From ec1a3e08354d27742e89fd8437a5a1d39153c53e Mon Sep 17 00:00:00 2001
From: Bruce Hill <bruce@bruce-hill.com>
Date: Thu, 20 May 2021 16:46:14 -0700
Subject: Fixes for char sets/ranges and cleanups

---
 grammars/bp.bp       |  2 +-
 grammars/builtins.bp |  3 +--
 pattern.c            | 17 +++++++++--------
 3 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/grammars/bp.bp b/grammars/bp.bp
index fc1e794..228f7d0 100644
--- a/grammars/bp.bp
+++ b/grammars/bp.bp
@@ -5,7 +5,7 @@
 # The grammar files provided with BP are not otherwise intended to be full
 # language grammars.
 
-Grammar: __ *(Def [__`;])%__ __ ($$ / (!)(..%\n$$ => "Could not parse this code"))
+Grammar: __ *(Def [__`;])%__ __ [(!)(+(./\n) => "Could not parse this code")]
 Def: @name=id __ `: __ (
       @definition=extended-pat
     / $$ (!)=>"No definition for rule"
diff --git a/grammars/builtins.bp b/grammars/builtins.bp
index cf642e2..888cc2c 100644
--- a/grammars/builtins.bp
+++ b/grammars/builtins.bp
@@ -17,8 +17,7 @@ braces: `{ ..%(\n/braces/string) `}
 parens: `( ..%(\n/parens/string) `)
 string: `" ..%string-escape `" / `' ..%string-escape `'
 string-escape: `\ (`x 2 Hex / 1-3 `0-7 / `u 1-4 Hex / .)
-left-id-edge: ^ / <(\x00-x7f!~(^^id-char)) / <((\xc0-xdf \x80-xbf)!~(^^id-char))
-    / <((\xe0-xef 2\x80-xbf)!~(^^id-char)) / <((\xf0-xf7 3\x80-xbf)!~(^^id-char))
+left-id-edge: !<id-char
 right-id-edge: !id-char
 id: left-id-edge !`0-9 !(keyword left-id-edge) +id-char
 id-char: `a-z,A-Z,_,0-9
diff --git a/pattern.c b/pattern.c
index 4e55928..d11fdcb 100644
--- a/pattern.c
+++ b/pattern.c
@@ -146,6 +146,7 @@ pat_t *chain_together(file_t *f, pat_t *first, pat_t *second)
         if (p->type == BP_UPTO) {
             p->args.multiple.first = second;
             p->min_matchlen = second->min_matchlen;
+            p->max_matchlen = -1;
             break;
         } else if (p->type == BP_CAPTURE) {
             p = p->args.capture.capture_pat;
@@ -250,10 +251,10 @@ static pat_t *_bp_simplepattern(file_t *f, const char *str)
                 const char *c1_loc = str;
                 str = next_char(f, c1_loc);
                 if (matchchar(&str, '-')) { // Range
-                    if (&str[-1] - c1_loc > 1 || next_char(f, str) > str+1)
-                        file_err(f, start, next_char(f, str), "Sorry, UTF-8 character ranges are not yet supported.");
-                    char c1 = *c1_loc;
-                    char c2 = *str;
+                    const char *c2_loc = str;
+                    if (next_char(f, c1_loc) > c1_loc+1 || next_char(f, c2_loc) > c2_loc+1)
+                        file_err(f, start, next_char(f, c2_loc), "Sorry, UTF-8 character ranges are not yet supported.");
+                    char c1 = *c1_loc, c2 = *c2_loc;
                     if (!c2 || c2 == '\n')
                         file_err(f, str, str, "There should be a character here to complete the character range.");
                     if (c1 > c2) { // Swap order
@@ -261,14 +262,14 @@ static pat_t *_bp_simplepattern(file_t *f, const char *str)
                         c1 = c2;
                         c2 = tmp;
                     }
-                    str = next_char(f, str);
-                    pat_t *pat = new_pat(f, start, str, 1, 1, BP_RANGE);
+                    str = next_char(f, c2_loc);
+                    pat_t *pat = new_pat(f, start == c1_loc - 1 ? start : c1_loc, str, 1, 1, BP_RANGE);
                     pat->args.range.low = (unsigned char)c1;
                     pat->args.range.high = (unsigned char)c2;
                     all = either_pat(f, all, pat);
                 } else {
-                    size_t len = (size_t)(str - start - 1);
-                    pat_t *pat = new_pat(f, start, str, len, (ssize_t)len, BP_STRING);
+                    size_t len = (size_t)(str - c1_loc);
+                    pat_t *pat = new_pat(f, c1_loc, str, len, (ssize_t)len, BP_STRING);
                     pat->args.string = c1_loc;
                     all = either_pat(f, all, pat);
                 }
-- 
cgit v1.2.3