[httperf] file descriptors

[Mark Nottingham]

Ah; I was looking at Adrian's fork. Sorry.



On 08/08/2008, at 4:20 PM, Theodore Bullock wrote:

> On Fri, 2008-08-08 at 15:58 +1000, Mark Nottingham wrote:
>> Right now that's specific to idleconn; if you don't enable it, it
>> doesn't check;
>>
>> if test "x$enable_idleconn" = xyes; then
>>     AC_CHECK_LIB([event],
>>         [event_init],
>>         ,
>>         AC_MSG_ERROR([libevent is required to build idleconn]) )
>> fi
>>
>> I'm tempted to commit the obvious fix, but considering that I know
>> less about AC than adrian does... ;)
>
> I just double checked this....  My original statement is correct, you
> might check if your sources are up to date.
>
> CVS has the following without the libevent test
>
> # Are we building the idleconn program?
> AC_ARG_ENABLE([idleconn],
> 	AS_HELP_STRING([--enable-idleconn],
> 		[Build the idleconn program (default=no), requires libevent]),
> 	[enable_idleconn=$enableval],
> 	[enable_idleconn=no])
>
> AM_CONDITIONAL(IDLECONN,
> 	[test "x$enable_idleconn" = xyes])
>
> # Checks for libraries.
> AC_CHECK_LIB(m, sqrt)
> AC_CHECK_LIB(crypto, main)
> AC_CHECK_LIB(ssl, SSL_version, , AC_MSG_WARN([SSL Disabled]) )
> AC_CHECK_LIB([event],
> 		[event_init],
> 		,
> 		[AC_MSG_ERROR([libevent is required to build idleconn])])
>
>
> -- 
> Ted Bullock <tbullock at canada.com>
> http://www.comlore.com
> High Voltage Software Engineer
>

--
Mark Nottingham       mnot at yahoo-inc.com