Date: Sat, 15 Feb 1997 19:34:59 -0800 (PST)
From: Dean Gaudet <dgaudet@arctic.org>
To: new-httpd@hyperreal.com
Subject: Re: [PATCH] my version of the mod_include outbut buffering (fwd)
In-Reply-To: <199702160329.WAA07258@shado.jaguNET.com>
Message-ID: <Pine.LNX.3.95dg2.970215193248.20778B-100000@twinlark.arctic.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: new-httpd-owner@apache.org
Precedence: bulk
Reply-To: new-httpd@hyperreal.com

I actually did a bunch of the work implementing the boyer-moore-horspool
stuff so that we could speed this up even more... but that's definately a
feature and not for 1.2. 

I'm running Marc's patch on one of my servers, so I give it a +1... but I
respect Rob's sentiment that this is a feature.
 
Dean

On Sat, 15 Feb 1997, Jim Jagielski wrote:

> +1...
> 
> It would be interesting to see if it still improves things,
> performance-wise, when using a larger SNDBUF size for the
> socket. Either way, it's win-win!
> 
> Marc Slemko wrote:
> > 
> > Repost.  On a large file with no SSIs but still being parsed by
> > mod_include it sped up output by two to three times on my FreeBSD system.
> > 
> > Index: mod_include.c
> > ===================================================================
> > RCS file: /export/home/cvs/apache/src/mod_include.c,v
> > retrieving revision 1.21
> > diff -c -r1.21 mod_include.c
> > *** mod_include.c	1997/01/20 04:28:13	1.21
> > --- mod_include.c	1997/02/08 02:49:56
> > ***************
> > *** 113,156 ****
> >       }
> >   }
> >   
> > ! #define GET_CHAR(f,c,r,p) \
> >    { \
> >      int i = getc(f); \
> > !    if(feof(f) || ferror(f) || (i == -1)) { \
> > !         pfclose(p,f); \
> > !         return r; \
> >      } \
> >      c = (char)i; \
> >    }
> >   
> > - /* --------------------------- Parser functions --------------------------- */
> > - 
> > - /* Grrrr... rputc makes this slow as all-get-out.  Elsewhere, it doesn't
> > -  * matter much, but this is an inner loop...
> > -  */
> > - 
> >   int find_string(FILE *in,char *str, request_rec *r, int printing) {
> >       int x,l=strlen(str),p;
> >       char c;
> >   
> >       p=0;
> >       while(1) {
> > !         GET_CHAR(in,c,1,r->pool);
> >           if(c == str[p]) {
> > !             if((++p) == l)
> >                   return 0;
> >           }
> >           else {
> >               if (printing) {
> >                   for(x=0;x<p;x++) {
> > !                     rputc(str[x],r);
> >                   }
> > !                 rputc(c,r);
> >               }
> >               p=0;
> >           }
> >       }
> >   }
> >   
> >   /*
> >    * decodes a string containing html entities or numeric character references.
> > --- 113,209 ----
> >       }
> >   }
> >   
> > ! 
> > ! 
> > ! /* --------------------------- Parser functions --------------------------- */
> > ! 
> > ! #define OUTBUFSIZE 4096
> > ! /* PUT_CHAR and FLUSH_BUF currently only work within the scope of 
> > !  * find_string(); they are hacks to avoid calling rputc for each and
> > !  * every character output.  A common set of buffering calls for this 
> > !  * type of output SHOULD be implemented.
> > !  */
> > ! #define PUT_CHAR(c,r) \
> > !  { \
> > !    outbuf[outind++] = c; \
> > !    if (outind == OUTBUFSIZE) { FLUSH_BUF(r) }; \
> > !  } 
> > ! 
> > ! /* there SHOULD be some error checking on the return value of
> > !  * rwrite, however it is unclear what the API for rwrite returning
> > !  * errors is and little can really be done to help the error in 
> > !  * any case.
> > !  */
> > ! #define FLUSH_BUF(r) \
> > !  { \
> > !    rwrite(outbuf, outind, r); \
> > !    outind = 0; \
> > !  }
> > ! 
> > ! /*
> > !  * f: file handle being read from
> > !  * c: character to read into
> > !  * ret: return value to use if input fails
> > !  * r: current request_rec
> > !  *
> > !  * This macro is redefined after find_string() for historical reasons
> > !  * to avoid too many code changes.  This is one of the many things
> > !  * that should be fixed.
> > !  */
> > ! #define GET_CHAR(f,c,ret,r) \
> >    { \
> >      int i = getc(f); \
> > !    if(i == EOF) { /* either EOF or error -- needs error handling if latter */ \
> > !        if (ferror(f)) \
> > ! 	   fprintf(stderr, "encountered error in GET_CHAR macro, mod_include.\n"); \
> > !        FLUSH_BUF(r); \
> > !        pfclose(r->pool,f); \
> > !        return ret; \
> >      } \
> >      c = (char)i; \
> >    }
> >   
> >   int find_string(FILE *in,char *str, request_rec *r, int printing) {
> >       int x,l=strlen(str),p;
> > +     char outbuf[OUTBUFSIZE];
> > +     int outind = 0;
> >       char c;
> >   
> >       p=0;
> >       while(1) {
> > !         GET_CHAR(in,c,1,r);
> >           if(c == str[p]) {
> > !             if((++p) == l) {
> > ! 		FLUSH_BUF(r);
> >                   return 0;
> > + 	    }
> >           }
> >           else {
> >               if (printing) {
> >                   for(x=0;x<p;x++) {
> > !                     PUT_CHAR(str[x],r);
> >                   }
> > !                 PUT_CHAR(c,r);
> >               }
> >               p=0;
> >           }
> >       }
> >   }
> > + 
> > + #undef FLUSH_BUF
> > + #undef PUT_CHAR
> > + #undef GET_CHAR
> > + #define GET_CHAR(f,c,r,p) \
> > +  { \
> > +    int i = getc(f); \
> > +    if(i == EOF) { /* either EOF or error -- needs error handling if latter */ \
> > +        if (ferror(f)) \
> > + 	   fprintf(stderr, "encountered error in GET_CHAR macro, mod_include.\n"); \
> > +        pfclose(p,f); \
> > +        return r; \
> > +    } \
> > +    c = (char)i; \
> > +  }
> >   
> >   /*
> >    * decodes a string containing html entities or numeric character references.
> > 
> 
> 
> -- 
> ====================================================================
>       Jim Jagielski            |       jaguNET Access Services
>      jim@jaguNET.com           |       http://www.jaguNET.com/
>                   "Not the Craw... the CRAW!"
>