explaining Bogofilter simply

[David Relson]

On Sun, 25 Jan 2004 21:11:03 +0100
Andreas Pardeike wrote:

> On 2004-01-25, at 19.05, David Relson wrote:
> 
> > That's how to get the score of a message line.  The "-H", for
> > suppressing header tags, is crucial in this case.
> 
> But isn't the result below contrary to this statement?
> 
> > shell> echo "Buy Viagra for free" | bogofilter -H -vvv
> > X-Bogosity: Yes, tests=bogofilter, spamicity=0.970665,
> > version=0.16.0
> >                                      n    pgood     pbad      fw    
> >                                      U
> > "head:for"                         779  0.044110  0.042813  0.492547
> > -"head:Buy"                           0  0.000000  0.000000 
> > 0.677171 +"head:Viagra"                        0  0.000000  0.000000
> >  0.677171 +
> > "head:free"                          1  0.000000  0.000113  0.996804
> > + N_P_Q_S_s_x_md                       3  1.37e-02  9.55e-01 
> > 9.71e-01
> >                                         1.00e-02  6.77e-01  0.100
> 
> > shell> echo "\r\n\r\nBuy Viagra for free" | bogofilter -H -vvv
> > X-Bogosity: No, tests=bogofilter, spamicity=0.701919, version=0.16.0
> >                                      n    pgood     pbad      fw    
> >                                      U
> > "for"                             9457  0.719393  0.330351  0.314697
> > +"free"                            1017  0.071059  0.042016 
> > 0.371578 +"Viagra"                           586  0.011330  0.054700
> >  0.828410 +
> > "Buy"                              331  0.005940  0.031370  0.840791
> > + N_P_Q_S_s_x_md                       4  3.52e-01  7.56e-01 
> > 7.02e-01
> >                                         1.00e-02  6.77e-01  0.100
> 
> ???
> 
> Andreas Pardeike

Andreas,

You're right!  '-H' is broken in the bogofilter version you're using. 
The flag's sense (true/false meaning) is wrong in 0.16 releases before
0.16.2.  

Here's the patch:


Modified Files:
	bogoconfig.c bogolexer.c 
Log Message:
Have -H disable header_line_markup.

Index: bogoconfig.c
===================================================================
RCS file: /cvsroot/bogofilter/bogofilter/src/bogoconfig.c,v
retrieving revision 1.131
retrieving revision 1.132
diff -u -d -r1.131 -r1.132
--- bogoconfig.c	9 Jan 2004 12:42:21 -0000	1.131
+++ bogoconfig.c	9 Jan 2004 12:43:30 -0000	1.132
@@ -349,6 +349,9 @@
 		  "\t  -c file - read specified config file.\n"
 		  "\t  -C      - don't read standard config files.\n"
 		  "\t  -d path - specify directory for wordlists.\n"
+#ifndef	ENABLE_DEPRECATED_CODE
+		  "\t  -H      - disables header line tagging.\n"
+#endif   
 		  "\t  -k size - set BerkeleyDB cache size (MB).\n"
 #ifdef	ENABLE_DEPRECATED_CODE
 		  "\t  -W      - use combined wordlist for spam and ham
tokens.\n"
@@ -660,7 +663,7 @@
 #ifdef	ENABLE_DEPRECATED_CODE
 	    header_degen = true;
 #else
-	    header_line_markup = true;
+	    header_line_markup = false;
 #endif
 	    break;
 

Index: bogolexer.c
===================================================================
RCS file: /cvsroot/bogofilter/bogofilter/src/bogolexer.c,v
retrieving revision 1.55
retrieving revision 1.56
diff -u -d -r1.55 -r1.56
--- bogolexer.c	9 Jan 2004 12:42:21 -0000	1.55
+++ bogolexer.c	9 Jan 2004 12:43:30 -0000	1.56
@@ -81,6 +81,9 @@
 	    "\t-v\t- set verbosity level.\n"
 	    "\t-c file\t- read specified config file.\n"
 	    "\t-C\t- don't read standard config files.\n"
+#ifndef	ENABLE_DEPRECATED_CODE
+	    "\t-H\t- disables header line tagging.\n"
+#endif   
 	    "\t-I file\t- read message from file instead of stdin.\n"
 	    "\t-x list\t- set debug flags.\n"
 	    "\t-D\t- direct debug output to stdout.\n");
@@ -119,7 +122,7 @@
 }
 
 #ifndef	ENABLE_DEPRECATED_CODE
-#define	OPTIONS	":c:CDhI:npP:qvVx:X:m"
+#define	OPTIONS	":c:CDhHI:npP:qvVx:X:m"
 #else
 #define	OPTIONS	":c:CDhHI:npP:qvVx:X:m"
 #endif
@@ -169,9 +172,9 @@
 	    help();
 	    exit(EX_OK);
 
-#ifdef	ENABLE_DEPRECATED_CODE
+#ifndef	ENABLE_DEPRECATED_CODE
 	case 'H':
-	    header_degen = true;
+	    header_line_markup = false;
 	    break;
 #endif