ab: Add missing longest request (100%) to CSV export. Submitted by: Marcin Fabrykowski <bugzilla fabrykowski.pl> Committed by: ylavic

ab: Use only one connection to determine working destination socket address.

ab: try all destination socket addresses returned by apr_sockaddr_info_get instead of failing on first one when not available.

ab: increase request and response header size to 8192 bytes, fix potential buffer-overflow in Server: header handling.

* support/ab.c: Fix crash caused by integer overflow when printing stats with lot of requests (for example -n 500000000).

ab: support custom HTTP method with -m argument. PR 56604. Example: ab -n 1 -m PING http://127.0.0.1/ Submitted By: Roman Jurkov <winfinit gmail com> Committed By: ylavic

* support/ab.c: Revert accidental version bump in r1580928. (Although it seems likely this *should* have been bumped sometime since 2006?)

* support/ab.c (write_request): Avoid redundant write(,,0) preceding the write() of the request.

follow-up to r1542533: fix format string (unsigned vs. size_t)

Check all memory allocations and abort on failure

ab: Fix potential buffer overflows when processing the T and X command-line options. PR: 55360 Submitted by: Mike Rumph <mike.rumph oracle.com> Reviewed by: trawick

Add a new -l parameter in order not to check the length of the responses. This can be usefull with dynamic pages. PR9945, PR27888, PR42040

Check if malloc succeeded PR54344 [Bill Parker, wp02855 gmail com]

3rd patch from PR45355

The 2 first patches from PR45355 have been ported

ab: add the possibility to define a socket timeout (-s timeout) [Guido Serra <zeph fsfe org>] Also some changes In usage(): - reword -t option to be more clear with the new -s - add missing -q - add the new -s option - reword some options to better match online documentation ab.xml claimed that -s was already there, but nothing in the code looks like that. So, I guess it only landed on the online doc...

Allow forced setting of TLS1.1 and TLS1.2 protocols with the -f command-line switch, and adapt the output to more accurately report what SSL/TLS protocol was negotiated (use SSL_get_version() instead of SSL_CIPHER_get_version()). PR: 53916 Submitted by: Nicolás Pernas Maradei <nico emutex com> Reviewed/amended by: Kaspar Brand

ab: Fix read failure when targeting SSL server.

When -B is specified, search for a host address in the same family as the specified outbound interface. If impossible, fail at name resolution instead of bind().

* support/ab.c (test, start_connect): Only bind the local end of the socket (and look up an address for it) if -B is specified.

If MaxMemFree is set, set SSL_MODE_RELEASE_BUFFERS in mod_ssl. Always set SSL_MODE_RELEASE_BUFFERS in ab. PR: 51618 Submitted by: Cristian Rodríguez <crrodriguez opensuse org>, Stefan Fritsch

More cleanup: Expand tabs and some more indentation fixes No functional change

Cleanup effort in prep for GA push: Trim trailing whitespace... no func change

Drop support for the RSA BSAFE SSL-C toolkit from configure, and remove #ifdef'ed code from mod_ssl and ab where applicable. Consensus for dropping support for SSL/TLS toolkits other than OpenSSL was reached on dev@httpd in June 2010 (message with ID <20100602162310.GA11156@redhat.com> and follow-ups).

Support specifying the local address to use. PR: 48930. Submitted by: Peter Schuller <scode spotify com>

Fix various "variable 'x' set but not used" warnings.

Improvements found by cppcheck: remove some unused variables and dead assignments, reduce the scope of some variables, add some parens to improve readability

mod_ssl, ab: Support OpenSSL compiled without SSLv2 support

* support/ab.c (test): Catch more calloc failures.

