From 1930098552665575539f12744677bcc9acebb570 Mon Sep 17 00:00:00 2001
From: Rob Landley <rob@landley.net>
Date: Thu, 29 Sep 2022 00:04:43 -0500
Subject: [PATCH] Fix grep fast path logic to handle initial \ and '^$'
 pattern.

---
 tests/grep.test   |  5 +++++
 toys/posix/grep.c | 16 ++++++++++------
 2 files changed, 15 insertions(+), 6 deletions(-)

diff --git a/tests/grep.test b/tests/grep.test
index d150987b..67cfecaa 100755
--- a/tests/grep.test
+++ b/tests/grep.test
@@ -204,3 +204,8 @@ seq 1 100002 | base64 > testfile
 testing "speed" "timeout 5 grep -f testfile testfile 2>/dev/null | wc -l" \
   "10332\n" "" ""
 rm -f testfile
+
+# Fast path tests
+
+testcmd 'initial \' '\\.jar' 'bell.jar\n' '' 'bell.jar\n'
+testcmd '^$' '^\$' '\n' '' 'one\n\ntwo\n'
diff --git a/toys/posix/grep.c b/toys/posix/grep.c
index c7e0064d..efaabe82 100644
--- a/toys/posix/grep.c
+++ b/toys/posix/grep.c
@@ -177,7 +177,7 @@ static void do_grep(int fd, char *name)
             if (ss!=start) continue;
             pp++;
           }
-          for (ii = 1; pp[ii] && ss[ii]; ii++) {
+          for (ii = 0; pp[ii] && ss[ii]; ii++) {
             if (!FLAG(F)) {
               if (pp[ii]=='.') continue;
               if (pp[ii]=='\\' && pp[ii+1]) pp++;
@@ -418,13 +418,16 @@ static void parse_regex(void)
 
   // Convert to regex where appropriate
   for (last = &TT.e; *last;) {
-    if ('.'!=*(s = (*last)->arg) && !FLAG(F)) for (; *s; s++) {
+    // Can we use the fast path?
+    s = (*last)->arg;
+    if ('.'!=*s && !FLAG(F) && strcmp(s, "^$")) for (; *s; s++) {
       if (*s=='\\') {
         if (!s[1] || !strchr(special, *++s)) break;
         if (!FLAG(E) && *s=='(') break;
       } else if (*s>127 || strchr(special+4, *s)) break;
     }
 
+    // Add entry to fast path (literal-ish match) or slow path (regexec)
     if (!*s || FLAG(F)) last = &((*last)->next);
     else {
       struct reg *shoe;
@@ -439,18 +442,19 @@ static void parse_regex(void)
   }
   dlist_terminate(TT.reg);
 
-  // Sort fixed patterns into buckets by first character
+  // Sort fast path patterns into buckets by first character
   for (al = TT.e; al; al = new) {
     new = al->next;
     key = '^'==*al->arg;
-    if ('$'==al->arg[key] && !al->arg[key+1]) key = 0;
-    else key = al->arg[key];
+    if ('\\'==al->arg[key]) key++;
+    else if ('$'==al->arg[key] && !al->arg[key+1]) key++;
+    key = al->arg[key];
     if (FLAG(i)) key = toupper(key);
     al->next = TT.fixed[key];
     TT.fixed[key] = al;
   }
 
-  // Sort each fixed pattern set by length so first hit is longest match
+  // Sort each fast path pattern set by length so first hit is longest match
   if (TT.e) for (key = 0; key<256; key++) {
     if (!TT.fixed[key]) continue;
     for (len = 0, al = TT.fixed[key]; al; al = al->next) len++;
-- 
2.39.2