curl/lib
Daniel Stenberg ed4eecc05e multi: don't expire timeouts at disonnect or done
The functions Curl_disconnect() and Curl_done() are both used within the
scope of a single request so they cannot be allowed to use
Curl_expire(... 0) to kill all timeouts as there are some timeouts that
are set before a request that are supposed to remain until the request
is done.

The timeouts are now instead cleared at curl_easy_cleanup() and when the
multi state machine changes a handle to the complete state.
2010-09-28 23:49:32 +02:00
..
.gitignore
amigaos.c
amigaos.h
arpa_telnet.h
base64.c
CMakeLists.txt
config-amigaos.h
config-mac.h
config-os400.h
config-riscos.h
config-symbian.h
config-tpf.h
config-vxworks.h
config-win32.h
config-win32ce.h
config.dos
connect.c
connect.h portabilty: use proper variable type to hold sockets 2010-09-06 00:02:54 +02:00
content_encoding.c
content_encoding.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
cookie.c
cookie.h
curl_addrinfo.c
curl_addrinfo.h
curl_base64.h
curl_config.h.cmake
curl_fnmatch.c
curl_fnmatch.h
curl_gethostname.c
curl_gethostname.h
curl_hmac.h
curl_ldap.h
curl_md4.h
curl_md5.h
curl_memory.h
curl_memrchr.c
curl_memrchr.h
curl_rand.c
curl_rand.h
curl_rtmp.c
curl_rtmp.h
curl_sspi.c
curl_sspi.h
curl_threads.c
curl_threads.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
curlx.h
dict.c Curl_setup_transfer: no longer returns anything 2010-04-16 23:43:04 +02:00
dict.h
easy.c duphandle: use ares_dup() 2010-09-21 16:41:14 +02:00
easyif.h
escape.c
escape.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
file.c file_range: remove unnecessary check for NULL pointer 2010-04-16 23:18:43 +02:00
file.h
fileinfo.c
fileinfo.h
firefox-db2pem.sh
formdata.c remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
formdata.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
ftp.c FTP: fix bad check of Curl_timeleft() return code 2010-09-18 23:08:32 +02:00
ftp.h FTP: WILDCARDMATCH/CHUNKING/FNMATCH added 2010-05-12 23:17:51 +02:00
ftplistparser.c ftp wildcard: FTP LIST parser FIX 2010-06-24 23:45:44 +02:00
ftplistparser.h
getenv.c
getinfo.c portabilty: use proper variable type to hold sockets 2010-09-06 00:02:54 +02:00
getinfo.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
gopher.c
gopher.h
gtls.c
gtls.h
hash.c remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
hash.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
hmac.c
hostares.c
hostasyn.c
hostip4.c
hostip6.c
hostip.c fix compiler warning: variable might be clobbered by longjmp or vfork 2010-04-13 16:34:13 +02:00
hostip.h
hostsyn.c
hostthre.c
http_chunks.c
http_chunks.h
http_digest.c
http_digest.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
http_negotiate.c
http_negotiate.h
http_ntlm.c
http_ntlm.h
http.c
http.h add_buffer_send: fix compiler warning 2010-07-24 22:52:35 +02:00
if2ip.c
if2ip.h
imap.c Curl_setup_transfer: no longer returns anything 2010-04-16 23:43:04 +02:00
imap.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
inet_ntop.c
inet_ntop.h
inet_pton.c
inet_pton.h
krb4.c
krb4.h
krb5.c
ldap.c
libcurl.def
libcurl.imp added missing curl_easy_pause to export list. 2009-08-16 15:03:58 +00:00
libcurl.plist Matt Veenstra updated to 7.12.3. Starting now, we'll update the version number 2004-12-08 23:09:23 +00:00
libcurl.rc
llist.c
llist.h
Makefile.am
makefile.amiga
Makefile.b32 remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
makefile.dj
Makefile.inc Forgot gopher.h in Makefile.inc 2010-08-25 14:19:58 +02:00
Makefile.m32 It is sufficient to pipe stderr to NUL to get rid of the nasty messages. 2010-08-18 16:08:20 +02:00
Makefile.netware
Makefile.riscos Makefile: add gopher.c file to build 2010-08-25 15:56:35 +02:00
Makefile.vc6 Makefile: add gopher.c file to build 2010-08-25 15:56:35 +02:00
Makefile.vxworks
Makefile.Watcom It is sufficient to pipe stderr to NUL to get rid of the nasty messages. 2010-08-18 16:08:20 +02:00
md4.c build: remove unneeded cast to (void *) 2010-08-04 19:05:09 +02:00
md5.c build: remove unneeded cast to (void *) 2010-08-04 19:05:09 +02:00
memdebug.c
memdebug.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
mk-ca-bundle.pl
mk-ca-bundle.vbs
mprintf.c dprintf_formatf: Value stored to 'left' is never read 2010-04-16 23:50:39 +02:00
msvcproj.foot
msvcproj.head
multi.c
multiif.h
netrc.c
netrc.h
nonblock.c remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
nonblock.h
nss.c http_ntlm: add support for NSS 2010-06-30 13:12:25 +02:00
nssg.h
nwlib.c
nwos.c
openldap.c LDAP: moved variable declaration to avoid compiler warn 2010-09-18 22:50:04 +02:00
parsedate.c
parsedate.h
pingpong.c build: add typecast to avoid warning 2010-08-02 18:53:34 +02:00
pingpong.h
polarssl.c
polarssl.h
pop3.c
pop3.h
progress.c
progress.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
qssl.c
qssl.h
rawstr.c remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
rawstr.h
README.ares
README.curl_off_t
README.curlx
README.encoding
README.hostip
README.httpauth HTTP "auth done right". See lib/README.httpauth 2004-11-24 16:11:35 +00:00
README.memoryleak
README.multi_socket update the section on timeouts 2010-04-12 11:09:55 +02:00
README.pingpong
README.pipelining
rtsp.c
rtsp.h
security.c
select.c
select.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
sendf.c
sendf.h sendrecv: split the I/O handling into private handler 2010-05-07 15:05:34 +02:00
setup_once.h
setup-os400.h
setup.h Moved S_ISREG define to setup as suggested by Dan. 2010-09-02 04:37:13 +02:00
share.c
share.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
slist.c remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
slist.h
smtp.c
smtp.h
sockaddr.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
socks_gssapi.c
socks_sspi.c
socks.c fix compiler warning: enumerated type mixed with another type 2010-05-31 16:36:05 +02:00
socks.h
speedcheck.c
speedcheck.h remove the CVSish $Id$ lines 2010-03-24 11:02:54 +01:00
splay.c OOM fixes in http_negociate.c and lib/splay.c 2010-05-26 00:40:26 +02:00
splay.h
ssh.c
ssh.h build: fix libssh2_scp_send64() availability 2010-08-05 16:27:39 +02:00
sslgen.c sendrecv: split the I/O handling into private handler 2010-05-07 15:05:34 +02:00
sslgen.h sendrecv: split the I/O handling into private handler 2010-05-07 15:05:34 +02:00
ssluse.c
ssluse.h
strdup.c
strdup.h
strequal.c
strequal.h
strerror.c
strerror.h
strtok.c
strtok.h
strtoofft.c
strtoofft.h
telnet.c moved vars into conditional since seems that winsock implementation doesnt use them. 2010-05-08 05:38:42 +02:00
telnet.h
tftp.c TFTP: re-indented the source code 2010-09-26 23:38:00 +02:00
tftp.h
timeval.c
timeval.h
transfer.c
transfer.h
url.c
url.h multi interface: missed storing connection time 2010-05-07 23:49:29 +02:00
urldata.h
vc6libcurl.dsw
vc8proj.foot
vc8proj.head
version.c
warnless.c
warnless.h
wildcard.c
wildcard.h

