httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <dgau...@arctic.org>
Subject Re: cvs commit: apache-2.0 STATUS
Date Thu, 19 Mar 1998 20:10:47 GMT


On Thu, 19 Mar 1998, Jim Jagielski wrote:

> My only concern is that we will deviate from the current "ideals"
> of having options for all systems. For example, we still support
> file-based scoreboards, when it's best to use shared memory.
> For shared mem, we support both SysV shmget and BSD mmap. Thus, we
> have all our options open. What I would really hate, and I'm hearing
> it already, is for us to say "well, if they don't have XXX and
> YYY and can't do this, f*ck 'em". Seems a big change from the old
> days...

NO NO NO!  The DESIGN supports all those old systems.  BUT FOLKS USING
THOSE SYSTEMS/OPTIONS WILL HAVE TO WRITE THE CODE!  The design is meant to
be modular and breaks the entire process down into enough granularity to
support the multiple process model! 

Sheesh.  Haven't I said this like a half dozen times already?  Why do you
keep on bringing it up like we're ignoring it? 

What you're hearing a bunch of us say is yeah we will support those old
systems in the design, but what turns our crank, what makes us interested,
what is best for apache, is if we move into the featureset of modern
operating systems.  So what if we can't support every damn feature under
the sun on Dynix/PTX 3.0.  What the hell does it matter?  There are only a
handful of those boxes left running... and I doubt any of them need high
performance feature-laden robust reliable webserving. 

Dean



Mime
View raw message