From 6d62de4f9aa16fb129513fa38c60158bfde36f16 Mon Sep 17 00:00:00 2001
From: =?utf8?q?Paul=20H=C3=A4nsch?= <paul@plutz.net>
Date: Tue, 19 Mar 2024 12:10:33 +0100
Subject: [PATCH] use tags (#positive and !negative) in search

---
 handlers/40_search.sh | 42 +++++++++++++++++++++++++++++++++++++++---
 themes/default.sh     |  2 +-
 2 files changed, 40 insertions(+), 4 deletions(-)

diff --git a/handlers/40_search.sh b/handlers/40_search.sh
index 1e33019..57117e3 100644
--- a/handlers/40_search.sh
+++ b/handlers/40_search.sh
@@ -21,15 +21,49 @@
 . "$_EXEC/cgilite/db23.sh"
 
 I="$_DATA/index"
+tags="$( GET q | awk '
+  BEGIN { # Field separator FS should include punctuation, including Unicode Block U+2000 - U+206F
+          if ( length("¡") == 1 )  # Utf-8 aware AWK
+          FS = "([] \\t\\n\\r!\"'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '[\342\200\200-\342\201\257]')"')+";
+          else                     # UTF-8 Hack
+          FS = "([] \\t\\n\\r!\"'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '\342\200[\200-\277]|\342\201[\201-\257]')"')+";
+          fi
+        }
+  { for (n = 1; n <= NF; n++) if ($n ~ /#[[:alnum:]_]+/) {
+      sub(/^#/,"",$n)
+      printf "%s	", toupper($n);
+    }
+  }
+')"
+
+ntags="$( GET q | awk '
+  BEGIN { # Field separator FS should include punctuation, including Unicode Block U+2000 - U+206F
+          if ( length("¡") == 1 )  # Utf-8 aware AWK
+          FS = "([] \\t\\n\\r\"#'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '[\342\200\200-\342\201\257]')"')+";
+          else                     # UTF-8 Hack
+          FS = "([] \\t\\n\\r\"#'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '\342\200[\200-\277]|\342\201[\201-\257]')"')+";
+          fi
+        }
+  { for (n = 1; n <= NF; n++) if ($n ~ /![[:alnum:]_]+/) {
+      sub(/^!/,"",$n)
+      printf "%s	", toupper($n);
+    }
+  }
+')"
+
 words="$( GET q | awk '
   BEGIN { # Field separator FS should include punctuation, including Unicode Block U+2000 - U+206F
           if ( length("¡") == 1 )  # Utf-8 aware AWK
-          FS = "([] \\t\\n\\r!\"#'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '[\342\200\200-\342\201\257]')"')+";
+          FS = "([] \\t\\n\\r\"#'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '[\342\200\200-\342\201\257]')"')+";
           else                     # UTF-8 Hack
-          FS = "([] \\t\\n\\r!\"#'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '\342\200[\200-\277]|\342\201[\201-\257]')"')+";
+          FS = "([] \\t\\n\\r\"#'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '\342\200[\200-\277]|\342\201[\201-\257]')"')+";
           fi
         }
-	{ for (n = 1; n <= NF; n++) printf "%s	", tolower($n); }
+  { for (n = 1; n <= NF; n++) if ($n !~ /![[:alnum:]_]+/) {
+      sub(/!/," ",$n)
+      printf "%s	", tolower($n);
+    }
+  }
 ')"
 
 searchteaser() {
@@ -89,6 +123,8 @@ done \
     [ "${page%/:*/}" = "${page%/:${LANGUAGE}/}" ] || continue
   fi
   acl_read "$page" || continue
+  has_tags "$page" $tags || continue
+  has_tag "$page" $ntags && continue
   printf '%s	%s\n' "$doc" "$(searchteaser "$(mdfile "$page")" $words)"
 done \
 | theme_search "${words%	}"
diff --git a/themes/default.sh b/themes/default.sh
index 033dd7f..0f45c59 100755
--- a/themes/default.sh
+++ b/themes/default.sh
@@ -142,7 +142,7 @@ theme_search(){
 	<article>
 	  <h1>$([ "$words" ] && _ "Search results" || _ "Search" )</h1>
 	  <form class="search" method="GET">
-	    <input type="search" name="q" value="$(HTML $words)"><button class="search" type="submit">$(_ Search)</button>
+	    <input type="search" name="q" value="$(GET q |HTML)"><button class="search" type="submit">$(_ Search)</button>
 	  </form>
 	  <ol class="searchresults">
 	    $(while read -r p t; do
-- 
2.39.2