On Tuesday 29 September 2009 01:33, Ian Wienand wrote:
> Hi,
>
> Although it sounds a little ridiculous for embedded systems, with the
> advent of big-disk NAS systems etc it's probably quite easy to
> overflow the 9 characters for the file size in ls. We noticed when
> people tried ls against tb sized files on VMFS partitions.
It is till is readable, though, right? It just does not line up anymore,
like this:
-rw-r--r-- 1 2049807 201 1073741824 Sep 28 16:14 big
-rwxr-xr-x 1 2049807 201 855368 Sep 28 16:13 busybox
-rwxr-xr-x 1 2049807 201 999754 Sep 28 16:22 busybox_old
> This implements a flexi-size for the size field. Although other
> fields could also be made flexi, it seems they're less likely to
> expand past their limits.
It needs to be at least conditional on ENABLE_DESKTOP,
or should get a new ENABLE_xxxx.> if (all_fmt & STYLE_LONG) {
> ncols = 1;
> +
> + /* find the longest file size */
> + for (i = 0; i < nfiles; i++) {
> + int w;
> + if (all_fmt & LS_DISP_HR) {
> + w =
> bb_mbstrlen(make_human_readable_str(dn[i]->dstat.st_size, 1, 0));
In this case, it never overflows 9 chars. Right?
Also, it is in ASCII, you can use just strlen().
> + } else {
> + off_t s = dn[i]->dstat.st_size;
> + w = 2; // 1 + 1 for padding
> + while (s > 10) {
> + w++; s /= 10;
> + }
> + }
> +
> + if (w > size_width) {
> + size_width = w;
> + }
> + }
> +
--
vda
_______________________________________________
busybox mailing list
[email protected]
http://lists.busybox.net/mailman/listinfo/busybox