Renamed optarg -> opt_arg to avoid name clashes with system-own optarg; removed leftover from IDE build times.

simplify walking through array of returned apr_pollfd_t

fix one error message, eliminate another that will never be issued (since !n occurs when status != APR_SUCCESS)

Fix usage output. PR: 45331

ab: Fix memory leak with -v2 and SSL. PR: 49383 Submitted by: Pavel Kankovsky <peak argo troja mff cuni cz>

ab was sending wrong number of requests when keepalive enabled. Thanks to Bryn Dole for the patch. PR: 48497

ab: Fix number of requests per second in HTML output being to large by a factor of 1000 PR: 48594

Add parentheses to min max macro functions to avoid future bugs.

Fix string constness to get rid of gcc compiler warnings by -Wwrite-strings.

update pollset to reflect changed connection state before closing the socket; otherwise the pollset-remove can crash (at least with select-based poll, since fd is now -1)

ab: Fix segfault in case the argument for -n is a very large number. PR: 47178 Submitted by: Philipp Hagemeister <oss phihag.de> Reviewed by: Stefan Fritsch

when a request body is sent, just display "Total body sent" in the summary instead of making up the past tense of an HTTP method (POST-ed or PUT)

* support/ab.c (output_results): Fix gcc warning on LP64 - field width passed via '*' must be an int.

refactor the representation of HTTP method and whether to send a body, fixing these insignificant bugs: * method name displayed in a debug message was wrong in some cases * specifying conflicting HTTP methods on the command-line displayed an invalid message in some cases

fixed compatibility to Openssl < 0.9.7g; sync'd compatibility to Openssl >= 0.9.9 with 2.2.x branch.

small code cleanups for POST/PUT, especially to combine common code exit if the file to POST/PUT can't be read

Fix a segfault in ab when run in verbose mode using https. PR: 46393 Submitted by: Stefan Fritsch <sf sfritsch.de>, Ryan Niebur Reviewed by: poirier

Add support for HTTP PUT to ab. Submiited by: Jeff Barnes <jbarnesweb yahoo.com>

ab: Fix broken error messages after resolver or connect() failures. The APR error code was truncated because ab used an incorrect data type.

*) ab: Fix a 100% CPU loop on platforms where a failed non-blocking connect returns EINPROGRESS and a subsequent poll() returns only POLLERR. Observed on HP-UX.

Use APR_POLLSET_NOCOPY for better performance with some pollset implementations.

repeat apr_pollset_poll() while it returns EINTR

remove unreachable code

* Remove unused variable.

Clean up more compiler emits. Add CHANGES entry, credit Kasper for the OpenSSL STACK fix

axe a fearful comment that was addressed before committing r749438

ab: Fix maintenance of the pollset to resolve EINPROGRESS errors with kqueue (BSD/OS X) and excessive CPU with event ports (Solaris). The apr_pollset API requires that a socket be removed from the pollset whenever there is a change to the requested events. The socket must also be removed when it is closed. This wasn't completely implemented. PR: 44584

The development trunk of OpenSSL has tightened up the type safety of the STACK construct and the functions that manipulate it. Make httpd trunk compile against OpenSSL HEAD as well as OpenSSL 0.9.8j. Also, get rid of some warnings.

* Change some "apr_palloc / memcpy" to apr_pstrmemdup PR: 39519 Submitted by: Christophe JAILLET <christophe.jaillet wanadoo.fr> * Remove unnecessary casts

clear out bogus whitespace

* Include <limits.h> earlier if available since INT_MAX is defined there on Windows and we need INT_MAX already for the definition of MAX_REQUESTS. PR: 45024

* Check for failure of apr_file_info_get call Reported By: BOYA SUN <boya.sun case.edu>

Improve client performance by clearing connection pool instead of destroying it. PR 40054 Submitted by: Brad Roberts <braddr puremagic.com>

Don't stop sending a request if EAGAIN is returned, which will only happen if both the write and subsequent wait are returning EAGAIN, and count posted bytes correctly when the initial write of a request is not complete. PR 10038, 38861, 39679 Submitted by: Patrick McManus <mcmanus datapower.com> (in 2003) Stefan Fleiter <stefan.fleiter web.de> (in 2006) and Davanum Srinivas (in 2006). Committed patch contains parts of all three, tweaked by Roy (2008).

