diff options
| author | alecpl <alec@alec.pl> | 2009-10-19 07:27:16 +0000 |
|---|---|---|
| committer | alecpl <alec@alec.pl> | 2009-10-19 07:27:16 +0000 |
| commit | 16dd372574ca0e03481b1ca65c65556eddb40be1 (patch) | |
| tree | f5f34243ee2990de27cb5e67641de43856c01e1b /program/steps | |
| parent | b6673c4e416affcfe644473cb8fc6009d04ee971 (diff) | |
- pspell checker: split text by [dot]+[non-word] instead of [dot] and don't check numeric strings
Diffstat (limited to 'program/steps')
| -rw-r--r-- | program/steps/mail/spell_pspell.inc | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/program/steps/mail/spell_pspell.inc b/program/steps/mail/spell_pspell.inc index 717b92ee5..0abdc9ea1 100644 --- a/program/steps/mail/spell_pspell.inc +++ b/program/steps/mail/spell_pspell.inc @@ -43,7 +43,7 @@ $text = substr($data, $left+6, $right-($left+6)); $text = html_entity_decode($text, ENT_QUOTES, RCMAIL_CHARSET); // tokenize -$words = preg_split('/[ !"#$%&()*+\\,-.\/\n:;<=>?@\[\]^_{|}]+/', $text, NULL, PREG_SPLIT_NO_EMPTY | PREG_SPLIT_OFFSET_CAPTURE ); +$words = preg_split('/[ !"#$%&()*+\\,\/\n:;<=>?@\[\]^_{|}-]+|\.[^\w]/', $text, NULL, PREG_SPLIT_NO_EMPTY | PREG_SPLIT_OFFSET_CAPTURE ); // init spellchecker $plink = pspell_new(get_input_value('lang', RCUBE_INPUT_GET), null, null, RCMAIL_CHARSET, PSPELL_FAST); @@ -56,7 +56,8 @@ foreach ($words as $w) { $word = trim($w[0]); $pos = $w[1] - $diff; $len = mb_strlen($word); - if ($word && $plink && !pspell_check($plink, $word)) { + if ($word && $plink && preg_match('/[^0-9\.]/', $word) + && !pspell_check($plink, $word)) { $suggestions = pspell_suggest($plink, $word); if (sizeof($suggestions)>10) $suggestions = array_slice($suggestions, 0, MAX_SUGGESTIONS); |