Re: [PATCH] fix unterminated sed cmd in export_report.pl

Re: [PATCH] fix unterminated sed cmd in export_report.pl

[Jiri Kosina]

On Tue, 10 May 2011, jim.cromie@gmail.com wrote:

> From: Jim Cromie <jim.cromie@gmail.com>
> 
> fix following err by escaping end-of-line $ in regex.
> perl /home/jimc/projects/lx/linux-2.6/scripts/export_report.pl > /dev/null
> sed: -e expression #1, char 5: unterminated `s' command
> sh: .mod.c/: not found
> 
> without this fix, SECTION 2 of report is empty.
> 
> Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
> ---
>  scripts/export_report.pl |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/scripts/export_report.pl b/scripts/export_report.pl
> index 04dce7c..91fa5a2 100644
> --- a/scripts/export_report.pl
> +++ b/scripts/export_report.pl
> @@ -50,7 +50,7 @@ sub usage {
>  
>  sub collectcfiles {
>      my @file
> -	= `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko$/.mod.c/`;
> +	= `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko\$/.mod.c/`;
>      chomp @file;
>      return @file;
>  }

Thanks for the fix, Jim.

Adding Michal and linux-kbuild@, this should rather go in through his 
tree.

-- 
Jiri Kosina
SUSE Labs

Re: [PATCH] fix unterminated sed cmd in export_report.pl

[Arnaud Lacombe]

Hi,

On Wed, May 11, 2011 at 5:27 AM, Jiri Kosina <jkosina@suse.cz> wrote:
> On Tue, 10 May 2011, jim.cromie@gmail.com wrote:
>
>> From: Jim Cromie <jim.cromie@gmail.com>
>>
>> fix following err by escaping end-of-line $ in regex.
>> perl /home/jimc/projects/lx/linux-2.6/scripts/export_report.pl > /dev/null
>> sed: -e expression #1, char 5: unterminated `s' command
>> sh: .mod.c/: not found
>>
>> without this fix, SECTION 2 of report is empty.
>>
>> Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
>> ---
>>  scripts/export_report.pl |    2 +-
>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/scripts/export_report.pl b/scripts/export_report.pl
>> index 04dce7c..91fa5a2 100644
>> --- a/scripts/export_report.pl
>> +++ b/scripts/export_report.pl
>> @@ -50,7 +50,7 @@ sub usage {
>>
>>  sub collectcfiles {
>>      my @file
>> -     = `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko$/.mod.c/`;
>> +     = `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko\$/.mod.c/`;
This is bogus. the first \ is still eaten by the shell. All what
sed(1) will see is the '.'.

Btw, that's an awful lot of command to do a simple task. grep(1) and
cat(1) can be done within sed(1), so I guess the patch should looks
like:

-     = `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko$/.mod.c/`;
+     = `sed '/\.ko$/!d; s/\.ko$/.mod.c/' .tmp_versions/*.mod`;

[not tested]

 - Arnaud

>>      chomp @file;
>>      return @file;
>>  }
>
> Thanks for the fix, Jim.
>
> Adding Michal and linux-kbuild@, this should rather go in through his
> tree.
>
> --
> Jiri Kosina
> SUSE Labs
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kbuild" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH] fix unterminated sed cmd in export_report.pl

[Arnaud Lacombe]

Hi

On Wed, May 11, 2011 at 5:16 PM, Arnaud Lacombe <lacombar@gmail.com> wrote:
> -     = `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko$/.mod.c/`;
> +     = `sed '/\.ko$/!d; s/\.ko$/.mod.c/' .tmp_versions/*.mod`;
>
Actually, my perl-fuu is not what it used to be, that one should be better:

+     = `sed '/\\.ko\$/!d; s/\\.ko\$/.mod.c/' .tmp_versions/*.mod`;

The backslash before the '.' should be escaped as well to be passed to
sed(1). I'll give it a try to confirm.