Overhaul ab.c stats collection and reporting to avoid integer truncation and time divisions within the test loop, retain native time resolution until output, remove unused data, avoid structure copies, consistently round milliseconds, and generally avoid losing accuracy of calculation due to type casts. Incidentally fixes output bug on gnuplot (seconds were being output as microseconds). It would make more sense to output gnuplot stats as microseconds after the start of test, but this change returns us to consistency with pre-apr versions of ab. PR: 44878, 44931.

* Use a 64 bit unsigned int instead of a signed long to count the bytes transferred to avoid integer overflows. PR: 44346

* Do not try to read non existing response bodies of HEAD requests. PR: 34275 Submitted by: Takashi Sato <serai lans-tv.com> Reviewed by: rpluem

Kill the compiler emit - casting struct * to char * (as defined in openssl 0.9.8e). Easiest is to cast through void *.

Explain that POST data should be sent as the correct MIME type. Submitted by Vincent Bray noodlet at gmail dot com, edited and reviewed by sctemme

Correct behavior of HTTP request headers sent by ab in presence of -H command- line overrides. Previously, ab would concatenate a supplied -H User-Agent: header to the existing one, and send duplicate headers if either -H Host: or -H Accept: were specified on the command line. Now, the default headers are not sent if they are overridden using the -H command-line flag. Submitted by: Arvind Srinivasan arvind.srinivasan sun.com Reviewed by: sctemme PR: 31268, 26554

The apr_port_t type is unsigned, but ab was using a signed format code in its reports. PR 42070. Submitted by Takashi Sato serai lans-tv.com, reviewed by sctemme.

ab: Add -r option to continue after socket receive errors. Submitted by: Filip Hanik <devlist hanik.com> Reviewed by: trawick

update license header text

A keepalive response need not neccessarily have included any content-length header, handle this case properly for 304 responses. PR39789

when compiling statistics, only interate accross requests which occured.

Update the copyright year in all .c, .h and .xml files

Update the copyright notice to reflect that changes have gone in this year

Correct a dumb typo

* Add siege-like behaviour to ApacheBench; output the results, as they have accrued so far, when the user interrupts with ctrl-c. As the signal handler is non-reentrant, we don't need volatiles, and the operations all look signal-safe. * Update the base version, since handling a signal differently is an external API change of sorts. * set the LastChangedRevision svn property, as ab.c has $Rev $ embedded, so it really should be.

Give users the ability to set the window size for apache bench runs. Useful for benchmarking largefile downloads.

* modules/ssl/ssl_scache_dc.c (ssl_scache_dc_retrieve): Fix warning with OpenSSL 0.9.8a. * support/ab.c (ssl_print_cb, ssl_print_cert_info): Fix warnings with OpenSSL 0.9.8a.

No functional Change: Removing trailing whitespace. This also means that "blank" lines consisting of just spaces or tabs are now really blank lines

* support/ab.c (test): Increment count of started requests when a non-blocking connect completes; avoids case where more requests are made than intended. PR: 36966

* support/ab.c: Rewrite SSL support so that it works reliably, and enable it by default when the --enable-ssl is passed to configure. Adds support for -Z and -f arguments to tweak SSL protocol options. Submitted by: Masaoki Kobayashi <masaoki techfirm.co.jp>

* support/ab.c (open_postfile): Use apr_file_read_full, tidy up error messages, remove redundant variable.s

* support/ab.c (test): Treat POST data as a binary blob not a C string. PR: 12981 Submitted by: Matthew H. Gerlach <mgerlach@lightsurf.com>

Fix and prevent some segfaults in ab: * support/ab.c (main): Fail if given concurrency level greater than number of requests, to prevent segfaults later. (ssl_print_cert_info): Use the correct buffer size. (ssl_start_connect): SSL_get_peer_cert_chain doesn't bump refcounts, so don't free the cert chain here. (test): Use both calloc parameters (unrelated cleanup).

