You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@apr.apache.org by cl...@apache.org on 2004/07/09 22:29:05 UTC
cvs commit: apr/threadproc/netware proc.c
clar 2004/07/09 13:29:05
Modified: threadproc/netware proc.c
Log:
Added bit mask operation for detach and addrspace fields to make possible overloading on fcts parameter.
Revision Changes Path
1.30 +2 -2 apr/threadproc/netware/proc.c
Index: proc.c
===================================================================
RCS file: /home/cvs/apr/threadproc/netware/proc.c,v
retrieving revision 1.29
retrieving revision 1.30
diff -u -r1.29 -r1.30
--- proc.c 14 Jun 2004 17:26:19 -0000 1.29
+++ proc.c 9 Jul 2004 20:29:05 -0000 1.30
@@ -179,7 +179,7 @@
APR_DECLARE(apr_status_t) apr_procattr_detach_set(apr_procattr_t *attr, apr_int32_t detach)
{
- attr->detached = detach;
+ attr->detached = (detach & 1);
return APR_SUCCESS;
}
@@ -268,7 +268,7 @@
APR_DECLARE(apr_status_t) apr_procattr_addrspace_set(apr_procattr_t *attr,
apr_int32_t addrspace)
{
- attr->addrspace = addrspace;
+ attr->addrspace = (addrspace & 2);
return APR_SUCCESS;
}
Re: cvs commit: apr/threadproc/netware proc.c
Posted by Joe Orton <jo...@redhat.com>.
On Fri, Jul 09, 2004 at 08:29:05PM -0000, clar@apache.org wrote:
> --- proc.c 14 Jun 2004 17:26:19 -0000 1.29
> +++ proc.c 9 Jul 2004 20:29:05 -0000 1.30
> @@ -179,7 +179,7 @@
>
> APR_DECLARE(apr_status_t) apr_procattr_detach_set(apr_procattr_t *attr, apr_int32_t detach)
> {
> - attr->detached = detach;
> + attr->detached = (detach & 1);
> return APR_SUCCESS;
This implements a different interface to that described in the header,
so -1 again. The point of the discussion we had is that you have to do
the messy work *inside mod_cgi*, because that's where the binary compat
requirement is. So mod_cgi should be doing e.g.:
apr_procattr_detach_set(procattr, e_info->detached & 1)
apr_procattr_addrspace_set(procattr, e_info->detached & 2)
and it needs to be documented in mod_cgi.h that the detached field is
used as a bitfield like this in the 2.0 branch. For mod_cgi in HEAD it
needn't be weird of course, the addrspace field could be left in there.
joe
Re: cvs commit: apr/threadproc/netware proc.c
Posted by Joe Orton <jo...@redhat.com>.
On Fri, Jul 09, 2004 at 08:29:05PM -0000, clar@apache.org wrote:
> --- proc.c 14 Jun 2004 17:26:19 -0000 1.29
> +++ proc.c 9 Jul 2004 20:29:05 -0000 1.30
> @@ -179,7 +179,7 @@
>
> APR_DECLARE(apr_status_t) apr_procattr_detach_set(apr_procattr_t *attr, apr_int32_t detach)
> {
> - attr->detached = detach;
> + attr->detached = (detach & 1);
> return APR_SUCCESS;
This implements a different interface to that described in the header,
so -1 again. The point of the discussion we had is that you have to do
the messy work *inside mod_cgi*, because that's where the binary compat
requirement is. So mod_cgi should be doing e.g.:
apr_procattr_detach_set(procattr, e_info->detached & 1)
apr_procattr_addrspace_set(procattr, e_info->detached & 2)
and it needs to be documented in mod_cgi.h that the detached field is
used as a bitfield like this in the 2.0 branch. For mod_cgi in HEAD it
needn't be weird of course, the addrspace field could be left in there.
joe