stack, to to bottom, now is: * `Curl_req_send()`: has parameter to indicate amount of header bytes, buffers all data. * `Curl_xfer_send()`: knows on which socket index to send, returns amount of bytes sent. * `Curl_conn_send()`: called with socket index, returns amount of bytes sent. In addition there is `Curl_req_flush()` for writing out all buffered bytes. `Curl_req_send()` is active for requests without body, `Curl_buffer_send()` still being used for others. This is because the special quirks need to be addressed in future parts: * `expect-100` handling * `Curl_fillreadbuffer()` needs to add directly to the new `data->req.sendbuf` * special body handlings, like `chunked` encodings and line end conversions will be moved into something like a Client Reader. In functions of the pattern `CURLcode xxx_send(..., ssize_t *written)`, replace the `ssize_t` with a `size_t`. It makes no sense to allow for negative values as the returned `CURLcode` already specifies error conditions. This allows easier handling of lengths without casting. Closes #12964
2024-02-15 23:22:53 +08:00
/**
* TRUE iff the request has sent all request headers and data.
*/
bool Curl_req_done_sending(struct Curl_easy *data);
/*
* Read more from client and flush all buffered request bytes.
* @return CURLE_OK on success or the error on the sending.
* Never returns CURLE_AGAIN.
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk The method for sending not just raw bytes, but bytes that are either "headers" or "body". The send abstraction stack, to to bottom, now is: * `Curl_req_send()`: has parameter to indicate amount of header bytes, buffers all data. * `Curl_xfer_send()`: knows on which socket index to send, returns amount of bytes sent. * `Curl_conn_send()`: called with socket index, returns amount of bytes sent. In addition there is `Curl_req_flush()` for writing out all buffered bytes. `Curl_req_send()` is active for requests without body, `Curl_buffer_send()` still being used for others. This is because the special quirks need to be addressed in future parts: * `expect-100` handling * `Curl_fillreadbuffer()` needs to add directly to the new `data->req.sendbuf` * special body handlings, like `chunked` encodings and line end conversions will be moved into something like a Client Reader. In functions of the pattern `CURLcode xxx_send(..., ssize_t *written)`, replace the `ssize_t` with a `size_t`. It makes no sense to allow for negative values as the returned `CURLcode` already specifies error conditions. This allows easier handling of lengths without casting. Closes #12964
2024-02-15 23:22:53 +08:00
*/
CURLcode Curl_req_send_more(struct Curl_easy *data);
lib: Curl_read/Curl_write clarifications - replace `Curl_read()`, `Curl_write()` and `Curl_nwrite()` to clarify when and at what level they operate - send/recv of transfer related data is now done via `Curl_xfer_send()/Curl_xfer_recv()` which no longer has socket/socketindex as parameter. It decides on the transfer setup of `conn->sockfd` and `conn->writesockfd` on which connection filter chain to operate. - send/recv on a specific connection filter chain is done via `Curl_conn_send()/Curl_conn_recv()` which get the socket index as parameter. - rename `Curl_setup_transfer()` to `Curl_xfer_setup()` for naming consistency - clarify that the special CURLE_AGAIN hangling to return `CURLE_OK` with length 0 only applies to `Curl_xfer_send()` and CURLE_AGAIN is returned by all other send() variants. - fix a bug in websocket `curl_ws_recv()` that mixed up data when it arrived in more than a single chunk The method for sending not just raw bytes, but bytes that are either "headers" or "body". The send abstraction stack, to to bottom, now is: * `Curl_req_send()`: has parameter to indicate amount of header bytes, buffers all data. * `Curl_xfer_send()`: knows on which socket index to send, returns amount of bytes sent. * `Curl_conn_send()`: called with socket index, returns amount of bytes sent. In addition there is `Curl_req_flush()` for writing out all buffered bytes. `Curl_req_send()` is active for requests without body, `Curl_buffer_send()` still being used for others. This is because the special quirks need to be addressed in future parts: * `expect-100` handling * `Curl_fillreadbuffer()` needs to add directly to the new `data->req.sendbuf` * special body handlings, like `chunked` encodings and line end conversions will be moved into something like a Client Reader. In functions of the pattern `CURLcode xxx_send(..., ssize_t *written)`, replace the `ssize_t` with a `size_t`. It makes no sense to allow for negative values as the returned `CURLcode` already specifies error conditions. This allows easier handling of lengths without casting. Closes #12964
2024-02-15 23:22:53 +08:00
/**
* TRUE iff the request wants to send, e.g. has buffered bytes.
*/
bool Curl_req_want_send(struct Curl_easy *data);
/**
* Stop sending any more request data to the server.
* Will clear the send buffer and mark request sending as done.
*/
CURLcode Curl_req_abort_sending(struct Curl_easy *data);
#endif /* HEADER_CURL_REQUEST_H */