Update copyright year to 2005 and standardize on current copyright owner line.

Update copyright notice to new format to show what needs to be done for all of trunk.

Small fix on ab's use of sprintf(). PR: 28204 Submitted by: Erik Weide <erik.weidel mplus-technologies.de>

Use standardized names for ISO-8859-1, see http://www.iana.org/assignments/character-sets (Using the name iso8859-1 may still work, because it is aliased in apr-iconv/ccs/charset.aliases)

replaced tabs with spaces and aligned indentation within brackets

ab mixing mean and median values, also wrong time values for start and connect when doing keep alive

Using ap_max() to ensure no negative values are stored in data structure.

Add Missing parenthesis !

Limit the concurrency to MAX_CONCURRENCY. Otherwise, ab may dump core (calloc fails) when a arbitrarily huge value is used.

using build in err function, thanks Jeff for pointing that.

added check on apr_pollset_create() return value to exit ab it case it fails. On NetWare using a concurrency higher than 64 is segfaulting because of FD_SETSIZE as a value of 64.

fix name of The Apache Software Foundation

fix copyright dates according the first cehck in

apply Apache License, Version 2.0

* support/ab.c (write_request): Fix gcc warning at -O2 in USE_SSL build.

update license to 2004.

Fix ab's SSL support a little: * support/ab.c (ssl_*): Fix function prototypes througout. (test): Use correct connection structure when using SSL.

switch to APR 1.0 API (which is still in flux) because of the changes to the argument lists of apr_mmap_dup and apr_socket_create, 2.1-dev won't build with apr and apr-util's 0.9 branch anymore

distinguish between these two states: . just issued connect(), polling for writability to find out when connect() completed . already connected, waiting for data from the server so polling for readability by only polling for writability during the first state, we avoid some logic in test() which inappropriately writes multiple requests to the server btw, the most basic SSL test segfaulted before and after this commit

restore a check for write failures which was lost for non-SSL builds

ab: Work over non-loopback on Unix again. (Broken as of 2.0.47 due to dependence on an APR bug which was fixed in 2.0.47.) PR: 21495

reject overlong credentials from the commandline Acked by: Cliff Woolley

Pick up getpid() prototype.

catch out of memory condition (ID 29 of reasoning report)

Fix NULL-pointer issue in ab when parsing an incomplete or non-HTTP response. PR: 21085 Submitted by: Glenn Nielsen <glenn@apache.org>

Absorb our ssl library-foo from the mod_ssl build changes to automatically generate an SSL/https: enabled ab binary.

finished that boring job: update license to 2003. Happy New Year! ;-))

%qd is not portable; use APR_INT64_FMT_T for apr_time_t

Eliminated some fixed-size buffers that didn't have proper overflow checks

Keep track of the recent changes for the new pollset API, since they may affect performance characteristics.

Switched ab to the new abstract pollset API (this will allow ab to take advantage of future apr_pollset_t scalability improvements)

Leon Brocard writes: > Hello, > > I love httpd2 and I really like apachebench but using the gnuplot > output option (-g) gave me this: > > starttime seconds ctime dtime ttime wait > Thu Aug 29 12:29:31 200 1030620571362758 0 54 54 0 > > It's truncating the year. Now according to > http://apr.apache.org/docs/apr/group__APR__Time.html#a15 "Unlike > ANSI/ISO C ctime(), apr_ctime() does not include a \n", so the code > and comment in ab.c are wrong, hence my attached patch to remove them. Submitted by: Leon Brocard <acme@astray.com>

Revert to CRYPTO_malloc_init for non-rsa [gotta read the diffs more closely the first time.]

