Re: Search results titles

[andy hopkins <ahopkins@eyemg.com>]

- got it w/out touching glimpse. Changing the temp filenames in makenh
did it for me.

(sub getnewname)

   if ($file=~/\.pl/) {
      $ext=".html";
   } elsif($file=~/\/$/){
      $ext=".html";
   }

thanks for all your help.

andy


Golda Velez wrote:
> 
> > The extraction of the HTML title depends on the suffix and is
> > hardcoded into the glimpseindex executable. One possible solution for
> > you would be to add .pl as an HTML-like suffix. To do that you would
> > need tomodify glimpse.h in the glimpse source tree and change
> > these lines from
> >
> > #define EXTRACT_INFO_SUFFIX {".htm", ".html", ".shtm", ".shtml"}
> > #define NUM_EXTRACT_INFO_SUFFIX       4
> >
> > #define EXTRACT_INFO_SUFFIX {".htm", ".html", ".shtm", ".shtml", ".pl" }
> > #define NUM_EXTRACT_INFO_SUFFIX       5
> 
> That should work if the URL ends in .pl, but for Andy's example
> 
>         http://server.domain/cgi-bin/program.pl?foo=bar&bar=foo
> 
> I'm not sure if it will.  He might have to make a hack to the extension-test section in filetype.c, the simplest would be to just comment out the test, like so
> 
> diff -r1.6.2.1 filetype.c
> 149,151c149,151
> <               for (i=0; i<NUM_EXTRACT_INFO_SUFFIX; i++) {
> <                       if (!strcasecmp(&tempname[name_len - strlen(extract_info_suffix[i])], extract_info_suffix[i])) break;
> <               }
> ---
> > #             for (i=0; i<NUM_EXTRACT_INFO_SUFFIX; i++) {
> > #                     if (!strcasecmp(&tempname[name_len - strlen(extract_info_suffix[i])], extract_info_suffix[i])) break;
> > #             }
> 153c153
> <               if (i < NUM_EXTRACT_INFO_SUFFIX) {
> ---
> > #             if (i < NUM_EXTRACT_INFO_SUFFIX) {
> 155c155
> <               }
> ---
> > #             }
> 
> This would just force glimpseindex to treat all files as HTML regardless of extension.
> 
> If we do look at a hack to webglimpse instead, I think the easiest would be to manipulate the temporary filenames.  Since a script has to be retrieved anyway, makenh could just store it in a file named N.html instead of N.program.pl?foo=bar&bar=foo as it currently does.
> 
> --Golda