Fix lack of ->name and ->location in DO REPEAT's getl_interface.

[pspp-builds.git] / src / language / lexer / ChangeLog

diff --git a/src/language/lexer/ChangeLog b/src/language/lexer/ChangeLog
index ebbf25441ca2b99b9bd2e26836f7eb66a66e87a5..db33299a4cd473398effd2f2ac0ac9d5415ac79b 100644 (file)
--- a/src/language/lexer/ChangeLog
+++ b/src/language/lexer/ChangeLog
@@ -1,3 +1,30 @@
+Sat Dec  2 21:19:50 2006  Ben Pfaff  <blp@gnu.org>
+
+       General clean-up.
+       
+       * lexer.c: (lex_token_name) Don't use a static buffer or, rather,
+       use a separate static buffer for each possible answer.
+       (lex_token_representation) Now use lex_token_name as building
+       block.  Previously this broke lex_force_match because it uses
+       lex_token_name in a call to lex_error, which in turn uses
+       lex_token_representation.
+       (lex_force_match_id) Use lex_match_id as building block, to
+       simplify.
+       (parse_id) Rewrite to work with modified lex_id_to_token and
+       lex_id_get_length.  The computation of rest_of_line is a bit of an
+       abomination but it will get fixed later.
+       
+Sat Dec  2 20:16:50 2006  Ben Pfaff  <blp@gnu.org>
+
+       * lexer.c (struct lexer): Change function signature for
+       `read_line' to take an "enum getl_syntax *" instead of "bool *".
+       (lex_create) Ditto, for argument.
+       (lex_preprocess_line) New function.
+       (lex_get_line_raw) New arg, to allow caller to obtain getl_syntax
+       of the line read.
+       (lex_get_line) Use lex_get_line_raw and lex_preprocess_line to
+       simplify.
+
 Sun Nov 19 09:20:42 2006  Ben Pfaff  <blp@gnu.org>
 
        * range-parser.c (parse_num_range): Because data_in takes an enum