HTTP Pipelining with libcurl
============================

Background

Since pipelining implies that one or more requests are sent to a server before
the previous response(s) have been received, we only support it for multi
interface use.

Considerations

When using the multi interface, you create one easy handle for each transfer.
Bascially any number of handles can be created, added and used with the multi
interface - simultaneously. It is an interface designed to allow many
simultaneous transfers while still using a single thread. Pipelining does not
change any of these details.

API

We've added a new option to curl_multi_setopt() called CURLMOPT_PIPELINING
that enables "attempted pipelining" and then all easy handles used on that
handle will attempt to use an existing pipeline.

Details

- A pipeline is only created if a previous connection exists to the same IP
  address that the new request is being made to use.

- Pipelines are only supported for HTTP(S) as no other currently supported
  protocol has features resemembling this, but we still name this feature
  plain 'pipelining' to possibly one day support it for other protocols as
  well.

- HTTP Pipelining is for GET and HEAD requests only.

- When a pipeline is in use, we must take precautions so that when used easy
  handles (i.e those who still wait for a response) are removed from the multi
  handle, we must deal with the outstanding response nicely.

- Explicitly asking for pipelining handle X and handle Y won't be supported.
  It isn't easy for an app to do this association. The lib should probably
  still resolve the second one properly to make sure that they actually _can_
  be considered for pipelining. Also, asking for explicit pipelining on handle
  X may be tricky when handle X get a closed connection.

- We need options to control max pipeline length, and probably how to behave
  if we reach that limit. As was discussed on the list, it can probably be
  made very complicated, so perhaps we can think of a way to pass all
  variables involved to a callback and let the application decide how to act
  in specific situations. Either way, these fancy options are only interesting
  to work on when everything is working and we have working apps to test with.