Re: cvs commit: apache-2.0/src/lib/apr/mmap/unix Makefile.in mmap.c mmap_h.h

[Bill Stoddard <st...@raleigh.ibm.com>]

>
>   ap_status_t ap_mmap_create(struct mmap_t **new, const char *fname,
>        ap_context_t *cont)
>   {
>
clip
>
>       if (stat(fname, &st) == -1)

>           /* we couldn't stat the file...probably doesn't exist! */
>           return APR_ENOFILE;
>       }
This will be a big performance hit if we use this function in place of the
mmap code already in http_core.c. We need a 'create' function that takes a
handle (or ap_file_t)  of an already opened file.

Bill

Re: cvs commit: apache-2.0/src/lib/apr/mmap/unix Makefile.in mmap.c mmap_h.h

[David Reid <ab...@dial.pipex.com>]

I see no problem with that.

david
----- Original Message -----
From: Bill Stoddard <st...@raleigh.ibm.com>
To: <ne...@apache.org>
Sent: 21 October 1999 05:17
Subject: Re: cvs commit: apache-2.0/src/lib/apr/mmap/unix Makefile.in mmap.c
mmap_h.h


> >
> >   ap_status_t ap_mmap_create(struct mmap_t **new, const char *fname,
> >        ap_context_t *cont)
> >   {
> >
> clip
> >
> >       if (stat(fname, &st) == -1)
>
> >           /* we couldn't stat the file...probably doesn't exist! */
> >           return APR_ENOFILE;
> >       }
> This will be a big performance hit if we use this function in place of the
> mmap code already in http_core.c. We need a 'create' function that takes a
> handle (or ap_file_t)  of an already opened file.
>
> Bill
>
>