| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60556 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The reason is really, really simple. If we ever choose to mix clib and
dos error codes, they criss-cross and don't line up, but they share the
same number space. As I wrote the new APR_IS_ERROR macros, I realized
we were about to shoot ourselves in the foot.
These changes nearly entirely affect Win32 only. The next big patch will
affect all of the rv == APR_ENOENT type problems throughout the system.
PR:
Obtained from:
Submitted by:
Reviewed by:
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60553 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
| |
problem on non-glibc linux machines. This patch solves the problem.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60540 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
systems have unsigned sockaddr len parameters; others have signed.
When passing the length by address (as to accept()) a warning is
generated if the sign is not correct.
This patch assumes that, if no native socklen_t is provided,
apr_socklen_t is always signed; that can be easily tweaked for
certain platforms later.
Inside APR, apr_socklen_t is only used currently in the Unix
implementation.
Submitted by: Victor J. Orlikowski <v.j.orlikowski@gte.net>
Reviewed by: Jeff Trawick
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60538 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60534 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60514 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60513 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
| |
Define TCP_NODELAY on OS/390. libc doesn't define it in a header file
we'd want to include.
TCP_NODELAY was busted all along on OS/390; we just didn't realize it.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60509 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Note that several areas have not been tested as they apply to MPMs or APR
code that I can't test.
PR:
Obtained from:
Submitted by:
Reviewed by:
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60508 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
setsockopt(SO_SNDRCVTIMEO). Applications will need to segment apr_sends
or set timeouts appropriately to accomodate slow clients.
Apache should be fine as it does not typically attempt to apr_send more
than 8192 bytes at once with default timeouts of 15 seconds.
apr_sendfile() is another matter... On Windows, TransmitFile will block
(or not trigger a completion event) until the entire file content is sent.
It is easy for a slow client to request a file than cannot be served before
the timeout expires. To accomodate these cases, I introduced the
MAX_SEGMENT_SIZE #define to define the maximum amount of data TransmitFile
will be asked to send at once. The default setting of 65536 should accomodate
most clients dialed in at 28K (or more) if i/o timeouts are configured
to be 30 seconds or more.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60506 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60502 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60486 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
see src/lib/apr/apr_compat.h for most details.
Also a few minor nits to get Win32 to build.
PR:
Obtained from:
Submitted by:
Reviewed by:
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60481 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
| |
apr_sendfile.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60474 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60473 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60470 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
| |
return the right information on the local socket address. This
change uses the Unix logic for delaying/avoiding getsockname() to
ensure that it is done when needed.
Submitted by: Gregory Nicholls <gnicholls@level8.com>
Reviewed by: Jeff Trawick
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60468 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
Submitted by: Gregory Nicholls <gnicholls@level8.com>
Reviewed by: Jeff Trawick
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60467 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
pass one, so don't fall over if hostname is NULL. The caller
should have called ap_set_remote_ipaddr() if hostname is NULL.
Submitted by: Gregory Nicholls <gnicholls@level8.com>
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60451 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
apr_string.h for apr_lib.h, fix some bad linkage declarations,
and some more comments on threadproc/win32/proc.c
PR:
Obtained from:
Submitted by:
Reviewed by:
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60444 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60426 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
create a new header for those functions. This is the first step to
removing the apr/lib directory completely, and moving those files/functions
to descriptive directories.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60398 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
set hdtr to the address of one that specifies no headers and trailers.
This prevents accessing low core while keeping the following code
simpler (no need to check hdtr != NULL later).
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60390 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
version from Dave Hill. It supports non-blocking sockets.
Submitted by: Dave Hill <ddhill@zk3.dec.com>
Reviewed by: Jeff Trawick
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60387 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The setsockopt() for TCP_CORK gets the failure.
Submitted by: Greg Ames
Reviewed by: Jeff Trawick, who tweaked Greg's latest patch
to avoid a compile warning and to work with
non-blocking sockets. We can't leave the
socket corked when we bail out when a write
would block. Not even shutdown will uncork it
(Linux kernel 2.2.5).
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60385 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
| |
is a start. Also, add an error message for machines where sendfile is
detected, but nobody has written ap_sendfile.
Submitted by: Dave Hill <ddhill@laurel.zk3.dec.com>
Reviewed by: Ryan Bloom
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60380 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
| |
error." this removes some of the error checking that was occurring in the
"user data" functions. also saw some in the "get/set OS type" functions and
nuked those; there are still checks for pool==NULL, though, since that would
end up making us malloc() rather than segfault'ing.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60318 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60317 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60315 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
Linux flavor: get it working with non-blocking sockets which
don't have a timeout
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60311 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
| |
all Unix flavors: fix the check for whether or not we have a timeout
FreeBSD flavor: make it work correct with a non-blocking socket that
doesn't have a timeout
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60310 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
| |
argument is properly initialized for all platforms.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60305 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60291 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
| |
PR:
Obtained from:
Submitted by:
Reviewed by:
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60289 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
| |
Defer/avoid getsockname() whenever possible.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60280 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
OS/390: just use Victor's AIX flavor... it works for OS/390 as well
FreeBSD: account for header bytes when we tell the kernel how much
to send; prior to this fix, we wouldn't send as much of the
file as desired if headers were provided
Note: APR_SO_TIMEOUT issues discussed on new-httpd today will be
addressed later. (one step at a time!)
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60269 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
with non-zero argument makes the socket non-blocking. BeOS and
OS/2 already worked this way.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60266 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
Submitted by: "Victor J. Orlikowski" <vjo@raleigh.ibm.com>
Reviewed by: Greg Stein
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60263 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
| |
APR can modify the ap_socket_t as it sees fit. It may choose to
modify the ap_socket_t on functions which only read from the ap_socket_t
conceptually.
Note: http_connection::ap_new_apr_connection() passes its ap_socket_t
arg to one of the changed functions, so const was removed there also.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60260 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60254 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
storage for the headers. This will not cause us too much of a problem in
practice for HTTP I think since keep alive connections do not tend to
stay up long.
OTOH, some benchmarking clients may keep a connection up for the duration
of a test run. In this case, need to configure Apache to kill the connection
after, say, 100 connections on the outside. Left the malloc code in but #if'ed
out in case we want to change it back...
Submitted by: Victor Orlikowski (with some munging by Bill)
Reviewed by: Bill Stoddard
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60249 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
patch changes it back on all platforms. This makes APR_HAS_SENDFILE look
like the rest of the APR function feature macros. Sorry about the
inadvertant change.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60244 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
restore the original timeout on exit.
This is very similar to the Win32 implementation Bill Stoddard a week ago.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60242 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
should also use that macro internally. This keeps us from checking for
multiple macros as we were doing in the SENDFILE case. It also means that
APR is definately building the same way that external programs expect it to.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60239 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
integers and we wanted 64 bit integers
Submitted by: Victor J. Orlikowski <vjo@raleigh.ibm.com>
Reviewed by: Ryan Bloom
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60238 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
Submitted by: Victor Orlikowski
Reviewed by: Bill Stoddard
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60229 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60214 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60205 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
and iol_getopt().
Do we need to move iol into APR?
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60204 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
| |
now. Cleaned up some Windows specific bugs with timing out connections as well.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr/trunk@60202 13f79535-47bb-0310-9956-ffa450edef68
|