Support rsaref compilation

fix a spelling problem

And fix the version number

Fix the long-standing bug that "ab -t10" would loop for 10000 seconds instead of 10 as documented. Also fix an off-by-one-second error, which did not matter much (in comparison to the 1000x error ;-)

http occurances to update for apr_socket_opt_get/set

We where mixing #if USE_SSL en #ifdef USE_SSL. Since there where many more #ifdef's than #if's - Converted the minority to the majority. Note however that the APR_* defines are usually set to '=1'. So this is not quite consistent. (I.e. CFLAGS=-DUSE_SSL or CFLAGS=-DUSE_SSL=1). Dw.

Bring Apache in line with APR's new poll implementation. This isn't optimal, because it still uses the compatibility functions to make it work, but it does work. The next step will be to take the memory for the pollsets under Apache's control.

If you want floating math, use float literals.

fix for compilers that don't support long long datatype

use the time conversion macros rather than dividing by APR_USEC_PER_SEC

A few more unused/unnecessary variables and surpress an uninitialized value warning.

Enable SSL negotation in ab.c. It is that, or remove all the entirely BOGUS ssl config options from ab.c. If it works, nifty, if not, we now have directives that have some effect, rather than no effect, for SSL. Submitted by: Madhu Mathihalli <madhusudan_mathihalli@hp.com>

Well, know that I'm aware that the ssl stuff was entirely premature, it's still worth committing the correct code to handle the port number.

Locks are good, provide locks. ab -s -c >1 is [was?] dying on Win32 while -s -c 1 works just fine. Still investigating.

There's really no reason for a -dev here.

Fix some minor formatting issues with ab. Part of this is in reference to PR 8544, the rest I noticed while testing the PR fix. [Paul J. Reder] Submitted By: Carlo Marcelo Arenas Belon Reviewed By: Paul J. Reder

Remove APR_WANT_SIGNAL from apr_want.h because code must include apr_signal.h in order to get consistent definitions.

During the 1.3->2.0 migragrion; ab its #defined VERSION own string was replaced by that of the base server distribution it sits in. Propably by accident. This is propably not a good idea - as ideally one would like to be able to compare ab runs as much as possible - even across releases of apache 2.0 - assuming ab or the dependent APR has not changed (note to self: we do not track APR in our version structure). This commit decouples the version strings for now. Though the actual value may be nicely confusing.

Document the recent AB changes since they have the potential to change the meaning of test results.

Major improvement in concurrent processing for AB: - Enable non-blocking connects. - Prevent quasi-blocking mode apr_recv (which would prevent AB from multiplexing over the entire descriptor set). - Catch other fatal apr_recv() errors. [This patch is slightly different than the one posted to the dev list, but regardless thanks to the many people who reviewed this.]

fixed a typo

Correct const'ness of argv in all support apps, and use the new apr_app_initialize over apr_initialize for win32, and other platforms that may wish to tweak 'apr-ized' application support (e.g. Netware?)

Made AB's use of the Host: header rfc2616 compliant as reported by Taisuke Yamada <tai@iij.ad.jp> PR: Obtained from: Mailing list Submitted by: Taisuke Yamada <tai@iij.ad.jp> Reviewed by: Dirk-Willem van Gulik

Removed email address on request - as obviscation was slammed down on the list - and Google is your friend in any case

What is "zop"? :)

Carriage returns??

Update our copyright for this year.

simplify the way we find the family of an apr_sockaddr_t

Fixed segfault in "-i" mode of ab

"totalcon / requests" is no longer double either, so %5e doesn't fly

I don't believe it. This actually fixes the rest of the type mismatches. Of course, I could have hosed everything, a second pair of eyes would be appreciated.

clean up some warnings on AIX since apr_size_t isn't always int, we should use APR_SIZE_T_FMT instead of %d

Some type correctness fixes for LARGEFILES platforms.

simplify some code in output_results() just a bit to avoid a bug in gcc 2.7.2.3 on FreeBSD 3.4; didn't hit the problem with -O2; did hit the problem with no -O

Now that it's apr'ized, ab is most definately in a 2.0 revision cycle, and definately needs to track the Apache revision (especially because of apr and apr-util!)

