Now we support building modules as shared objects on all major platforms.
Puhhh....


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80508 13f79535-47bb-0310-9956-ffa450edef68

(BTW: compares to the linker nightmare I had under AIX today the
 OSF1 platform is really a pleasure)


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80507 13f79535-47bb-0310-9956-ffa450edef68

which currently is global without need.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80506 13f79535-47bb-0310-9956-ffa450edef68

confuses some Make variants) and slightly sort the objects file list for
cosmetic issues.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80505 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80504 13f79535-47bb-0310-9956-ffa450edef68

knows how to enable shared object support and to
avoid PR's because they had used SharedModule but
not included mod_so itself :-(


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80503 13f79535-47bb-0310-9956-ffa450edef68

- OBJS needs to be declared manually here, because Configure
  does not set it but our `$(OBJS): Makefile' dependency
  line needs OBJS to be defined. If not at least HP/UX make
  complains.

- consistently use $(LDFLAGS) and $(LIBS) even if not
  really needed for rotatelogs

- move TARGETS define in front of the targets itsself
  because some Make variants don't like defines in the
  middle of the targets section


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80502 13f79535-47bb-0310-9956-ffa450edef68

T_AT isn't needed... and it's existence made the scheme stuff a bit wrong


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80501 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80500 13f79535-47bb-0310-9956-ffa450edef68

issues to be resolved.  This is "NULL-compression" for invoke_handler()
similar to what happened in run_method() ages ago.  This helps the
server scale when more modules are added.

Submitted by:   Dmitry Khrustalev <dima@bog.msu.su>

Oh yeah and some CHANGES cleanup.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80499 13f79535-47bb-0310-9956-ffa450edef68

support for the situation where all modules of a modules/ subdir are build as
shared objects (mod_example).


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80498 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80497 13f79535-47bb-0310-9956-ffa450edef68

parse_uri().  Guys this is probably something that should be done
elsewhere.  Like maybe in directory_walk().


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80496 13f79535-47bb-0310-9956-ffa450edef68

version is over two orders of magnitude faster based on timing trials
requesting the test page from mod_test_util_uri.  It's 50% faster overall
when doing a zb /index.html with the default index.html.

I'm still resiting the urge to hardcode i386 assembly language with
a C fallback for the rest of the world ;)


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80495 13f79535-47bb-0310-9956-ffa450edef68

performance work.  Nothing but the basics.  I think it's an excellent
counter-point to our kitchen-and-sink config files that are so damn
confusing to even me... too many comments in them, too many things
to look at.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80494 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80493 13f79535-47bb-0310-9956-ffa450edef68

exclusion list because we have forgotten to bind the regex
to its boundaries.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80492 13f79535-47bb-0310-9956-ffa450edef68

to find a better solution to our mime_find_cf() cross-call
problem with mod_proxy...


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80491 13f79535-47bb-0310-9956-ffa450edef68

Pointed out by Martin.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80490 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80489 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80488 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80487 13f79535-47bb-0310-9956-ffa450edef68

where there's no need to pstrdup the arguments.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80486 13f79535-47bb-0310-9956-ffa450edef68

mod_rewrite or mod_autoindex.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80485 13f79535-47bb-0310-9956-ffa450edef68

declare the parse_uri_components/util_uri stuff part of the API.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80484 13f79535-47bb-0310-9956-ffa450edef68

NOTE!  I notice that log_error is being deleted here... and I'm too tired
to figure out what in UpdateHide is causing that.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80483 13f79535-47bb-0310-9956-ffa450edef68

the values in content_type, handler, content_encoding, ... were lowercase.
I'm not referring to mod_negotiation here -- there were other cases where
strcmp() was being used.  But in addition, mod_negotiation could be
convinced to call str_tolower() on r->content_language(s), which could
possibly modify a "read-only" string... which wastes at least one memory
page per child.

Clean all that up by declaring that the content fields in request_rec
(and related fields elsewhere) must be lowercase, and must not be modified
in place.  Naturally I chose this because it's more efficient... rather
than going around needlessly pstrdup()ing and strcasecmp()ing everywhere.

There are a few other tweaks in here I couldn't resist, along the lines
of getting rid of unneeded pstrdup()s.  Plus a new function
set_string_slot_lower().


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80482 13f79535-47bb-0310-9956-ffa450edef68

DirectoryIndex at config time.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80481 13f79535-47bb-0310-9956-ffa450edef68

looking like:

  #define SERVER_SUBVERSION \"whatever\"

When including this in a file you end up with an unterminated string
constant.  We want it to look like this instead:

  #define SERVER_SUBVERSION "whatever"

The backslashes are only needed in the actual -D argument.

This little patch makes sure that any \" from a -D is replaced with just "
when ap_config.h is generated.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80479 13f79535-47bb-0310-9956-ffa450edef68

Obtained from:
Submitted by:	Jim Jagielski
Reviewed by:
If Configuration.tmpl is more recent that Configuration, Configure
will now complain.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80476 13f79535-47bb-0310-9956-ffa450edef68

default to having them all below HTTPD_ROOT, so we may as well just do that.

PR: 1894
Obtained from:
Submitted by:   Jason Riedy <ejr@cise.ufl.edu>, Marc Slemko
Reviewed by:


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80475 13f79535-47bb-0310-9956-ffa450edef68

loadind shared objects is now supported ;-): IRIX


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80474 13f79535-47bb-0310-9956-ffa450edef68

Add "text/css" MIME type to default mime.types file, per suggestion by
Hakon Lie <howcome@w3.org>, in message <3.0.3.32.19980308204255.009adc80@localhost>

	Brian


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80472 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80469 13f79535-47bb-0310-9956-ffa450edef68

BTW: I received SIGBUS all the time in dlopen() and after wasting
     a lot of time debugging this problem I discovered that SunOS
     doesn't like it when the mod_xx.so files are stripped on
     installation. So, never strip them on SunOS.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80468 13f79535-47bb-0310-9956-ffa450edef68

Cleanup of the HTML markup code
(Thanks to Marc and Martin for hints)


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80467 13f79535-47bb-0310-9956-ffa450edef68

two reasons:

- LDFLAGS has traditionally meant "flags to use when linking an
    executable", and such flags can be different than those for making
    a shared library

- LDFLAGS has traditionally been flags supplied to the cc executable for
    linking, not flags supplied to the ld executable... cc translates
    various flags before passing them to ld.  For example,
    "gcc -Wl,M" passes -M to ld.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80465 13f79535-47bb-0310-9956-ffa450edef68

eventually.  signals are a mess.  It's bogus for mod_include to be
doing a hard_timeout() and then running subrequests and all that other
fun stuff.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80464 13f79535-47bb-0310-9956-ffa450edef68

PR:
Obtained from:
Submitted by:
Reviewed by:


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80463 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@80462 13f79535-47bb-0310-9956-ffa450edef68