patch 8.2.5028: syntax regexp matching can be slow
Problem: Syntax regexp matching can be slow.
Solution: Adjust the counters for checking the timeout to check about once
per msec. (closes #10487, closes #2712)
diff --git a/src/regexp_bt.c b/src/regexp_bt.c
index 698ff04..5bf0e0c 100644
--- a/src/regexp_bt.c
+++ b/src/regexp_bt.c
@@ -3271,8 +3271,10 @@
break;
}
#ifdef FEAT_RELTIME
- // Check for timeout once in a 100 times to avoid overhead.
- if (tm != NULL && ++tm_count == 100)
+ // Check for timeout once in 250 times to avoid excessive overhead from
+ // reading the clock. The value has been picked to check about once
+ // per msec on a modern CPU.
+ if (tm != NULL && ++tm_count == 250)
{
tm_count = 0;
if (profile_passed_limit(tm))
@@ -3313,7 +3315,7 @@
op = OP(scan);
// Check for character class with NL added.
if (!rex.reg_line_lbr && WITH_NL(op) && REG_MULTI
- && *rex.input == NUL && rex.lnum <= rex.reg_maxline)
+ && *rex.input == NUL && rex.lnum <= rex.reg_maxline)
{
reg_nextline();
}
@@ -4990,8 +4992,10 @@
else
++col;
#ifdef FEAT_RELTIME
- // Check for timeout once in a twenty times to avoid overhead.
- if (tm != NULL && ++tm_count == 20)
+ // Check for timeout once in 500 times to avoid excessive overhead
+ // from reading the clock. The value has been picked to check
+ // about once per msec on a modern CPU.
+ if (tm != NULL && ++tm_count == 500)
{
tm_count = 0;
if (profile_passed_limit(tm))