 - Arnaud

Re: [PATCH] fix unterminated sed cmd in export_report.pl

[Jim Cromie]

On Wed, May 11, 2011 at 3:26 PM, Arnaud Lacombe <lacombar@gmail.com> wrote:
> Hi
>
> On Wed, May 11, 2011 at 5:16 PM, Arnaud Lacombe <lacombar@gmail.com> wrote:
>> -     = `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko$/.mod.c/`;
>> +     = `sed '/\.ko$/!d; s/\.ko$/.mod.c/' .tmp_versions/*.mod`;
>>
> Actually, my perl-fuu is not what it used to be, that one should be better:
>
> +     = `sed '/\\.ko\$/!d; s/\\.ko\$/.mod.c/' .tmp_versions/*.mod`;
>
> The backslash before the '.' should be escaped as well to be passed to
> sed(1). I'll give it a try to confirm.
>
>  - Arnaud
>


I agree w your complicated-construct sentiment re the cmd-pipeline.

this used to be done with a grep,
undone by


commit 91416cfdf98bdbc828fd3e5ca7208beba5979d63
Author: Stephen Hemminger <shemminger@vyatta.com>
Date:   Mon Feb 22 15:17:22 2010 -0800

    export_report: fix perl warnings

    Use local file handles, use three argument open.
    Don't modify arguments in perl grep (use sed instead)

    Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
    Acked-by: WANG Cong <amwang@redhat.com>
    Cc: Michal Marek <mmarek@suse.cz>
    Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
    Signed-off-by: Michal Marek <mmarek@suse.cz>

diff --git a/scripts/export_report.pl b/scripts/export_report.pl
index 705b5ba..04dce7c 100644
--- a/scripts/export_report.pl
+++ b/scripts/export_report.pl
@@ -49,10 +49,10 @@ sub usage {
 }

 sub collectcfiles {
-        my @file = `cat .tmp_versions/*.mod | grep '.*\.ko\$'`;
-        @file = grep {s/\.ko/.mod.c/} @file;
-       chomp @file;
-        return @file;
+    my @file
+       = `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko$/.mod.c/`;
+    chomp @file;
+    return @file;
 }


the backslash escape noise in the subshell pipeline can be avoided
by returning to the grep, with less process spawning too
(not that this is executed often)
I'll try that.

Re: [PATCH] fix unterminated sed cmd in export_report.pl

[Stephen Hemminger]

On Wed, 11 May 2011 16:26:23 -0600
Jim Cromie <jim.cromie@gmail.com> wrote:

> On Wed, May 11, 2011 at 3:26 PM, Arnaud Lacombe <lacombar@gmail.com> wrote:
> > Hi
> >
> > On Wed, May 11, 2011 at 5:16 PM, Arnaud Lacombe <lacombar@gmail.com> wrote:
> >> -     = `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko$/.mod.c/`;
> >> +     = `sed '/\.ko$/!d; s/\.ko$/.mod.c/' .tmp_versions/*.mod`;
> >>
> > Actually, my perl-fuu is not what it used to be, that one should be better:
> >
> > +     = `sed '/\\.ko\$/!d; s/\\.ko\$/.mod.c/' .tmp_versions/*.mod`;
> >
> > The backslash before the '.' should be escaped as well to be passed to
> > sed(1). I'll give it a try to confirm.
> >
> >  - Arnaud
> >
> 
> 
> I agree w your complicated-construct sentiment re the cmd-pipeline.
> 
> this used to be done with a grep,
> undone by
> 
> 
> commit 91416cfdf98bdbc828fd3e5ca7208beba5979d63
> Author: Stephen Hemminger <shemminger@vyatta.com>
> Date:   Mon Feb 22 15:17:22 2010 -0800
> 
>     export_report: fix perl warnings
> 
>     Use local file handles, use three argument open.
>     Don't modify arguments in perl grep (use sed instead)
> 
>     Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
>     Acked-by: WANG Cong <amwang@redhat.com>
>     Cc: Michal Marek <mmarek@suse.cz>
>     Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
>     Signed-off-by: Michal Marek <mmarek@suse.cz>
> 
> diff --git a/scripts/export_report.pl b/scripts/export_report.pl
> index 705b5ba..04dce7c 100644
> --- a/scripts/export_report.pl
> +++ b/scripts/export_report.pl
> @@ -49,10 +49,10 @@ sub usage {
>  }
> 
>  sub collectcfiles {
> -        my @file = `cat .tmp_versions/*.mod | grep '.*\.ko\$'`;
> -        @file = grep {s/\.ko/.mod.c/} @file;
> -       chomp @file;
> -        return @file;
> +    my @file
> +       = `cat .tmp_versions/*.mod | grep '.*\.ko\$' | sed s/\.ko$/.mod.c/`;
> +    chomp @file;
> +    return @file;
>  }
> 
> 
> the backslash escape noise in the subshell pipeline can be avoided
> by returning to the grep, with less process spawning too
> (not that this is executed often)
> I'll try that.

Even better would be doing the work in perl rather than using
shell for this.