You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@httpd.apache.org by "William A. Rowe, Jr." <wr...@rowe-clan.net> on 2006/01/12 21:57:09 UTC

Any concerns with 2.2.1 by the weekend?

I'm offering to RM this package late friday or sometime Saturday if there's
no serious objection.  Fixing or not fixing partial page results is obviously
not a regression, although I'm happy to wait an extra couple days if the fix
has been agreed upon, committed and is only undergoing some extra bashing-upon.

Part of the Win32 binaries issue is that the -src.zip still required too much
tweaking to be called the official version, by the time i'd finished build/test
I'd had to make too many other 'minor tweaks'.  I'd feel much warmer releasing
2.2.1 as binaries on Win32, as the build system (now) corresponds to the actual
product.

I can do one of three things to eliminate awk for generating .rc files;

  * use one common module .rc file with some /D inputs for module name, etc.

  * create the dozens of .rc files in svn

  * have the buildconf create all the .rc files

I'm leaning twords the first option.

Bill

Re: Any concerns with 2.2.1 by the weekend?

Posted by Brian Akins <br...@turner.com>.

William A. Rowe, Jr. wrote:
> I'm offering to RM this package late friday or sometime Saturday if there's
> no serious objection.  Fixing or not fixing partial page results is 
> obviously
> not a regression, although I'm happy to wait an extra couple days if the 
> fix
> has been agreed upon, committed and is only undergoing some extra 
> bashing-upon.

I'd like the "partial page" fix to be in 2.2.1 if possible.


-- 
Brian Akins
Lead Systems Engineer
CNN Internet Technologies

Re: Any concerns with 2.2.1 by the weekend?

Posted by Ruediger Pluem <rp...@apache.org>.


On 01/12/2006 09:57 PM, William A. Rowe, Jr. wrote:
> I'm offering to RM this package late friday or sometime Saturday if there's
> no serious objection.  Fixing or not fixing partial page results is
> obviously
> not a regression, although I'm happy to wait an extra couple days if the
> fix
> has been agreed upon, committed and is only undergoing some extra
> bashing-upon.

I think the following thing from the backport list must be solved in 2.2.1
(maybe with different patch if the one proposed is not ok):

- mod_cache: Fix PR38017 (mod_cache not working in reverse proxy setup?)
  This is a regression against 2.0.x and makes it impossible to use mod_cache
  on reverse proxies. I just added this to the showstopper section of the
  STATUS file.

The following from backport list should be fixed:

- mod_ssl: Fix PR37791 (CVEID: CAN-2005-3357) (SEGV if the client is
                        connection plain to a SSL enabled port)
  Although minor, it is a security issue and should be fixed plus it only misses one vote.

- The partial pages issue. Although improvements could be made to the current patch
  I think it solves the actual problem suitable. Improvements could be done later.

Anyway I think at least PR38017 and PR37791 can be sorted out very quickly.

Apart from the points above I am fine with releasing 2.2.1.
Thanks for volunteering as RM Bill.


Regards

Rüdiger

Re: Any concerns with 2.2.1 by the weekend?

Posted by Jim Jagielski <ji...@jaguNET.com>.

I would have liked to do it, but there's no way I
could volunteer for this weekend (a few family things
popped up with my Dad in the hospital and my
Uncle/Godfather passing away)... But my comments
about 2.2.1 are:

    1. I think the proxy fixes should be part of 2.2.1,
       but I think that, especially now with all the
       revisions that we worked through in trunk,
       they require more testing. So maybe a coupla
       extra days would be nice.

