Date: Thu, 17 Jun 2004 22:29:46 +0200 From: "Cyrille Lefevre" <clefevre-lists@9online.fr> To: <ports@freebsd.org> Cc: Dan Langille <dan@langille.org> Subject: Re: HEADS UP - master/slave ports Message-ID: <065a01c454a9$d5c32a00$7890a8c0@dyndns.org> References: <40D1C3C1.26556.7C76FE30@localhost>
next in thread | previous in thread | raw e-mail | index | archive | help
"Dan Langille" <dan@langille.org> wrote:
> On 17 Jun 2004 at 16:50, Cyrille Lefevre wrote:
[snip]
> > there is no need to fork any sub-processes :
> >
> > # Try to determine if we are a slave port. These variables are used by
> > # FreshPorts and portsmon, but not yet by the ports framework itself.
> > _MASTERDIR= ${MASTERDIR:C,/([^/]+)/\.\.,,:C,/([^/]+)/\.\.,,}
> >
> > .if ${_MASTERDIR} != ${.CURDIR}
> > IS_SLAVE_PORT?= yes
> > # take your pick :
> > # _PORTSDIR= ${_MASTERDIR:C|([^/]+/[^/]+)/?$||}
> > # MASTERPORT= ${_MASTERDIR:C|${_PORTSDIR}||}
>
> real 15m24.018s
> user 10m39.785s
> sys 4m35.054s
>
> > # or
> > MASTERPORT= ${_MASTERDIR:H:T}/${_MASTERDIR:T}
>
> real 15m23.343s
> user 10m39.088s
> sys 4m34.888s
>
> > .else
> > IS_SLAVE_PORT?= no
> > MASTERPORT=
> > .endif
>
> These appear to be slightly faster than the previous stats I posted.
> Everything is pratically identical in user time. sys is where the
> savings are being made.
it's strange that this code (w/ no fork) is not much faster than 2 forks !
Ruslan, do you have any idea about that ?
CC -make maintainer
Cyrille Lefevre.
--
mailto:clefevre-lists@9online.fr
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?065a01c454a9$d5c32a00$7890a8c0>