If we've spent the effort of getting some connections, when we die with an error at least inform the user how many we completed before dying. Tidy up some formatting...

A smart answer to users attempting the -s arg to ab.

fix the type of parms to apr_send() and apr_recv()

Truncate the working set in ab by introducing a connection pool.

Hmmm... editing with tabs != 8, eeuh. Just cleaning and fixing casts.

Fix some casting

Changes to back out inherit flag from apr_os_sock_make() and apr_socket_create()

Added an inherit flag to apr_socket_create and other socket creation functions. This allows APR programs to specify that a socket should be passed to any child processes that are created. The inherit flag is only meaningful if programs use apr_process_create(). This also adds a couple of macros that allow APR types to set and unset the inherit flag. This also fixes Apache to use the new API.

Fix spelling mistake.

The patch fixes 2 things. * apr_terminate now has the right declaration so it can be used directly in atexit * Win2K gets a divde by zero error when heartbeatres is zero (reproduce via support\Debug\ab.exe -k -c 6 -n 100 localhost/ Submitted by: Ian Holsman <ianh@cnet.com> Reviewed by: Jeff Trawick

Make ab work again by changing its native types to apr types and formats. Submitted by: Justin Erenkrantz <jerenkrantz@ebuilt.com> Reviewed by: Roy Fielding

Uhh, we really need the prototypes for standard string functions.

Make live simpler - as we are using strchr already

Doing last minute alphabettically sorting things to look nice is not always a good idea.

make sure we have the prototype for index() (fixes a Solaris warning)

consistent types; avoid warning on AIX for mismatch between printf format and arg

"void *res = signal(blah)" isn't cool 'cause void * is not compatible with a function pointer; but we don't need to save the old signal pointer anyway so drop the variable

keep ab from dividing by zero when heartbeatres is zero

Cleanup all the warnings introduced in AB yesterday.

Commit early and often. This should merge all changes into ab from 1.3 into 2.0. SSL work not completed; I need to move some of the autoconf detection of mod_wtls to a more central place first. Any suggestions appreciated :-) PR: Obtained from: Submitted by: Reviewed by:

Either MacOS-X is much more sensitive for border conditions - or I am just being unlucky. This allows for apr_send() to do consequitive partial write's; by keeping a buffer and TO within ab itself. Ab was already doing that for read anyway. PR: Obtained from: Submitted by: Reviewed by:

First stab at encapsulating the apr_send() to make it a little more pushy. Ran into this whilst' porting the -1.3 changes forward. PR: Obtained from: Submitted by: Reviewed by:

Prevent cobbering of lenght for subsequent sends. Still not quite right; we ough to call apr_send again if len != requested_len - but then our timeout could be len * timeout if we happen to only stuff one byte down the pipe each time. PR: Obtained from: Submitted by: Reviewed by:

Use "Basic" authentication instead of "basic" in ab, as the spec says we should. PR: 7294 Submitted by: Andre Breiler <andre.breiler@rd.bbc.co.uk>

Update copyright to 2001

renaming various functions for consistency sake see: http://apr.apache.org/~dougm/apr_rename.pl PR: Obtained from: Submitted by: Reviewed by:

close the post file after reading it

The changes required for the APR_FINFO_wanted argument to apr_stat/lstat/getfileinfo. These are -NOT- optimal, they are simply the required changes to get the server working. The size of the patch is a warning about how we need to really look at what we are trying to accomplish with all of these stat/lstat calls.

The big change. This is part 3 of the apr-util symbols rename, please see the first commit of srclib/apr-util/include (cvs apr-util/include) for the quick glance at symbols changed.

