The final, inconsequential patch which makes the diff for request
  body processing between trunk/ and 2.1.x/ very easy to read.

  This patch consists of nothing except whitespace changes, and the
  change from b to bb for the local bucket brigade variable name,
  to disambiguate it from a bucket.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230744 13f79535-47bb-0310-9956-ffa450edef68

  Fix some very minor nits in ap_proxy_http_request() which make it
  quite simple to diff to the current trunk/ code, ensuring the most
  accurate review possible.

  Consists only of whitespace and line spacing changes, and a change
  from 'bb' to 'header_brigade' for this local argument variable's name,
  so there are no functional changes here.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230741 13f79535-47bb-0310-9956-ffa450edef68

  Sync to trunk/, add an extra measure of paranoia to the cl + te case.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230738 13f79535-47bb-0310-9956-ffa450edef68

  Commit a comment from trunk/, this brings spool_reqbody_cl in sync
  with trunk/.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230737 13f79535-47bb-0310-9956-ffa450edef68

  Backport r230735, we need not look at the first bucket for EOS, 
  because the outer while loop protected us from that case.  

  Backport the header brigade changes as it's impossible to have
  a body request waiting for a final send.  Look at seen_eos to 
  flush us in the request body loop, and handle the only exception,
  (header_brigade), outside of that loop.

  This brings stream_reqbody_cl in sync with the trunk.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230736 13f79535-47bb-0310-9956-ffa450edef68

  Backport the fix to handling Connection: close.  The existing code
  was impossible to follow; the new code sets up p_conn->close correctly,
  and uses that evaluated value to inject the appropriate choice immediately
  before passing the request to the backend server.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230729 13f79535-47bb-0310-9956-ffa450edef68

  Backport the corrected C-L streamed, v.s. spooled, v.s. T-E: chunked
  selection logic from httpd trunk/.  This now correctly chooses the
  most portable mechanism (e.g. C-L) when we can see the entire body,
  even for chunked bodies from the client, falls back on spool cl when
  it's necessary, and uses chunked when we have faith in it.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230728 13f79535-47bb-0310-9956-ffa450edef68

  Backport the rejection of non-'chunked' values in the transfer
  encoding; we simply don't know what, exactly to do with them.

  Backport the reporting of 'both C-L and T-E' when we encounter
  this edge case, setting the connection up to close down once
  we finish (perhaps we were victims of a request splitting attack).


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230727 13f79535-47bb-0310-9956-ffa450edef68

  Backport the fix to an edge case, it's now possible for a primary
  request which has a body in spite of what was determined by the
  header parsing; this would usually be due to an input filter between
  the client request and mod_proxy.  Add another consideration, and
  force the C-L determination if we saw bytes in already.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230726 13f79535-47bb-0310-9956-ffa450edef68

  Backport the fix to an edge case; it's now much more efficient to 
  entirely skip request body determinations for subrequests, stuff
  in an EOS and we are off to the races, ready to create a body-less
  proxied subrequest.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230725 13f79535-47bb-0310-9956-ffa450edef68

  Backport the body pre-fetch code from trunk/, modifying the three
  stream_te / stream_cl / spool_cl functions to presume an input_brigade
  before fighting with fetching additional body content.

  We will be using the bytes_read in a later commit to make a better
  decision about choosing to send a C-L or chunked body to the backend.

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230724 13f79535-47bb-0310-9956-ffa450edef68

  Backport yet another security fix; if stream_cl exceeds the 'stated' 
  CL which proxy_request_body asked us to send, then we have to quit 
  forwarding any more bytes (we won't even pass the header if we 
  hadn't yet.)

  Closes an HTTP Request splitting edge case.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230722 13f79535-47bb-0310-9956-ffa450edef68

  Fix a buglett in backport 230718; get our selection of brigades correct.

  (tis what happens when our code is this far out-of-sync with trunk, sorry.)

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230720 13f79535-47bb-0310-9956-ffa450edef68

  Backport fix r230718; a soon-to-be impossible edge case; we are always
  sending a body (even an 'empty' body) using stream_chunked, so we must
  always send the trailing "0" end of body marker.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230719 13f79535-47bb-0310-9956-ffa450edef68

  Finish up the header brigades up-front; we simply know ahead of time if
  they are necessary or not.  This changes one behavior; the stream_chunked
  now always sends a transfer encoded body; even if it's nothing but an
  empty body.

