that, but it was wrong anyway. Noticed by Ovies Brabson

Submitted by:   Ovies Brabson <oviesb@us.ibm.com>
Reviewed by:    Martin Kraemer


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84260 13f79535-47bb-0310-9956-ffa450edef68

The *ebcdic*_strictly() kludge is now no longer required.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84259 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84258 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84257 13f79535-47bb-0310-9956-ffa450edef68

Submitted by: Honza Jirousek <honza@ecn.cz>

The author writes:
 couple notes:

 1) My Netscape Communicator 4.6 for Linux lists Czech language under code
 Czech(cs). I'm not sure what standards rule this, but it's quite likely wrong.
 Since the split of Czechoslovakia into two independent countries about 7 years
 ago, cz is an ISO code for Czech and sk for Slovak. Top level DNS domain cs
 (which caused so much trouble to mail users from computer science departments
 in US universities, who used to cut the domain part of local e-mail addreses
 short (e.g. abc@cs.berkeley.edu-&gt;abc@cs) and were surprised to get messages
 bounced by some Czechoslovakian mail servers) and other uses of the code cs
 should be deprecated by now. I assume, however, that this is of no
 significance to Apache, which handles thye country codes in httpd.conf only.

 2) Note that while Slovak language uses the same character set as Czech and
 while it's a fairly close to Czech, it is not identical and this translation
 will not pass for a Slovak version of the page. Slovak users may likely use
 the language negotiation feature, though, and set the order &quot;sk,cz,en&quot; in
 their browsers.

 3) There are two encoding for non-asci Czech characters. Supposed Internet
 standard is iso-8559-2, but most windows browsers use windows-1250 (and thus
 many Czech web pages are also created in this character set). Most browsers
 (on all platforms) can handle both encodings, but have to be told which one to
 use, as no good &quot;default&quot; can be assumed. One way is to configure Apache to
 send proper character set in HTTP headers (this has to be set locally based on
 website owner's preference), another is to include the character set in a META
 tag. I think the latter would work better for default welcome page and so I
 did in the page I'm sending.

 Honza Jirousek


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84256 13f79535-47bb-0310-9956-ffa450edef68

Next step is replacement of all xxx_strictly stuff by xxx
(and renaming the os_toascii_strictly[] table to os_toascii[])


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84255 13f79535-47bb-0310-9956-ffa450edef68

the code (somewhat) more readable, and improves the portability
to character sets other than ASCII (e.g., EBCDIC).
This patch results in no functional change whatsoever on ASCII machines,
but allows EBCDIC platforms to live without the ebcdic2ascii_strictly()
kludge.

Submitted by:   Paul Gilmartin <pg@sweng.stortek.com>
Reviewed by:    Jim Jgielski, Martin Kraemer


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84254 13f79535-47bb-0310-9956-ffa450edef68

- better comparing of request-uri with uri parameter in Authorization
  header
- added a check for a MUST condition in the spec
- fixed SEGV


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84253 13f79535-47bb-0310-9956-ffa450edef68

Submitted by:	Joe Moenich <moenich@us.ibm.com>


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84250 13f79535-47bb-0310-9956-ffa450edef68

	Add ability to handle arbitrary methods to Script directive.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84248 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84246 13f79535-47bb-0310-9956-ffa450edef68

	Add Youichirou Koga's AddCharset patch to what's available.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84245 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84243 13f79535-47bb-0310-9956-ffa450edef68

This change results in fewer OS dependencies than the old (hardcoded)
"#if defined() &&..." list

Submitted by:   Ovies Brabson <oviesb@us.ibm.com>
Reviewed by:    Martin Kraemer


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84242 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84241 13f79535-47bb-0310-9956-ffa450edef68

Otherwise, you can't install a module into a package root.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84238 13f79535-47bb-0310-9956-ffa450edef68

overrides for CFG_SBINDIR and CFG_TARGET have been evaluated.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84237 13f79535-47bb-0310-9956-ffa450edef68

the prototype bugs ("const char *"  vs.  "char * const"  & the like)


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84236 13f79535-47bb-0310-9956-ffa450edef68

prevent successful compilation on OS/390

Submitted by:	Ovies Brabson <oviesb@us.ibm.com>
Reviewed by:	Martin Kraemer


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84235 13f79535-47bb-0310-9956-ffa450edef68

	Note a couple of submitted patches.

Submitted by:	Raymond S Brand <rsbx@rsbx.net>
		and Andrew Ford <andrew@icarus.demon.co.uk>


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84234 13f79535-47bb-0310-9956-ffa450edef68

all methods using the ap_*_client_block() functions, most notably
POST and PUT. The functionality to switch input between protocol
parts (chunks) and (possibly binary) data had been missing all
the time, making chunked PUT impossible until now.

Hello TPF developers: I think these five lines in os/tpf/os.c
ought to be removed:
    else{
           if (r->method_number == M_PUT)
               ap_bsetflag(r->connection->client, B_ASCII2EBCDIC, 0);
               /* don't translate non-text files to EBCDIC */
    }
Also, your change to Vincent's mod_put should be undone.

In a next step, I will try to make EBCDIC conversion completely
configurable. Any volunteers for helping (designing/implementing)?


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84232 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84222 13f79535-47bb-0310-9956-ffa450edef68

mod_rewrite's expansion of expression back-references.

