httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jim Jagielski <...@jaguNET.com>
Subject Re: [PATCH] AcceptMutex and SingleListen runtime
Date Thu, 23 Aug 2001 14:29:46 GMT
Thanks for the feedback...
At 10:29 AM +0200 8/23/01, Martin Kraemer wrote:
>
>Shouldn't these be
>> +#define DEFAULT_SERIALIZED_ACCEPT_METHOD "fcntl"
>> +#define DEFAULT_SERIALIZED_ACCEPT_METHOD "flock"
>then?
>

The reason I have them as such is so that the actual C code takes
care of the quotes. I like to avoid situations where people need
to escape quotes at the command-line when compiling... But either
way works for me :)

>
>Can these 4 pointers be put into a structure? Then it is easier
>to assign either a pointer to an initialized structure, or
>assign the whole initialized structure en bloc. And it's better
>for encapsulation and object orientation.
>
> >  #ifdef USE_XXXXX_SERIALIZED_ACCEPT
>>      printf(" -D USE_XXXXX_SERIALIZED_ACCEPT\n");
>>  #endif

Yeah... I looking into using the apr method, which stores these in
pointers and then dereferences them as needed. The only reason I didn't
was to avoid all the extra "stuff" required to define the struct and
then change all the others calls to use it... But yes, I agree that
it's most likely worth it.

>And I'd also recommend the name change from USE_* to HAVE_*
>although I think we need some way of backward compatibility layer,
>because it has been an established and documented interface.
>
>> 	ptrans = ap_make_sub_pool(pconf);
>> +
>> +	init_single_listen(ap_single_listen);

Yeah... that's why I kept the USE_ stuff... Making a name change
would break current custom compiles (which could break anyway, but
at least it keeps things consistant).

Or maybe you're suggesting we keep USE_ to define the *default*
method and then create HAVE_ which lists the available methods.
So basically we would replace all USE_XXXXX_SERIALIZED_ACCEPT with
HAVE_XXXXX_SERIALIZED_ACCEPT, and the have USE_ set the default.
Hmmmm... I think I like that :)

My main goal was to have a small enough patch that added the
functionality without so many changes that people rejected it,
hence the limited general changes ;)

If the above changes are implemented, anyone foresee a -1 ??


>What happens to the char * init_single_listen() return value? errors?
>
>   Martin
>--
><Martin.Kraemer@Fujitsu-Siemens.com>    |       Fujitsu Siemens
>       <martin@apache.org>              |   81730  Munich,  Germany


-- 
===========================================================================
   Jim Jagielski   [|]   jim@jaguNET.com   [|]   http://www.jaguNET.com/
      "A society that will trade a little liberty for a little order
                   will lose both and deserve neither"

Mime
View raw message