For POST and PUT, libcurl adds by default a "Expect: 100-continue"
header and wait for the server to reply with a 100 response.
However, since the server never does that (LL comment in the
code "I'm not sure what it means") it's up to libcurl to continue
anyway after a while and that part is apparently bugged, causing
people not to be able to login sometimes. But suppressing the
header, libcurl doesn't wait for it and we worked around this bug.
The commit introduces a much improved CurlEasyRequest::setPost
method that enforces the above for POST, but that also no longer
uses CURLOPT_COPYPOSTFIELDS (but CURLOPT_POSTFIELDS), no longer
making a copy of the body of what we are going to send to the server.
Instead it uses a new object, derived from AIPostField, to keep track
of this data and to dispose of it once the transaction is complete
(and no sooner).
Copied from llcorehttp, we now also always set a "Connection:
keep-alive" and "Keep-alive: 300" header for POST.
This was already done for texture downloads (HttpOpRequest),
but now we do it always :p. Might need to be changed in the
future, but currently those headers are ignored by the server
anyway.
* Do not include aithreadid.h from debug.h, because the latter is
included everywhere (from linden_common.h) and aithreadid.h is
heavy (includes among others windows.h).
* On windows, thread local members cannot be exported.
Also adds a more robust interface for setopt that does
type checking based on the options used. This fixes one
bug where a F32 was passed and interpreted as long.
In many cases a U32 or S32 was passed as long, which
would fail (only) on a 64bit non-windows big endian machine.
Fixes an exit crash when the user clicks on Login and immediately
on Quit, because termination of the LLAuth curl request statemachine
access the CurlRequestTimeOut debug setting.
Apart from just really cleaning things up and moving
everything into one class regarding thread IDs (ie,
is_main_thread(), comparing ID's etc), this also
fixes an obscure bug where LL was casting thread ID's
to U32 and then compared those to find out if it
the same thread. It's theoretically possible that
such fails on a 64bit OS.
By generalizing the interface, I adopted the use
of a thread-local cache for the current thread ID
as used by LLMutex et al, so now all code benefits
from that. The idea was even extended to now also
be used for is_main_thread() tests and even resetting
a thread ID to the ID of the current thread.
renamed cwdebug/debug_libcurl.cc -> llmessage/debug_libcurl.cpp
and cwdebug/debug_libcurl.h -> llmessage/debug_libcurl.h,
because debug_libcurl.cpp does curl calls that do ares and
openssl calls, so we need to link with those libraries.
llmessage is already linking with those libraries, and contains
the main entry point aicurl.h, so it's a suitable place to put
this.
Bug fix: must always include llpreprocessor.h before including
curl/curl.h.
Bug fix: Added #include "debug_libcurl.h" to hipporestrequest.cpp
and llurlsimstring.cpp which I missed before because they
included "curl/curl.h" instead of <curl/curl.h>. Same in
llwaterparammanager.cpp, but removed include there because it
isn't needed.
Now test DEBUG_CURLIO before including debug_curlio, that
seems better, because otherwise it would make more sense to
replace all #include <curl/curl.h> with #include "mycurl.h"
and then do it there-- but I didn't want to do that.
Bug fix: we undef-ed CURLOPT_DNS_USE_GLOBAL_CACHE, while really
that is an enum, not a macro.
Fixed DEBUG_WINDOWS_CODE_ON_LINUX again by adding a hack for
ioctlsocket(), not instantiating dumb_socketpair unless
DEBUG_WINDOWS_CODE_ON_LINUX is defined and removing again ^M's
introduced with the new windows non-blocking code.
Also changed the type of flags passed to fcntl to int (was long).
curl_easy_setopt expects a long int. Before this patch,
uninitialized memory was read, leading to extreme long
time out, instead of the intended disabled time out (so,
in practise this patch has little effect).
This bug was discovered with the previous commit.
Basically, cmake doesn't support linking static libs into a shared lib.
The correct way is to just specify source files in subdirectories
directly as source files of the shared library. This patch changes that.
Also, after this commit, when DEBUG_CURLIO is defined, every call to
libcurl is printed to llinfos (or to dc::curl when using libcwd).
