From bebc7faef73b5943279e692706950c1d6cda7f42 Mon Sep 17 00:00:00 2001
From: Victor Chahuneau <vchahune@cs.cmu.edu>
Date: Thu, 17 May 2012 10:45:55 -0400
Subject: Fix JSON parser for unicode, empty feature vectors

---
 creg/json_feature_map_lexer.ll | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

(limited to 'creg/json_feature_map_lexer.ll')

diff --git a/creg/json_feature_map_lexer.ll b/creg/json_feature_map_lexer.ll
index cbb6d9a9..f9ce7977 100644
--- a/creg/json_feature_map_lexer.ll
+++ b/creg/json_feature_map_lexer.ll
@@ -77,6 +77,11 @@ UNESCAPED_CH [^\"\\\b\n\r\f\t]
 
 <JSON>{WS}*{LCB}{WS}*                    { BEGIN(PREVAL); }
 
+<JSON>{WS}*{LCB}{WS}*{RCB}\n*            {const SparseVector<float> x;
+                                         json_fmap_callback(instid, x, json_fmap_callback_extra);
+                                         curfeat = 0;
+                                         BEGIN(INITIAL);}
+
 <PREVAL>\"                               { BEGIN(STRING); spos=0; }
 
 <STRING>\"                               { featname[spos] = 0;
@@ -92,7 +97,8 @@ UNESCAPED_CH [^\"\\\b\n\r\f\t]
 <STRING>\\n                              { }
 <STRING>\\r                              { }
 <STRING>\\t                              { }
-<STRING>\\u{HEX_D}{HEX_D}{HEX_D}{HEX_D}  { abort();
+<STRING>\\u{HEX_D}{HEX_D}{HEX_D}{HEX_D}  { uint16_t hex = strtol(&yytext[2], NULL, 16);
+                                           spos += unicode_escape_to_utf8(hex, 0, &featname[spos++])-1;
                                          }
 
 <JSONVAL>{WS}*:{WS}*                     { BEGIN(DOUBLE); }
@@ -129,4 +135,3 @@ int main() {
   JSONFeatureMapLexer::ReadRules(&std::cin, cb, NULL);
 }
 #endif
-
-- 
cgit v1.2.3