(no subject)

[bobf]

> The following commands:
>    awk '{print substr($0,1,8)}' bigfile >crud
>        and
>    awk '{print length($0,1,8)}' bigfile >crud
>
> Seem to lock up awk on Plan9.   If I launch either one there will be a
> flurry of hard disk activity then nothing.  A "ps" will reveal the
> process is "ready", but the crud file stops growing at 8647 for the
> first and 8192 for the second.   Rebooting doesn't change these numbers.

illegal utf values caused an infinite loop in the length() function.
thanks to rsc for finding this.

apply the following patch to /sys/src/cmd/awk/re.c:

301c301
< 	int i, j;
---
> 	int i;
304,309c304,305
< 	for (i = 0, end = s+n; *s && s < end; i++){
< 		j = mblen(s, n);
< 		if(j <= 0)
< 			j = 1;
< 		s += j;
< 	}
---
> 	for (i = 0, end = s+n; *s && s < end; i++)
> 		s += mblen(s, n);