On 12/03/2013 03:08 PM, Jaromir Capik wrote:
Sure ... the guys are working really hard to make the tests absolutely bulletproof. We REALLY want to offer excellent quality products. And we hope it has a positive impact on all included software products :] The uninitialized 'pos' is also harmless, but the bug is there and I'll fix it by moving the increment to the body of the next condition.
Jaromir, As promised, here's the top part. My compliments to your QA folks... Jim
>From acc1341ae1b0bd7c319cb9592612910899a7d88c Mon Sep 17 00:00:00 2001 From: Jim Warner <james.warner@xxxxxxxxxxx> Date: Tue, 3 Dec 2013 15:00:00 -0600 Subject: [PATCH] top: do not forget the fscanf %s terminating null byte Reference(s): //www.freelists.org/post/procps/procpsng-339-defects-found Signed-off-by: Jim Warner <james.warner@xxxxxxxxxxx> --- top/top.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/top/top.c b/top/top.c index 4d1d912..bb15d01 100644 --- a/top/top.c +++ b/top/top.c @@ -3445,8 +3445,8 @@ static void configs_read (void) { WIN_t *w = &Winstk[i]; p = fmtmk(N_fmt(RC_bad_entry_fmt), i+1, Rc_name); - // note: "fieldscur=%__s" on next line should equal PFLAGSSIZ ! - if (2 != fscanf(fp, "%3s\tfieldscur=%80s\n" + // note: "fieldscur=%__s" on next line should equal (PFLAGSSIZ -1) ! + if (2 != fscanf(fp, "%3s\tfieldscur=%79s\n" , w->rc.winname, w->rc.fieldscur)) goto default_or_error; #if PFLAGSSIZ > 80 -- 1.8.3.2