You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by rb...@covalent.net on 2000/05/24 16:39:53 UTC
Re: [STATUS] (apache-2.0) Wed May 24 23:45:12 EDT 2000
Okay, which of these can go away. :-)
> RELEASE SHOWSTOPPERS:
> * Win32: Test access logging with multiple threads. Will the
> native file I/O calls serialize automagically like the
> CRT calls or do we need to add region locking each time
> we access the logs?
> Status:
Shouldn't we know the status of this already? Can somebody check their
logs on NT to see if this is working?
> * The handler hook needs to use the new registration system
I thought we agreed this was wrong.
> RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
> * There are still a number of places in the code where we are
> loosing error status (i.e. throwing away the error returned by a
> system call and replacing it with a generic error code)
How many of these are left? Where are they?
> * Switch to autoconf, et al. for configuration.
> Status: Manoj has placed an initial version into the 2.0
> repository. Todos include:
Can this be removed, it's already a show stopper.
> * MaxRequestsPerChild measures connections, not requests.
> Until someone has a better way, we'll probably just rename it
> "MaxConnectionsPerChild".
Is this REALLY An issue?
Ryan
_______________________________________________________________________________
Ryan Bloom rbb@apache.org
406 29th St.
San Francisco, CA 94131
-------------------------------------------------------------------------------
Re: [STATUS] (apache-2.0) Wed May 24 23:45:12 EDT 2000
Posted by Bill Stoddard <re...@attglobal.net>.
>
> Okay, which of these can go away. :-)
>
> > RELEASE SHOWSTOPPERS:
> > * Win32: Test access logging with multiple threads. Will the
> > native file I/O calls serialize automagically like the
> > CRT calls or do we need to add region locking each time
> > we access the logs?
> > Status:
>
> Shouldn't we know the status of this already? Can somebody check their
> logs on NT to see if this is working?
>
It's broken. Either because of cross process sync problems or cross thread
problems. I fooled with LockFile/UnlockFile for awhile w/o ever getting it
to work.
>
> > * MaxRequestsPerChild measures connections, not requests.
> > Until someone has a better way, we'll probably just rename it
> > "MaxConnectionsPerChild".
>
> Is this REALLY An issue?
IMHO, no. This is not an issue.
Repost of old 2.0 suexec patch
Posted by Manoj Kasichainula <ma...@io.com>.
On Wed, May 24, 2000 at 04:35:46PM -0700, rbb@covalent.net wrote:
> I was looking through it from the archives earlier today. Could you
> either re-post or commit it? I would personally rather you committed it,
> because that gives people more of a chance to hack on it. (Plus it
> removes an item from the STATUS file). I think you are more likely to get
> results by committing these days. We are in a commit->review stage. :-)
Well, until I have a chance to actually try the patch out (maybe
tonight, hopefully in the next day or two), I've just attached the
original message.
Re: [STATUS] (apache-2.0) Wed May 24 23:45:12 EDT 2000
Posted by rb...@covalent.net.
I was looking through it from the archives earlier today. Could you
either re-post or commit it? I would personally rather you committed it,
because that gives people more of a chance to hack on it. (Plus it
removes an item from the STATUS file). I think you are more likely to get
results by committing these days. We are in a commit->review stage. :-)
Ryan
On Thu, 25 May 2000, Manoj Kasichainula wrote:
> On Wed, May 24, 2000 at 04:25:30PM -0700, rbb@covalent.net wrote:
> > Cool. :-) Hey Manoj, you posted a patch to make SuExec work a while ago,
> > what ever happened with that?
>
> It's waiting. I wanted someone to audit the concept before committing
> because it's security-critical, and I'm not a big fan of it's current
> design.
>
> The patch should be in the archives; I can repost if anyone requests.
>
_______________________________________________________________________________
Ryan Bloom rbb@apache.org
406 29th St.
San Francisco, CA 94131
-------------------------------------------------------------------------------
Re: [STATUS] (apache-2.0) Wed May 24 23:45:12 EDT 2000
Posted by Manoj Kasichainula <ma...@io.com>.
On Wed, May 24, 2000 at 04:25:30PM -0700, rbb@covalent.net wrote:
> Cool. :-) Hey Manoj, you posted a patch to make SuExec work a while ago,
> what ever happened with that?
It's waiting. I wanted someone to audit the concept before committing
because it's security-critical, and I'm not a big fan of it's current
design.
The patch should be in the archives; I can repost if anyone requests.
Re: [STATUS] (apache-2.0) Wed May 24 23:45:12 EDT 2000
Posted by rb...@covalent.net.
