You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Aaron Bannert <aa...@clove.org> on 2001/10/16 03:28:48 UTC
move worker to new APR lock API?
I'd like to move the worker MPM to APR's new lock API, can I get some +1s?
This shouldn't be a problem for platforms that haven't fully implemented
the new API, since the worker MPM is a predominately unix MPM, and that
platform is the most supported under the new API. Not to mention that
worker is the only place in httpd that uses the new condition variable
functions.
-aaron
Re: move worker to new APR lock API?
Posted by Aaron Bannert <aa...@clove.org>.
> > I'd like to move the worker MPM to APR's new lock API, can I get some +1s?
> >
> > This shouldn't be a problem for platforms that haven't fully implemented
> > the new API, since the worker MPM is a predominately unix MPM, and that
> > platform is the most supported under the new API. Not to mention that
> > worker is the only place in httpd that uses the new condition variable
> > functions.
>
> ++1
The worker MPM has now been completely ported to the new lock API.
I'd like to start migrating other aspects of httpd over to the new methods.
As of this weekend APR now has complete implementations of the
apr_thread_mutex_t API on all supported platforms. The next step in my
plan for lock API conversion will be to move apr_lock_t/APR_INTRAPROCESS
locks to the new methods. This will include various modules and MPMs,
and should be fairly painless.
-aaron
Re: move worker to new APR lock API?
Posted by Ryan Bloom <rb...@covalent.net>.
On Monday 15 October 2001 06:28 pm, Aaron Bannert wrote:
> I'd like to move the worker MPM to APR's new lock API, can I get some +1s?
>
> This shouldn't be a problem for platforms that haven't fully implemented
> the new API, since the worker MPM is a predominately unix MPM, and that
> platform is the most supported under the new API. Not to mention that
> worker is the only place in httpd that uses the new condition variable
> functions.
++1
Ryan
______________________________________________________________
Ryan Bloom rbb@apache.org
Covalent Technologies rbb@covalent.net
--------------------------------------------------------------
Re: move worker to new APR lock API?
Posted by Ryan Bloom <rb...@covalent.net>.
On Monday 15 October 2001 06:33 pm, Justin Erenkrantz wrote:
> On Mon, Oct 15, 2001 at 06:28:48PM -0700, Aaron Bannert wrote:
> > I'd like to move the worker MPM to APR's new lock API, can I get some
> > +1s?
> >
> > This shouldn't be a problem for platforms that haven't fully implemented
> > the new API, since the worker MPM is a predominately unix MPM, and that
> > platform is the most supported under the new API. Not to mention that
> > worker is the only place in httpd that uses the new condition variable
> > functions.
>
> +1 in concept.
>
> FWIW, is there any reason why (in concept at least) that other platforms
> couldn't use worker? Is there any part of worker that doesn't rely on
> APR? -- justin
Worker forks processes. Until I get fork() ported to Windows, using worker
on Windows is a non-starter.
Ryan
______________________________________________________________
Ryan Bloom rbb@apache.org
Covalent Technologies rbb@covalent.net
--------------------------------------------------------------
Re: move worker to new APR lock API?
Posted by Brian Havard <br...@kheldar.apana.org.au>.
On Mon, 15 Oct 2001 18:33:08 -0700, Justin Erenkrantz wrote:
>On Mon, Oct 15, 2001 at 06:28:48PM -0700, Aaron Bannert wrote:
>> I'd like to move the worker MPM to APR's new lock API, can I get some +1s?
>>
>> This shouldn't be a problem for platforms that haven't fully implemented
>> the new API, since the worker MPM is a predominately unix MPM, and that
>> platform is the most supported under the new API. Not to mention that
>> worker is the only place in httpd that uses the new condition variable
>> functions.
>
>+1 in concept.
>
>FWIW, is there any reason why (in concept at least) that other platforms
>couldn't use worker? Is there any part of worker that doesn't rely on
>APR? -- justin
fork() is the killer. Most non-unix platforms either can't do it at all or
emulate it badly.
--
______________________________________________________________________________
| Brian Havard | "He is not the messiah! |
| brianh@kheldar.apana.org.au | He's a very naughty boy!" - Life of Brian |
------------------------------------------------------------------------------
Re: move worker to new APR lock API?
Posted by Justin Erenkrantz <je...@ebuilt.com>.
On Mon, Oct 15, 2001 at 06:28:48PM -0700, Aaron Bannert wrote:
> I'd like to move the worker MPM to APR's new lock API, can I get some +1s?
>
> This shouldn't be a problem for platforms that haven't fully implemented
> the new API, since the worker MPM is a predominately unix MPM, and that
> platform is the most supported under the new API. Not to mention that
> worker is the only place in httpd that uses the new condition variable
> functions.
+1 in concept.
FWIW, is there any reason why (in concept at least) that other platforms
couldn't use worker? Is there any part of worker that doesn't rely on
APR? -- justin