Pages

[Gregg]

Well, Paul Q is rolling the 2.3.2 alpha as I type this, I plan to tackle it tonight, does Mario or Mika wanna copy to mess with?

[mario]

Sure I wanna try it out. I read the status file and hope that it runs smooth. I couldn't view major changes to 2.2

[Gregg]

I should have tried before I posted as it was DOA at linking libhttpd with pcre. I threw a few things at it but never had success. I think I'll take up bird watching instead 

[Gregg]

Changes with Apache 2.2.12
 
  *) mod_disk_cache: The module now turns off sendfile support if
     'EnableSendfile off' is defined globally. PR 41218.
     [Lars Eilebrecht, Issac Goldstand]

*) prefork: Fix child process hang during graceful restart/stop in
    configurations with multiple listening sockets.  PR 42829.  [Joe Orton,
    Jeff Trawick]

  *) mod_ssl: Add SSLRenegBufferSize directive to allow changing the
     size of the buffer used for the request-body where necessary
     during a per-dir renegotiation.  PR 39243.  [Joe Orton]
This one affects SNI

  *) mod_rewrite: Introduce DiscardPathInfo|DPI flag to stop the troublesome
     way that per-directory rewrites append the previous notion of PATH_INFO
     to each substitution before evaluating subsequent rules.
     PR38642 [Eric Covener]

  *) mod_authnz_ldap: Reduce number of initialization debug messages and make
     information more clear. PR 46342 [Dan Poirier]

  *) mod_cache: Introduce 'no-cache' per-request environment variable
     to prevent the saving of an otherwise cacheable response.
     [Eric Covener]

  *) core: Translate the status line to ASCII on EBCDIC platforms in
     ap_send_interim_response() and for locally generated "100 Continue"
     responses.  [Eric Covener]

  *) CGI: return 504 (Gateway timeout) rather than 500 when a script
     times out before returning status line/headers.
     PR 42190 [Nick Kew]

  *) prefork: Log an error instead of segfaulting when child startup fails
     due to pollset creation failures.  PR 46467.  [Jeff Trawick]

  *) mod_ext_filter: fix error handling when the filter prog fails to start,
     and introduce an onfail configuration option to abort the request
     or to remove the broken filter and continue.
     PR 41120 [Nick Kew]

  *) mod_include: support generating non-ASCII characters as entities in SSI
     PR 25202 [Nick Kew]

  *) core/utils: Enhance ap_escape_html API to support escaping non-ASCII chars
     [Nick Kew]

  *) mod_rewrite: fix "B" flag breakage by reverting r589343
     PR 45529 [Bob Ionescu <bobsiegen googlemail.com>]

  *) mod_cgid: fix segfault problem on solaris.
     PR 39332 [Masaoki Kobayashi <masaoki techfirm.co.jp>, Jeff Trawick]

  *) mod_ldap: Avoid a segfault when result->rc is checked in uldap_connection_init
     when result is NULL. This could happen if LDAP initialization failed.
     PR 45994.  [Dan Poirier <poirier pobox.com>]

  *) Set Listen protocol to "https" if port is set to 443 and no proto is specified
     (as documented but not implemented). PR 46066  [Dan Poirier <poirier pobox.com>]

  *) mod_cache: Correctly save Content-Encoding of cachable entity. PR 46401
     [Dan Poirier <poirier pobox.com>]

  *) Output -M and -S dumps (modules and vhosts) to stdout instead of stderr.
     PR 42571 and PR 44266 (dup).  [Dan Poirier <poirier pobox.com>]

  *) mod_cache: When an explicit Expires or Cache-Control header is set, cache
     normally non-cacheable response statuses. PR 46346.
     [Alex Polvi <alex polvi.net>]

Changes with Apache 2.2.11

[mario]

so 2.3.2 == 2.2.12 ?

[Gregg]

no no no no .. 2.3.2 != 2.2.12

2.3.x is derived each time straight from Trunk, 2.2 has is a branch taken off trunk way back when.

You can see it broke down on the SVN
http://svn.apache.org/viewvc/httpd/httpd/

You see the Trunk folder and branches, 1.3, 2.0 & 2.2 reside there.
Things however are tested in trunk, if deemed worthy or necessary, get backported to the branches from Trunk (sort-of some 2.2 and anything 2.0 usually required some manual manipulation to fit proper if the diffs are offset too much or have too much fuzz [whatever that means]).

[Gregg]

Tom succeeded at this and has sent me his notes on building the beast. Will give it a try again this week and see how it goes.

[mario]

Will that build also include the SNI patch?

[Gregg]

SNI is in 2.3.x, it was there in 2.3.0 .. it just didn't work. I think some of the reason is I do not think SSL is finished. It seemed not finished anyway. There have been quite some changes to SNI lately all of which are in the demo I set up. It is actually 2.2.12-dev .. I just don't want "dev" in the signature so I changed it to 11+, which is still correct .... technically 

[mario]

Did you correct it with mod_sec or in the source?

There was a discussion in apache.dev list about incompatible browser which can't use sni. Is there a list of that browser?
I like SNI cause it safes a lot of IPs   

Do you build 2.3 with VS6 oder VS9? 

In trunk I saw mod_heartbeatmonitor which is also intressting.

[Gregg]

1. Did you correct it with mod_sec or in the source?

2. There was a discussion in apache.dev list about incompatible browser which can't use sni. Is there a list of that browser?
I like SNI cause it safes a lot of IPs   

3. Do you build 2.3 with VS6 oder VS9? 

4. In trunk I saw mod_heartbeatmonitor which is also intressting.

1. Source: ap_release.h. I don't like to cheat lol

2. See "Support" at: http://en.wikipedia.org/wiki/Server_Name_Indication

3. VS6 first, always, less problems typically, but not always. If I succeed on 6 and get brave I go try on 9

4. mod_ratelimit too ... even if it does what I think it does (mod_bw), it's a definite plus. I'd like to see the features of mod_cband as well.


speaking of mod_bw .. do we have it? *scratching head*