When is a void fn(void) not a void fn(void)? It's a win32'ism, since the APR_DECLARE (excluding _NONSTD) uses the pascal convention. Really irrelevant in a function that never needs to clean up (atexit), but the compiler complains. Possible workarounds, but that would break non-c lanaguge code from linking into apr.

fix a warning on platforms where apr_size_t isn't int

Fix an endless loop in ab which occurred when ab was posting and the server dropped the connection unexpectedly.

open_postfile(): fix some return codes along the lines of OtherBill's comment fix some bad conditional logic for when to check if we read the wrong number of bytes

get rid of a non-portable use of errno don't return errno as an error flag after an APR function fails; return something guaranteed to be non-zero on all platforms

Cleanups so that the support programs build cleanly. The Win32 values need to be sanity checked. Submitted by: Cliff Woolley <cliffwoolley@yahoo.com> Reviewed by: Ryan Bloom

All the supports now build on APR (+APRUTIL), no reliance on httpd.

This program should run entirely without ap_config support from Apache, given that it is apr'ized

ab: use apr_parse_addr_port() in parse_url(). httpd.exp: Add the new APR symbol apr_parse_addr_port.

Fix some leftover dirtyness from the conversion. All is building over in Win32 land.

Make ApacheBench compile cleanly again

use APR_SIZE_T_FMT to avoid a warning on AIX

Use "const char * const *" for process->argv (which is the correct const-ness since we sometimes put "some string" in there, and also the CRT's argv). propagate this change within http_main and mpm/winnt/ (also correct some other const type usage within the MPM). fix ab's call to parse_url() which removed a const to actually manipulate an arg from the CRT's argv (indirectly via opt->arg). no idea how this has avoided segfaulting.

Get ab to compile cleanly after the apr_initopt() change.

*) Compensate for recent changes in the APR headers. Specifically, some files need to specifically include stdio.h, or a particular apr_*.h header. *) Adjust callers of apr_create_process() to deal with the extra "const" *) Add "const" to args of ap_os_create_privileged_process()

Don't use APR_INET6 unless APR_HAVE_INET6 is defined.

Update ab to accept URLs with IPv6 literal address strings (in the format described in RFC 2732), and to build Host header fields in the same format. This allows IPv6 literal address strings to be used with ab. This support has been tested against Apache 1.3 with the KAME patch, but Apache 2.0 does not yet work with this format of the Host header field.

Call apr_create_socket() instead of apr_create_tcp_socket() (deprecated).

APR: Change apr_connect() to take apr_sockaddr_t instead of hostname. Add generic apr_create_socket(). Add apr_getaddrinfo() for doing hostname resolution/address string parsing and building apr_sockaddr_t. Submitted by: David Reid Reviewed by: Jeff Trawick

Fix a lot of the fallback from the apr_ssize_t to apr_size_t change Submitted by: Victor J. Orlikowski <v.j.orlikowski@gte.net>

Get ab to build again by changing apr_set_remote_port() to apr_set_port().

Get rid of some warnings on Solaris 8 by using APR's ctype macros: logresolve.c: In function `main': logresolve.c:327: warning: subscript has type `char' ab.c: In function `main': ab.c:1074: warning: subscript has type `char' ab.c:1087: warning: subscript has type `char'

Some fixes and a general tidy up of ab. With these changes ab can now support concurrency levels > 1 against the loopback on BONE and will actually use all the connections it starts.

The lots of little ones... APR_IS_STATUS_condition(rv) conditional macros replacing the majority of fallible rv == APR_condition tests. But there are lots more to fix, these are the obvious ones that already did proper canonical error conversion.

Fix the types of the variables passed as the len parm to bucket->read() (in various places). In ab, change the declaration of argv so that it can be passed into apr_initopt() without warning. Submitted by: Victor J. Orlikowski <v.j.orlikowski@gte.net> Reviewed and modified slightly by: Jeff Trawick

Tweak apr_getopt() and its use in ab to avoid compiler warnings. (Note: apr_initopt() and/or its callers still need tweaking.) Submitted by: Bill Rowe Reviewed by: Jeff Trawick

Fix ApacheBench for the apr_initopt/apr_getopt change. This illustrates abusing the apr_getopt_t for direct access at the working structure, just as we did with the global vars. By my global search, there should be no more apr_opt*** symbols out there. PR: Obtained from: Submitted by: Reviewed by:

A few more (last?) ap_xlate->apr_xlate changes covering stuff not completely handled in the big apr rename last week.

