Russ Allbery wrote...
> Christoph Biedl <cbiedl@gmx.de> writes:
> 
> > One thing I miss after a first check of the manpages: I'd like to store
> > these "flagged" articles (i.e. rejected by cleanfeed but _not_ dropped,
> > is there a better word for this?) in a separate buffer. This buffer may
> > be rather small, e.g. rotate in 24 hours. That's enough time to retrieve
> > the article in case I'd like to know why cleanfeed didn't like it but it
> > wouldn't consume to much disk space/history/etc. Otherwise "Binary in
> > non-binary group" articles would have an impact on the retenition time
> > of regular articles.
> 
> > Is it possible to use this flag as a storage criteria? A very short
> > glance into the sources suggests this should be not too difficult.
> 
> It's not currently possible to do this.  It's certainly something that
> would be desirable to add, although I'm not sure the best general way to
> do it. 

See the attached patch 
applies to INN 2.4.2, "works for me".
applies to CURRENT after a small modification, but not tested yet.

The patch adds a "filtered:" clause to the storage.conf definitions that 
matches if the article was rejected by the Perl or Python filter (or TCL
which I've ignored for known reasons).

Two comments when using "dontfilterrejected" after the patch applied:

1. Regarding storage.conf: Since the default value for "filtered:" is
   "false", not "don't care", there must be at least two statements in
   the storage.conf. One with "filtered: true", one with the default
   "filtered: false" (may be omitted). Else innd will throttle due to a
   missing storage class. 
2. No messages about rejections to syslog were written. This has been
   changed since I'd still like to see how the perl filter is working.
   Since "rejecting" is a wrong term I've changed that to "filtering".
   Therefore innreport needed a tweak, too.

Enjoy,
    Christoph


-- Attached file included as plaintext by Ecartis --

diff -bur testnews.OLD/doc/man/storage.conf.5 testnews/doc/man/storage.conf.5
--- testnews.OLD/doc/man/storage.conf.5	2004-12-22 05:21:19.000000000 +0100
+++ testnews/doc/man/storage.conf.5	2005-08-07 01:19:47.000000000 +0200
@@ -36,6 +36,7 @@
 	expires: <mintime>[,<maxtime>]
 	options: <options>
 	exactmatch: <bool>
+	filtered: <bool>
 }
 
 .fi
@@ -122,6 +123,14 @@
 described above.)  This is a boolan value and ``true'', ``yes''
 and ``on'' are usable to enable this key.  The case of these values is not
 significant.  The default is false.
+.TP
+.B filtered
+If this key is set to ``true'', the article must have been rejected by the
+Perl/Python filter.  This also requires that ``dontfilterrejected'' is set 
+to ``true'' in
+.IR inn.conf .
+This is a boolan value and ``true'', ``yes'' and ``on'' are usable to enable
+this key.  The case of these values is not significant.  The default is false.
 .PP
 If an article matches all of the constraints of an entry, it is stored via
 that storage method and is associated with that <storage_class>.  This
diff -bur testnews.OLD/include/storage.h testnews/include/storage.h
--- testnews.OLD/include/storage.h	2004-12-22 05:21:19.000000000 +0100
+++ testnews/include/storage.h	2005-08-06 20:47:15.000000000 +0200
@@ -44,6 +44,7 @@
   void           *private;   /* A pointer to method specific data */
   time_t         arrived;    /* The time when the article arrived */
   time_t         expires;    /* The time when the article will be expired */
+  bool           filtered;   /* Article was marked by the filter */
   char           *groups;    /* Where Newsgroups header starts */
   int            groupslen;  /* Length of Newsgroups header */
   TOKEN          *token;     /* A pointer to the article's TOKEN */
diff -bur testnews.OLD/innd/art.c testnews/innd/art.c
--- testnews.OLD/innd/art.c	2005-01-29 02:05:24.000000000 +0100
+++ testnews/innd/art.c	2005-08-07 10:53:25.000000000 +0200
@@ -518,6 +518,7 @@
   arth.arrived = (time_t)0;
   arth.token = (TOKEN *)NULL;
   arth.expires = data->Expires;
+  arth.filtered = data->Filtered;
   if (innconf->storeonxref) {
     arth.groups = data->Replic;
     arth.groupslen = data->ReplicLength;
@@ -1886,6 +1887,10 @@
   TMRstop(TMR_PYTHON);
   if (filterrc != NULL) {
     if (innconf->dontrejectfiltered) {
+      snprintf(cp->Error, sizeof(cp->Error), "%d %.200s", NNTP_REJECTIT_VAL,
+               filterrc);
+      syslog(L_NOTICE, "filtering[python] %s %s", HDR(HDR__MESSAGE_ID),
+             cp->Error);
       Filtered = true;
     } else {
       snprintf(cp->Error, sizeof(cp->Error), "%d %.200s", NNTP_REJECTIT_VAL,
@@ -1913,6 +1918,10 @@
   if (filterrc) {
     if (innconf->dontrejectfiltered) {
       Filtered = true;
+      snprintf(cp->Error, sizeof(cp->Error), "%d %.200s", NNTP_REJECTIT_VAL,
+	       filterrc);
+      syslog(L_NOTICE, "filtering[perl] %s %s", HDR(HDR__MESSAGE_ID),
+	     cp->Error);
     } else {
       snprintf(cp->Error, sizeof(cp->Error), "%d %.200s", NNTP_REJECTIT_VAL,
                filterrc);
@@ -2295,6 +2304,7 @@
   for (i = 0; (ngp = GroupPointers[i]) != NULL; i++)
     ngp->PostCount = 0;
 
