Subject: [PATCH] scripts: kernel-doc: Check existence of FILE arg

Currently, when no FILE argument is given following switches such
as -man, -rst, and -none, kernel-doc ends up in the error of (long
msg from perl folded):

    Use of uninitialized value $ARGV[0] in pattern match (m//)
    at ./scripts/kernel-doc line 438.

Fix this by adding an existence check at the bottom of the while
loop parsing command switches and call usage() if there remains
no argument.

While at it, fix inconsistent indent by spaces in the previous
block.

Signed-off-by: Akira Yokosawa <akiyks@gmail.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: linux-doc@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
---
Hi Jon,

I noticed this minor issue while reviewing Tomasz's patch series.
This patch is relative to docs-next as of 2022.02.22.

        Thanks, Akira
--
 scripts/kernel-doc | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/scripts/kernel-doc b/scripts/kernel-doc
index 3106b7536b89..faefe2977f0e 100755
--- a/scripts/kernel-doc
+++ b/scripts/kernel-doc
@@ -494,7 +494,11 @@ while ($ARGV[0] =~ m/^--?(.*)/) {
 	}
     } else {
 	# Unknown argument
-        usage();
+	usage();
+    }
+    if ($#ARGV < 0){
+	print "No FILE!\n";
+	usage();
     }
 }
 

base-commit: b62ef3a1cca0553613adce16515f3640400725b4
-- 
2.17.1

On Wed, 23 Feb 2022 15:07:30 +0900
Akira Yokosawa <akiyks@gmail.com> wrote:

> Currently, when no FILE argument is given following switches such
> as -man, -rst, and -none, kernel-doc ends up in the error of (long
> msg from perl folded):
> 
>     Use of uninitialized value $ARGV[0] in pattern match (m//)
>     at ./scripts/kernel-doc line 438.

It's a warning not an error.

On Wed, 23 Feb 2022 14:47:13 +0100,
Tomasz Warniełło wrote:
> On Wed, 23 Feb 2022 15:07:30 +0900
> Akira Yokosawa <akiyks@gmail.com> wrote:
> 
>> Currently, when no FILE argument is given following switches such
>> as -man, -rst, and -none, kernel-doc ends up in the error of (long
>> msg from perl folded):
>>
>>     Use of uninitialized value $ARGV[0] in pattern match (m//)
>>     at ./scripts/kernel-doc line 438.
> 
> It's a warning not an error.

Good catch!

Will amend in v2 with proper attribution.

        Thanks Akira

On Wed, 23 Feb 2022 22:59:01 +0900
Akira Yokosawa <akiyks@gmail.com> wrote:

> On Wed, 23 Feb 2022 14:47:13 +0100,
> Tomasz Warniełło wrote:
> > On Wed, 23 Feb 2022 15:07:30 +0900
> > Akira Yokosawa <akiyks@gmail.com> wrote:
> >   
> >> Currently, when no FILE argument is given following switches such
> >> as -man, -rst, and -none, kernel-doc ends up in the error of (long
> >> msg from perl folded):
> >>
> >>     Use of uninitialized value $ARGV[0] in pattern match (m//)
> >>     at ./scripts/kernel-doc line 438.  
> > 
> > It's a warning not an error.  
> 
> Good catch!
> 
> Will amend in v2 with proper attribution.
> 
>         Thanks Akira

I'd include this in a series regulating all the CLI syntax - all those
mutual exclusions and arg parsing error handling, etc. At the moment they
exist only in the documentation.

But I'm not sure this is the right moment to do this. Have you seen
my RFC thread? I'd wait until the water stabilises and see the result.
This script may be phased out - how soon, I don't know. Unless you want
to do this purely for sport.

Tomasz

Tomasz Warniełło <tomasz.warniello@gmail.com> writes:

> But I'm not sure this is the right moment to do this. Have you seen
> my RFC thread? I'd wait until the water stabilises and see the result.
> This script may be phased out - how soon, I don't know. Unless you want
> to do this purely for sport.

Unless somebody steps up to do the work, I don't see that phasing out
happening any time really soon.  I know I'll not be able to find the
time to take on a project like that...so I certainly wouldn't be turning
away improvements to the current script on that basis.

Thanks,

jon