On Tue, 2016-06-21 at 21:38 +0200, Cyril Brulebois wrote: > Hi, > > Juerg Haefliger <juerg.haefliger@hpe.com> (2016-06-10): > > Package: kernel-wedge > > Version: 2.94 > > Severity: normal > > Tags: patch > > > > Module inclusion with wildcards should honor the KW_CHECK_NONFATAL > > env variable and not error out if it's set and a wildcard > > directory is not found. > > > > This is in line with how missing non-optional modules are treated > > when KW_CHECK_NONFATAL is set. > > > > diff --git a/commands/preprocess b/commands/preprocess > > index 045903b..1a54632 100755 > > --- a/commands/preprocess > > +++ b/commands/preprocess > > @@ -35,7 +35,7 @@ sub expandwildcards { > > if (! -d "$moddir/$subdir") { > > if (-d "$moddir/kernel/$subdir") { > > $subdir = "kernel/$subdir"; > > - } elsif ($checkdir) { > > + } elsif ($checkdir && !length($ENV{KW_CHECK_NONFATAL})) > > { > > die "pattern $pattern refers to nonexistent > > subdirectory"; > > } else { > > return (); > > Thanks for the report+patch. > > Ben, since you've been pushing most code to kernel-wedge recently, do you concur? I agree that this error shouldn't be fatal in case that variable is set, but there should still be a warning. Additionally, the variable test should exists() as well as length() to avoid a Perl warning when the variable is not defined at all. (find- dups gets away with this because the embedded Perl script does not enable warnings.) Ben. -- Ben Hutchings [W]e found...that it wasn't as easy to get programs right as we had thought. ... I realized that a large part of my life from then on was going to be spent in finding mistakes in my own programs. - Maurice Wilkes, 1949
Attachment:
signature.asc
Description: This is a digitally signed message part