git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80735 13f79535-47bb-0310-9956-ffa450edef68

Submitted by:	Lars Eilebrecht


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80734 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80733 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80732 13f79535-47bb-0310-9956-ffa450edef68

with the new Apache Autoconf-style Interface (APACI) while adjusting some
documentation and CVS files for it.

APACI was already tested by me under the following platforms:

  FreeBSD   2.1.5-RELEASE   [en1]         +DSO
  FreeBSD   2.2.2-RELEASE   [taz]         +DSO
  FreeBSD   2.2.5-RELEASE   [en4]         +DSO
  Solaris   2.5.1           [sunti5]      +DSO
  Solaris   2.6             [en5]         +DSO
  SunOS     4.1.3           [sunbi1]      +DSO
  Linux     2.0.33 / Deb131 [gw1]         +DSO
  Linux     2.0.33 / RH4.1  [rcs41]       +DSO
  DEC OSF/1 4.0             [rcs12]       +DSO
  IRIX      6.2             [sgihalle14]  -
  HPUX      10.20           [rcs31]       -
  AIX       4.2             [rcs7]        -

with the test procedure command

  $ ./configure --prefix=/tmp/apache --enable-module=most [--enable-shared=max]
  $ make
  $ make install
  $ vi /tmp/apache/etc/httpd.conf
  [edit Port 80->8000]
  $ /tmp/apache/sbin/apachectl start
  $ lynx -source http://localhost:8000/
  $ /tmp/apache/sbin/apachectl stop

We should test more in the next days before 1.3b6 will go out.  Let us hope
that our users now really honor my hacking and the groups decision effort for
APACI. Puuhhhhhhh....  ;-)


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80731 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80730 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80729 13f79535-47bb-0310-9956-ffa450edef68

the callers hand apapi_vformatter pointers into their own private buffers,
and are called back for flushing when appropriate.  This is more efficient
and seems somewhat more clean.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80728 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80727 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80726 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80725 13f79535-47bb-0310-9956-ffa450edef68

remove HAVE_CVT because we need threadsafe functions


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80724 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80723 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80722 13f79535-47bb-0310-9956-ffa450edef68

was defined in a parent directory.

Submitted by:	Lars Eilebrecht


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80721 13f79535-47bb-0310-9956-ffa450edef68

hack, so this define is no longer useable thus there is no need for
documenting it any longer.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80719 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80718 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80717 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80716 13f79535-47bb-0310-9956-ffa450edef68

since october.

- apapi_vformatter() is generic printf-style routine with arbitrary
  output
- replaces ap_snprintf(), eliminate the possibility of HAVE_SNPRINTF
- replaces bprintf() and bvprintf()
- add new psprintf()/pvsprintf() which allocate the string from the
  pool -- no static limitations
- use psprintf() in a bunch of places


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80715 13f79535-47bb-0310-9956-ffa450edef68

is necessary on at least Solaris where the /etc/rc?.d scripts
are run with these signals ignored, and "SIG_IGN" settings are
maintained across exec().

PR:             2009
Submitted by:   Rein Tollevik <reint@sys.sol.no>


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80714 13f79535-47bb-0310-9956-ffa450edef68

Submitted by: Rein Tollevik <reint@sys.sol.no>
Reviewed by: Ralf S. Engelschall
PR: 2010


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80713 13f79535-47bb-0310-9956-ffa450edef68

it affects our general c-t-r process and how some of us knew that
this was bound to happen


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80712 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80711 13f79535-47bb-0310-9956-ffa450edef68

fit the requirements of mod_proxy again.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80710 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80709 13f79535-47bb-0310-9956-ffa450edef68

when using shared objects. Because these lines have to be in front of any
directives which are contained in the module itself it is recommended to load
those modules at the top of httpd.conf.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80708 13f79535-47bb-0310-9956-ffa450edef68

- avoid gcc complains by define main as at the other places in the two new
  generation programs
- make uri_delim as all-in-one as test_char (#include in the middle of a
  struct definition is a little bit ugly)
- make the generated files readable even not editable
- remove double-defined dependencies in Makefile.tmpl (because they are
  already part of the "make depend" generated list)

PS: There seems to be another problem here: The files are
    generated in main/ and included with #include "...".
    Although we only have -I../include and -I../os/unix but
    not -I. ! Is it really portable for all compilers to
    assume the CWD is searched for #include "..." even
    if no -I. is used? I personally would change it so
    that the files are generated in ../include/. Comments?


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80707 13f79535-47bb-0310-9956-ffa450edef68

us as well because they can avert "bug" reports.  Fix a few more... there's
dozens more that could do with improvements.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80706 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80705 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80704 13f79535-47bb-0310-9956-ffa450edef68

Submitted by:	jim <jim@nlenx.com>


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80703 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80701 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80700 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80699 13f79535-47bb-0310-9956-ffa450edef68

We need to mark various bitfields as signed to follow the spec.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80698 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80697 13f79535-47bb-0310-9956-ffa450edef68

naming each beta release to come later, but, in short,
I think it just will cause major confusion for the outside world


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80696 13f79535-47bb-0310-9956-ffa450edef68

fixes I've been thinking about are somewhere deep in my mind I can't
find at the moment.  I'm ready to "throw in the towel" so that unix
1.3.0 can get out there, and get into reviewer's hot little hands.  I
know of a few groups that want to benchmark it, but really would prefer
to benchmark something released.  I really don't want to see us with
win32 tied to unix release schedule forever... it's pointless.

It's all just names anyhow.  Whether we call it a beta or a current
release it's just a name.  It's time we at least had a "release candidate"
to use the 1.2 terminology... I'm with Roy, I want to drop the x.ybz
designations.  I want x.y.z, a single numbering scheme.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80695 13f79535-47bb-0310-9956-ffa450edef68

for graceful restarts.  Given that the only usage of SIGURG is now
protected by a #ifdef we don't really need the definition.  ... similarly
for BeOS, why define SIGBUS SIGSEGV?


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80694 13f79535-47bb-0310-9956-ffa450edef68