Submitted by: Cliff Woolley <jwoolley@wlu.edu>
Reviewed by: Ralf S. Engelschall
PR: 4766, 5389


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84221 13f79535-47bb-0310-9956-ffa450edef68

current version (and the patched older version) doesn't need this any
more.  We don't want to delete the code from unix/os.c, however,
because OpenStep's dyld needs it.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84219 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84218 13f79535-47bb-0310-9956-ffa450edef68

(and WILL, if he had apache-1.3.[0-9] before) have a file called
index.html in his .../htdocs/
Accept the presence of either index.html or index.html.en as
a clue for an existing htdocs tree.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84216 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84212 13f79535-47bb-0310-9956-ffa450edef68

	Add the long-awaited IndexOptions DescriptionWidth keyword.
	It was waiting for a way to avoid breaking HTML tags, but
	that's already been done as much as possible -- we can't
	do anything about breaking HTML elements without putting
	a full parser in.

PR:		2324


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84211 13f79535-47bb-0310-9956-ffa450edef68

for the presence of index.html, while it's been renamed to index.html.en
a couple of weeks ago.

Also, didn't we want to move /usr/local/apache/htdocs/manual/ out of
the user's htdocs/ tree to /usr/local/apache/manual/
and Alias /manual/ to /usr/local/apache/manual/ ?
And perhaps note the presence of the i18n versions of the entry page
in the CHANGES file?  [none of these are done in this commit]


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84209 13f79535-47bb-0310-9956-ffa450edef68

representation of the special characters, not the latin1
representation. This would result in invalid URI references
for, e.g., filenames generated by mod_autoindex.c
(when they had special chars in them)


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84208 13f79535-47bb-0310-9956-ffa450edef68

between implied conversion for protocol parts and configured
conversion for document data was missing. The effect of this was that
PDF files could not be read by Acrobat Reader (which sends long
lists of byte ranges in each request) when the server was apache
on ebcdic machines.

Submitted by:	Oliver Reh <Oliver.Reh@FAEDV-N.Bayern.de>
Reviewed by:	Martin Kraemer


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84207 13f79535-47bb-0310-9956-ffa450edef68

	Add a FoldersFirst keyword to the IndexOptions directive, which
	causes subdirectories to always appear first in FancyIndexed
	listings.

Reviewed by:	Ryan Bloom, Martin Kraemer, Jim Jagielski


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84206 13f79535-47bb-0310-9956-ffa450edef68

than write(), and it doesn't have a maximum transfer size of 16kB.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84204 13f79535-47bb-0310-9956-ffa450edef68

the check fails


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84195 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84184 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84183 13f79535-47bb-0310-9956-ffa450edef68

to allow one to override the corresponding variables from config.layout.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84182 13f79535-47bb-0310-9956-ffa450edef68

for supporting Berkeley-DB/3.x....


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84181 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84180 13f79535-47bb-0310-9956-ffa450edef68

More details were described in my original posting two weeks ago:

-----------------------------------------------------------------------
Today (on my birthday ;) I've hacked on and cleaned up snprintf
implementations for a forthcoming string library and discovered the following
bug in ap_snprintf.  If one tries to print a "long long" value which is just a
little bit lower than the maximimum "unsigned long" (not "unsigned long
long"!) value, ap_snprintf treats the value as a (signed) "long" instead of a
"long long":

| $ cat test.c
| #include <stdio.h>
| #include "ap_config.h"
| #include "ap.h"
|
| int main(int argc, char *argv[])
| {
|     char buf[1024];
|
|     ap_snprintf(buf, sizeof(buf), "%qd %qd", 4294967290ULL, 4294967297ULL);
|     printf("ap_snprintf: \"%s\"\n", buf);
|
|     sprintf(buf, "%qd %qd", 4294967290ULL, 4294967297ULL);
|     printf("sprintf:     \"%s\"\n", buf);
| }
| $ cc -I../os/unix -I../include -o test test.c libap.a && ./test
| ap_snprintf: "-6 4294967297"
| sprintf:     "4294967290 4294967297"

The reason is a wrong optimization test in ap_snprintf.c which
can be fixed this way:

Index: ap_snprintf.c
===================================================================
RCS file: /e/apache/REPOS/apache-1.3/src/ap/ap_snprintf.c,v
retrieving revision 1.37
diff -u -r1.37 ap_snprintf.c
--- ap_snprintf.c   1999/10/21 20:44:10 1.37
+++ ap_snprintf.c   1999/11/17 12:22:11
@@ -414,7 +413,7 @@
      * If the value is less than the maximum unsigned long value,
      * then we know we aren't using quads, so use the faster function
      */
-    if (num <= ULONG_MAX)
+    if (num <= ULONG_MAX && is_unsigned)
        return(conv_10( (wide_int)num, is_unsigned, is_negative,
           buf_end, len));

With this patch ap_snprintf and the (FreeBSD) vendor sprintf print the same
values:

| $ cc -I../os/unix -I../include -o test test.c libap.a && ./test
| ap_snprintf: "4294967290 4294967297"
| sprintf:     "4294967290 4294967297"

I think the patch is correct, although I have to admit that because of time
I've not checked the whole quad-fiddling in ap_snprintf.c now. So I do not
commit this myself. Instead someone else should review this first.
-----------------------------------------------------------------------


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@84179 13f79535-47bb-0310-9956-ffa450edef68