  Also, Fix a bugglet in backporting the input_brigade argument; need the
  bucket_alloc already.

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230716 13f79535-47bb-0310-9956-ffa450edef68

  Pre-commit a simple bit from r219224, pass an initialized input_brigade
  off to the spool/stream functions.  Effectively a no-op so far, but in
  a couple of patches, the reason becomes obvious.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230714 13f79535-47bb-0310-9956-ffa450edef68

  Backport a logging change to make things clearer; includes jorton's r224721 fix.

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230712 13f79535-47bb-0310-9956-ffa450edef68

  Backport part of r216156 and very minor bits of r218978;

  Collapse the handling of Content-Length / Transfer-Encoding header
  values into the headers loop saving a bit of cpu and making the
  results clearer, handle the (r->main) subrequest case a bit more 
  efficiently as well, cutting out some duplicate tests, and a mild
  whitespace issue in the If-Unmodified-Since test line.

  Fixes a subrequest bug where subrequests *still* would attempt to
  read Transfer-Encoding: chunked request bodies.

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230711 13f79535-47bb-0310-9956-ffa450edef68

  Backport part of r216156;

  send_request_body makes it very difficult to follow all the mistakes
  in this code.  Fold send_request_body into ap_proxy_http_request
  so that proxy_http_request makes all of the elections.



git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230709 13f79535-47bb-0310-9956-ffa450edef68

  End apr_natstrcasecmp, backporting part of r216111 (there may remain
  other apr_natstrcasecmp abuse in other files).

  Still unsure if apr_strnatcasecmp of the hostname is deliberate to
  handle leading 0's in ip addresses, so leaving that single use case.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230708 13f79535-47bb-0310-9956-ffa450edef68

  Now backport whitespace-only changes from r209836, making the rest of
  the patches more legible, and the resulting code comparable   to httpd/trunk/modules/proxy/mod_proxy_http.c.

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230706 13f79535-47bb-0310-9956-ffa450edef68

  Prepare to backport r209836 whitespace changes; this is the one
  minor code change to declare *buf once, obliterating the need for
  a number of code blocks throughout this function.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230704 13f79535-47bb-0310-9956-ffa450edef68

  Create proxy-reqbody patch evaluation branch.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody-2.0.x@230703 13f79535-47bb-0310-9956-ffa450edef68

* modules/ldap/util_ldap_cache.c (util_ldap_cache_init): Use the
actual available size of the shm segment not the requested size.
Ensure the requested size is aligned.  Check errors from apr_rmm_init.

Reviewed by: jorton, minfrin, bnicholes


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@230474 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@230472 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@230455 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227521 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227435 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227423 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227422 13f79535-47bb-0310-9956-ffa450edef68

Translation into Spanish of the Authentication, Authorization and Access Control tutorial.
Translation made by Gustavo Nerea (informatica.gustavo @ gmail.com)
Translation reviewed by Jesus Blanco (blanco @ bitrock.com)



git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227421 13f79535-47bb-0310-9956-ffa450edef68

* modules/ldap/util_ldap_cache_mgr.c (util_ald_cache_insert): Fix a
cache corruption case: ensure that there is room in the cache for a
copy of the payload before inserting the node.

PR: 34209
Reviewed by: jorton, bnicholes, minfrin


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227332 13f79535-47bb-0310-9956-ffa450edef68

* server/log.c (piped_log_spawn): Return the APR error code,
as expected by piped_log_maintenance (and fixing "Unknown
error" messages in error path there).
(ap_open_piped_log): Expect an APR error code.

* server/log.c (ap_open_piped_log): Remove errno handling.

* server/log.c (piped_log_spawn): Don't leak an fd for each
invocation.

Reviewed by: jorton, trawick, wrowe


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227329 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227269 13f79535-47bb-0310-9956-ffa450edef68

connection timeout on the connection, not globally, avoiding segfaults
in older version of OpenLDAP (ITS #3487).

Submitted by: Brad Nicholes
Reviewed by: jorton, wrowe


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227253 13f79535-47bb-0310-9956-ffa450edef68

- propose two more mod_ldap fixes


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227213 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@227040 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@225786 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@225488 13f79535-47bb-0310-9956-ffa450edef68

  Vetoed the first attempt at a 2.1-dev mod_proxy_http request body
  backport - so note this as a showstopper rather than reverting the code.

  Also note the RFC bug that the -trace.patch corrects.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.0.x@219782 13f79535-47bb-0310-9956-ffa450edef68