    2. Monday is a holiday for many US companies.
       So timing might be an issue.

    3. Showstoppers exist, but I haven't had time to
       investigate them, but we can't ignore them :)

On Jan 12, 2006, at 3:57 PM, William A. Rowe, Jr. wrote:

> I'm offering to RM this package late friday or sometime Saturday if  
> there's
> no serious objection.  Fixing or not fixing partial page results is  
> obviously
> not a regression, although I'm happy to wait an extra couple days  
> if the fix
> has been agreed upon, committed and is only undergoing some extra  
> bashing-upon.
>
> Part of the Win32 binaries issue is that the -src.zip still  
> required too much
> tweaking to be called the official version, by the time i'd  
> finished build/test
> I'd had to make too many other 'minor tweaks'.  I'd feel much  
> warmer releasing
> 2.2.1 as binaries on Win32, as the build system (now) corresponds  
> to the actual
> product.
>
> I can do one of three things to eliminate awk for generating .rc  
> files;
>
>  * use one common module .rc file with some /D inputs for module  
> name, etc.
>
>  * create the dozens of .rc files in svn
>
>  * have the buildconf create all the .rc files
>
> I'm leaning twords the first option.
>
> Bill
>

Re: Any concerns with 2.2.1 by the weekend?

Posted by Nick Kew <ni...@webthing.com>.

On Friday 13 January 2006 07:30, Mladen Turk wrote:
> William A. Rowe, Jr. wrote:
> > I'm offering to RM this package late friday or sometime Saturday if
> > there's no serious objection.  Fixing or not fixing partial page results
> > is obviously
>
> Backport of ajp Cookie2 fix from HEAD.
>
>      * mod_proxy: Fix Cookie2 header problems that originates back
>        from mod_jk. Cookie2 was always sent as Cookie.
>        http://svn.apache.org/viewcvs.cgi?rev=358769&view=rev

+1 on your fix there (just entered into STATUS).  We have a handful of
approved backports, though that one needs another +1.

I note we have a showstopper PR38017 marked regarding mod_cache.
Rüdiger's fix looks OK, but I don't feel comfortable voting without figuring
out how it interacts with virtualhosts.  And I still think there's far too
much complexity for questionable gain in the whole quick-handler thing,
but that's another story.

-- 
Nick Kew

Re: Searching mail archives (Re: Any concerns with 2.2.1 by the weekend?)

Posted by Maxime Petazzoni <ma...@bulix.org>.

* Nick Kew <ni...@webthing.com> [2006-01-13 13:39:14]:

> To follow that up, I went to mail-archives to look for the
> message-ID (I don't keep a private archive of this list), but I
> don't find a search facility.  What am I missing?

The SoC project httpd-mbox-search did not succeed. I did not get any
code from the corresponding student.

Anyone heard of it after September 1st?

- Sam
-- 
Maxime Petazzoni (http://www.bulix.org)
 -- gone crazy, back soon. leave message.

Searching mail archives (Re: Any concerns with 2.2.1 by the weekend?)

Posted by Nick Kew <ni...@webthing.com>.

Just looking at STATUS, there's a fix:

    * HTTP: If a connection aborts while waiting for a chunked line,
      flag the connection as errored out.
      http://svn.apache.org/viewcvs.cgi?rev=354630&view=rev
      Message-ID: <43...@web.turner.com>
      +1: jerenkrantz, jim, wrowe

Looking at that, I wondered why the return value from ap_pass_brigade
is being discarded without handling, propagation or reporting round
about line 324 of http_filters.c in Justin's patch.

To follow that up, I went to mail-archives to look for the message-ID
(I don't keep a private archive of this list), but I don't find a search
facility.  What am I missing?

-- 
Nick Kew

Re: Any concerns with 2.2.1 by the weekend?

Posted by Mladen Turk <mt...@apache.org>.

William A. Rowe, Jr. wrote:
> I'm offering to RM this package late friday or sometime Saturday if there's
> no serious objection.  Fixing or not fixing partial page results is 
> obviously

Backport of ajp Cookie2 fix from HEAD.

     * mod_proxy: Fix Cookie2 header problems that originates back
       from mod_jk. Cookie2 was always sent as Cookie.
       http://svn.apache.org/viewcvs.cgi?rev=358769&view=rev


Regards,
Mladen.