> Sent: Sunday, November 16, 2014 at 1:27 AM > From: "Philip Guenther" <[email protected]> > To: "Kamil Rytarowski" <[email protected]> > Cc: [email protected] > Subject: Re: mg(1): Second set of patches > > > On Sat, 15 Nov 2014, Kamil Rytarowski wrote: > > I'm attaching two enhancements against mg(1): > > > > 0001-Include-limits.h-for-INT_MIN-and-INT_MAX.patch > > 0002-Comparison-of-array-bp-b_fname-not-equal-to-a-null-p.patch > > The latter patch was committed *months* ago. Please verify you're working > from a current tree! (That probably also explains a rejected chunk in one > of your other diffs that I had to handle manually...)
I have synced my sources with current OpenBSD CVS. > > > For the former, I think the diff I correct, but doesn't go far enough. > The reason <limits.h> isn't needed on OpenBSD is that our <sys/param.h> is > pulling in the requisite #defines...but mg doesn't have a good reason for > needing <sys/param.h>! So here's a slightly bigger diff that eliminates > that by switching from MAXPATHLEN to PATH_MAX and expanding the one use of > MAX(). > > oks? Good idea. Please go for it, this MAX() was problematic at SunOS. Best regards,