Fix some problems with the apr conversion so that APACHE_XLATE builds work again.

Remaining cleanup of ap_ -> apr_ and AP_ -> APR_ transformation... 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:

prefix libapr functions and types with apr_

Don't use perror() to report the failure of an APR function. Introduce apr_err() for that purpose.

ab would start up more connections than needed, then quit when the desired number were finished. Also fixed a logic error involving ab keepalives. Submitted by: Victor J. Orlikowski Reviewed by: Roy Fielding

Protect system header files with the appropriate macros.

PR: Obtained from: Submitted by: Reviewed by: One reason ap_config.h can be a problem. Change VERSION->AB_VERSION to avoid ap_config.h #undef.

Remove a warning about ap_base64encode not being defined anywhere.

ab: Fix a command-line processing bug; track bad headers in err_response; support reading headers up to 2K. (In the original patch it was 8K instead of 2K; it isn't clear to me that 8K is really needed or appropriate.) Submitted by: Ask Bjoern Hansen <ask@valueclick.com> Reviewed by: Jeff Trawick

APR-ize the CHARSET_EBCDIC support in ap_base64encode() and ap_base64decode(). The app (e.g., Apache, ab) must call a function to set up translation.

Fix some bugs (mostly lost 1.3 code) in ab's command-line processing.

add missing ';' in non-ASCII path

Win32: install ab.exe fix dependency in Apache.dsw (ab is dependent on aprlib and ap) fix ab and htdigest projects so that they find all header files define ap_signal() in apr.hw - this is o.k. for casual use as with cmd-line programs use API_VAR_EXPORT as appropriate in getopt.c general: ab doesn't need to declare ap_optarg/ap_optind; fix a warning htdigest needs to call ap_initialize() to avoid segfault at startup

Build ab on Win32. Fix some minor ab APR-ization issues.

Repair ap_poll() on UNIX systems where APR's poll flags have different values than the system's poll flags (i.e., call get_revents() to translate flags before returning to the caller; otherwise, they are quite meaningless :) ). Fix ab's usage of the new ap_poll() and repair some APR-ization breakage (e.g., not setting port prior to ap_connect()). Uninteresting changes: Change toascii to to_ascii in ab.c to avoid a conflict with MSVC++'s toascii() (I enabled the NOT_ASCII logic in ab on Win32 recently to verify that ap_xlate_X() returned the correct error codes and got warnings on toascii. get_revent() was reformatted unintentionally. Sorry.

I have merged the two ab files together, the one from src/support and src/lib/apr/test. This means that the ab program in the support directory is now portable using APR. This has only gone through the barest of testing, and needs to be tested much better.

Update to Apache Software License version 1.1

Fix all the License issues. Including: s/Apache Group/Apache Software Foundation/ s/1999/2000/ s/Sascha's license/ASF license

Add CVS Revision number to ab.c to allow for detecting changes which were not marked by a version number update. Additionally, display base repository name (apache-2.0) Also, merge revision 1.34 of the 1.3 branch: Made sure ApacheBench (ab) performs no more requests than specified on command line (option -n). Submitted by: Jim Cox <jc@superlink.net> Reviewed by: Ralf S. Engelschall PR: 4839

These changes allow the support programs to compile on BeOS. The biggest change is in ab.c where the functions that BeOS needs that aren't the standard format are hidden behind a new ab_* define. Tested on FreeBSD and BeOS.

Re-sync to most up-to-date ab.c version 1.3b

Changed pools to contexts. Tested with prefork and pthread mpm's. I'll check this out tomorrow and make sure everything was checked in correctly.

Apache 1.3.9 baseline for the Apache 2.0 repository. Obtained from: Apache 1.3.9 (minus unused files), tag APACHE_1_3_9 Submitted by: Apache Group