+  data->Filtered = Filtered;
   token = ARTstore(cp);
   /* change trailing '\r\n' to '\0\n' of all system header */
   for (i = 0 ; i < MAX_ARTHEADER ; i++) {
diff -bur testnews.OLD/innd/innd.h testnews/innd/innd.h
--- testnews.OLD/innd/innd.h	2004-12-22 05:21:19.000000000 +0100
+++ testnews/innd/innd.h	2005-08-06 20:51:22.000000000 +0200
@@ -197,6 +197,7 @@
   int		  HeaderLines;		/* number of header lines */
   long		  BytesValue;		/* size of stored article, "\r\n" is
 					   counted as 1 byte */
+  bool		  Filtered;		/* Marked by filter */
   char		  Bytes[16];		/* generated Bytes header */
   int		  BytesLength;		/* generated Bytes header length */
   char	      *   BytesHeader;		/* where Bytes header begins in
diff -bur testnews.OLD/scripts/innreport_inn.pm testnews/scripts/innreport_inn.pm
--- testnews.OLD/scripts/innreport_inn.pm	2004-12-22 05:21:19.000000000 +0100
+++ testnews/scripts/innreport_inn.pm	2005-08-07 00:27:02.000000000 +0200
@@ -244,8 +244,8 @@
       # SERVER cancelled +id
       return 1 if $left =~ /^SERVER cancelled /o;
     }
-    # rejecting[perl]
-    if ($left =~ /^rejecting\[perl\] <[^>]+> \d+ (.*)/o) {
+    # rejecting[perl]/filtering[perl]
+    if ($left =~ /^(?:rejecting|filtering)\[perl\] <[^>]+> \d+ (.*)/o) {
       $innd_filter_perl{$1}++;
       return 1;
     }
diff -bur testnews.OLD/storage/interface.c testnews/storage/interface.c
--- testnews.OLD/storage/interface.c	2004-12-22 05:21:19.000000000 +0100
+++ testnews/storage/interface.c	2005-08-07 01:11:43.000000000 +0200
@@ -334,6 +334,7 @@
 #define SMexpire  14
 #define SMoptions 15
 #define SMexactmatch 16
+#define SMfiltered 17
 
 static CONFTOKEN smtoks[] = {
   { SMlbrace,	"{" },
@@ -345,6 +346,7 @@
   { SMexpire,	"expires:" },
   { SMoptions,	"options:" },
   { SMexactmatch,	"exactmatch:" },
+  { SMfiltered,	"filtered:" },
   { 0, 0 }
 };
 
@@ -371,6 +373,7 @@
     char		*options = 0;
     int			inbrace;
     bool		exactmatch = false;
+    bool		filtered = false;
 
     /* if innconf isn't already read in, do so. */
     if (innconf == NULL) {
@@ -423,6 +426,7 @@
 	    minexpire = 0;
 	    maxexpire = 0;
 	    exactmatch = false;
+	    filtered = false;
 
 	} else {
 	    type = tok->type;
@@ -470,6 +474,12 @@
                         || strcasecmp(p, "on") == 0)
 			exactmatch = true;
 		    break;
+		  case SMfiltered:
+		    if (strcasecmp(p, "true") == 0
+                        || strcasecmp(p, "yes") == 0
+                        || strcasecmp(p, "on") == 0)
+			filtered = true;
+		    break;
 		  default:
 		    SMseterror(SMERR_CONFIG, "Unknown keyword in method declaration");
 		    syslog(L_ERROR, "SM Unknown keyword in method declaration, line %d: %s", f->lineno, tok->name);
@@ -512,6 +522,7 @@
 	    sub->minexpire = minexpire;
 	    sub->maxexpire = maxexpire;
 	    sub->exactmatch = exactmatch;
+	    sub->filtered = filtered;
 
 	    free(method);
 	    method = 0;
@@ -694,7 +705,8 @@
 	    (!sub->minexpire || article.expires >= sub->minexpire) &&
 	    (!sub->maxexpire || (article.expires <= sub->maxexpire)) &&
 	    MatchGroups(article.groups, article.groupslen, sub->pattern,
-                        sub->exactmatch)) {
+                        sub->exactmatch) &&
+	    (sub->filtered == article.filtered)) {
 	    if (InitMethod(typetoindex[sub->type]))
 		return sub;
 	}
diff -bur testnews.OLD/storage/interface.h testnews/storage/interface.h
--- testnews.OLD/storage/interface.h	2004-12-22 05:21:19.000000000 +0100
+++ testnews/storage/interface.h	2005-08-06 20:46:17.000000000 +0200
@@ -45,6 +45,7 @@
 				        method */
     bool		exactmatch;  /* all newsgroups to which article belongs
 				        should match the patterns */
+    bool		filtered;    /* article was marked by the filter */
     struct __S_SUB__   *next;
 } STORAGE_SUB;