httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ryan Bloom <...@raleigh.ibm.com>
Subject change over to APR.
Date Thu, 29 Apr 1999 19:10:24 GMT

I thought I would bring this topic up.  I am a bit worried about when we
have to do the change over to have the hybrid server using APR.  APR
development is coming along at a good rate right now, so this has been on
my radar for a little while.  The basic issue is that changing Apache to
use APR is a relatively big deal.

If we try to change apache to use apr_file_t's instead of int's, we can't
do it in just one file, the change has to be made in EVERY file, and it
has to be done in the same patch, or it won't compile until all the
patches are done.  Same thing goes for sockets, threads, processes, etc.

I don't think we want to try to make these changes while everybody is
modifying the hybrid server.  We definately don't want to make these
changes before 1.3.7 is released, because porting the patches from 1.3.7
to 2.0 would be next to impossible with the apr code in the server.

I have three possibilities, and I thought I would get the groups reaction
to them.

1)  Ignore this message, and just accept that immediately after Apache is
moved into 2.0, we have to change it to using APR, and it could be broken
for a while and not compiling while the change is made.

2)  Stop development of Apache-2.0 for some time after Apache-1.3.7 is
released.  Hopefully this time will be relatively quick.  During this
time, we change everything to use APR calls where applicable.  Where it
would most likely be me and/or Manoj and/or Bill making the change-over.

3)  Make the change-over before we release 1.3.7, as APR libraries are
finished. 

I REALLY hate option #3, but I am putting it in for completness.  I would
hate to be the person who had to deal with changing the patch to get 1.3.7
to the current state of the hybrid server.

I am perfectly happy doing things either of the other two ways, but I was
hoping to get the groups opinion, so that when it did finally happen,
everybody knew either why it was broken for weeks, or why development had
stopped for a time.  

Any thoughts?

Ryan

_______________________________________________________________________
Ryan Bloom		rbb@raleigh.ibm.com
4205 S Miami Blvd	
RTP, NC 27709		It's a beautiful sight to see good dancers 
			doing simple steps.  It's a painful sight to
			see beginners doing complicated patterns.	



Mime
View raw message