Ignore "misc" stem

IgKh · Nov 7, 2022 · f4832c9 · f4832c9
1 parent 40e4da6
commit f4832c9
Show file tree

Hide file tree

Showing 4 changed files with 14 additions and 1 deletion.
diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
@@ -9,7 +9,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        tag: [9.6-bullseye, 10-bullseye, 11-bullseye, 12-bullseye, 13-bullseye, 14-bullseye]
+        tag: [9.6-bullseye, 10-bullseye, 11-bullseye, 12-bullseye, 13-bullseye, 14-bullseye, 15-bullseye]
 
     runs-on: ubuntu-latest
     container:

diff --git a/dict_hspell.c b/dict_hspell.c
@@ -31,6 +31,8 @@ typedef struct DictHspell
 }
 DictHspell;
 
+static const unsigned char MISC_STEM[] = {0xF9, 0xE5, 0xF0, 0xE5, 0xFA, 0};
+
 static List* stem_list;
 
 PG_FUNCTION_INFO_V1(dhspell_init);
@@ -146,6 +148,10 @@ hspell_callback(const char* word, const char* baseword, int preflen, int prefspe
             continue;
         }
 
+        if (memcmp(stem_text, MISC_STEM, sizeof(MISC_STEM)) == 0) {
+            continue;
+        }
+
         if (!list_member_ptr(stem_list, stem_text)) {
             stem_list = lappend(stem_list, stem_text);
         }

diff --git a/expected/dict_hspell.out b/expected/dict_hspell.out
@@ -17,6 +17,12 @@ SELECT ts_lexize('hspell', 'איןמילהכזו');
  {איןמילהכזו}
 (1 row)
 
+SELECT ts_lexize('hspell', 'ישראל');
+ ts_lexize 
+-----------
+ {ישראל}
+(1 row)
+
 SELECT ts_lexize('hspell', 'English');
  ts_lexize 
 -----------

diff --git a/sql/dict_hspell.sql b/sql/dict_hspell.sql
@@ -3,6 +3,7 @@ CREATE EXTENSION pg_hspell;
 SELECT ts_lexize('hspell', 'הרכבת');
 SELECT ts_lexize('hspell', 'רק');
 SELECT ts_lexize('hspell', 'איןמילהכזו');
+SELECT ts_lexize('hspell', 'ישראל');
 
 SELECT ts_lexize('hspell', 'English');
 SELECT ts_lexize('hspell', 'Русский');