> > Glad I'm not imagining things. If nobody stands up to volunteer to change
> > the hook system, I'm going to move this from showstoppers to "WOULD BE
> > REAL NICE TO WRAP THESE UP:" and clarify what would be needed to fix it
> > correctly. BTW, I am personally -1 for this change. The current system
> > works, and makes sense. I think modifying it to work will end up being a
> > hell of a hack. :-)
>
> I think Greg is saying just remove that item from STATUS. I +1 that (I
> added the item in the first place because I was confused).
Cool. :-) Hey Manoj, you posted a patch to make SuExec work a while ago,
what ever happened with that?
Ryan
_______________________________________________________________________________
Ryan Bloom rbb@apache.org
406 29th St.
San Francisco, CA 94131
-------------------------------------------------------------------------------
Re: [STATUS] (apache-2.0) Wed May 24 23:45:12 EDT 2000
Posted by Manoj Kasichainula <ma...@io.com>.
On Wed, May 24, 2000 at 04:21:50PM -0700, rbb@covalent.net wrote:
> On Thu, 25 May 2000, Greg Stein wrote:
>
> > Yah, it's wrong. The hook system says "somebody please do something." For
> > the handler, we want "hey YOU. handle this." Things such as the
> > create/merge config functions, and the handler, are *directed*, so the
> > hook system doesn't work for them.
>
> Glad I'm not imagining things. If nobody stands up to volunteer to change
> the hook system, I'm going to move this from showstoppers to "WOULD BE
> REAL NICE TO WRAP THESE UP:" and clarify what would be needed to fix it
> correctly. BTW, I am personally -1 for this change. The current system
> works, and makes sense. I think modifying it to work will end up being a
> hell of a hack. :-)
I think Greg is saying just remove that item from STATUS. I +1 that (I
added the item in the first place because I was confused).
Re: [STATUS] (apache-2.0) Wed May 24 23:45:12 EDT 2000
Posted by Greg Stein <gs...@lyra.org>.
On Wed, 24 May 2000 rbb@covalent.net wrote:
> On Thu, 25 May 2000, Greg Stein wrote:
> > On Wed, 24 May 2000 rbb@covalent.net wrote:
> > >...
> > > > * The handler hook needs to use the new registration system
> > >
> > > I thought we agreed this was wrong.
> >
> > Yah, it's wrong. The hook system says "somebody please do something." For
> > the handler, we want "hey YOU. handle this." Things such as the
> > create/merge config functions, and the handler, are *directed*, so the
> > hook system doesn't work for them.
>
> Glad I'm not imagining things. If nobody stands up to volunteer to change
> the hook system, I'm going to move this from showstoppers to "WOULD BE
> REAL NICE TO WRAP THESE UP:" and clarify what would be needed to fix it
> correctly. BTW, I am personally -1 for this change. The current system
> works, and makes sense. I think modifying it to work will end up being a
> hell of a hack. :-)
"modifying it to work" ?? whatthehell? :-)
If it works, then it works. Just axe the line.
Cheers,
-g
--
Greg Stein, http://www.lyra.org/
Re: [STATUS] (apache-2.0) Wed May 24 23:45:12 EDT 2000
Posted by rb...@covalent.net.
On Thu, 25 May 2000, Greg Stein wrote:
> On Wed, 24 May 2000 rbb@covalent.net wrote:
> >...
> > > * The handler hook needs to use the new registration system
> >
> > I thought we agreed this was wrong.
>
> Yah, it's wrong. The hook system says "somebody please do something." For
> the handler, we want "hey YOU. handle this." Things such as the
> create/merge config functions, and the handler, are *directed*, so the
> hook system doesn't work for them.
Glad I'm not imagining things. If nobody stands up to volunteer to change
the hook system, I'm going to move this from showstoppers to "WOULD BE
REAL NICE TO WRAP THESE UP:" and clarify what would be needed to fix it
correctly. BTW, I am personally -1 for this change. The current system
works, and makes sense. I think modifying it to work will end up being a
hell of a hack. :-)
Ryan
_______________________________________________________________________________
Ryan Bloom rbb@apache.org
406 29th St.
San Francisco, CA 94131
-------------------------------------------------------------------------------
Re: [STATUS] (apache-2.0) Wed May 24 23:45:12 EDT 2000
Posted by Greg Stein <gs...@lyra.org>.
On Wed, 24 May 2000 rbb@covalent.net wrote:
>...
> > * The handler hook needs to use the new registration system
>
> I thought we agreed this was wrong.
Yah, it's wrong. The hook system says "somebody please do something." For
the handler, we want "hey YOU. handle this." Things such as the
create/merge config functions, and the handler, are *directed*, so the
hook system doesn't work for them.
Cheers,
-g
--
Greg Stein, http://www.lyra.org/