Skip to contentSkip to conversation

Subject checkversion: perl cleanup

PATCH In review Ralf Baechle 6 participants 23 messages Feb 19, 2010 – Feb 22, 2010 11 messages not in the archive

lore
Share & tools
lore
#5 lore raw

Re: [PATCH 01/10] checkstack: fix perlcritic warnings

In reply to a message not in the archive
Quoted 2 lines · Stephen Hemminger

Turn on strict checking.
Get rid of prototypes (prototypes are broken in Perl)

NAK, please check perldoc of sort().
The prototype there is needed by sort().

Quoted 1 line

Fix syntax of declaration.

Yeah, this is fine.

BTW, your $subject needs to be fixed too.

#17 lore raw

Re: [PATCH 04/10] namespace: perlcritic warnings

In reply to a message not in the archive

You do more than what you said, e.g. you replaced a return in a subroutine with die() which I think is not correct.

Please describe every of your changes.

Thanks.

#14 lore raw

Re: [PATCH 02/10] checkincludes: fix perlcritic warnings

In reply to Cong Wang
Quoted 2 lines · On Sat, 20 Feb 2010 11:05:00 +0800 Cong Wang <[email protected]> wrote:

This looks fine, but your $subject needs to be fixed.

Sorry, I use netdev standard subject name format. (Inherited from Jeff Garzik)

#6 lore raw

PATCH 1/1 scripts: improve checkstack

In reply to Cong Wang

Turn on strict checking, and get rid of annoying use of prototype. Fix syntax error in declaration

Use efficient sort algorithm by using schwartzian transform.
http://en.wikipedia.org/wiki/Schwartzian_transform

scripts/checkstack.pl +7 -10
@@ -21,6 +21,8 @@
 #
 #	TODO :	Port to all architectures (one regex per arch)
 
+use strict;
+
 # check for arch
 #
 # $re is used for two matches:
@@ -104,19 +106,11 @@ my (@stack, $re, $dre, $x, $xs);
 	}
 }
 
-sub bysize($) {
-	my ($asize, $bsize);
-	($asize = $a) =~ s/.*:	*(.*)$/$1/;
-	($bsize = $b) =~ s/.*:	*(.*)$/$1/;
-	$bsize <=> $asize
-}
-
 #
 # main()
 #
 my $funcre = qr/^$x* <(.*)>:$/;
-my $func;
-my $file, $lastslash;
+my ($func, $file, $lastslash);
 
 while (my $line = <STDIN>) {
 	if ($line =~ m/$funcre/) {
@@ -173,4 +167,7 @@ while (my $line = <STDIN>) {
 	}
 }
 
-print sort bysize @stack;
+# Use Schwartzian transform to sort by last field (size)
+print map { $_->[0] }
+      sort { $b->[1] <=> $a->[1] }
+      map { [$_, /:\t*(\d+)$/] } @stack;
#18 lore raw

Re: [PATCH 04/10] namespace: perlcritic warnings

In reply to Cong Wang

If objdump program is missing program should die with message to standard error and exit with non zero status, not continue on!

#15 lore raw

Re: [PATCH 02/10] checkincludes: fix perlcritic warnings

In reply to Stephen Hemminger
Quoted 7 lines · Stephen Hemminger
Quoted 2 lines · On Sat, 20 Feb 2010 11:05:00 +0800 Cong Wang <[email protected]> wrote:

This looks fine, but your $subject needs to be fixed.

Sorry, I use netdev standard subject name format. (Inherited from Jeff Garzik)

Oh, sorry, I misunderstood it. Then it is fine.

#7 lore raw

Re: [PATCH] scripts: improve checkstack

In reply to Stephen Hemminger
Quoted 5 lines · Stephen Hemminger

Turn on strict checking, and get rid of annoying use of prototype. Fix syntax error in declaration

Use efficient sort algorithm by using schwartzian transform.
http://en.wikipedia.org/wiki/Schwartzian_transform

Yeah, the idea is good, this can also make it more perlish. ;)

But...

scripts/checkstack.pl while (my $line = <STDIN>) { · line 173

Quoted 5 lines
scripts/checkstack.pl +4 -1
-print sort bysize @stack;
+# Use Schwartzian transform to sort by last field (size)
+print map { $_->[0] }
+      sort { $b->[1] <=> $a->[1] }
+      map { [$_, /:\t*(\d+)$/] } @stack;

This regex here is not strictly the same as before.

Can we just keep the original regex? If not, what's wrong?

Thanks.

#19 lore raw

Re: [PATCH 04/10] namespace: perlcritic warnings

In reply to Stephen Hemminger
Quoted 2 lines · Stephen Hemminger

If objdump program is missing program should die with message to standard error and exit with non zero status, not continue on!

Oh, right, makes sense.

#20 lore raw

Re: [PATCH 04/10] namespace: perlcritic warnings

In reply to a message not in the archive

On Sat, Feb 20, 2010 at 01:52, Stephen Hemminger <[email protected]> wrote:> Use local file handle not global.> Make loop and other variables local in scope.>> Signed-off-by: Stephen Hemminger <[email protected]>>>> --- a/scripts/namespace.pl      2010-02-19 08:59:00.432184740 -0800> +++ b/scripts/namespace.pl      2010-02-19 09:05:52.672644797 -0800> @@ -175,12 +175,11 @@ sub do_nm>        }>        if (! -e "$source.c" && ! -e "$source.S") {>                # No obvious source, exclude the object if it is conglomerate> -               if (! open(OBJDUMPDATA, "$objdump $basename|")) {> -                       printf STDERR "$objdump $fullname failed $!\n";> -                       return;> -               }> +               open(my $objdumpdata, "$objdump $basename|")> +                   or die "$objdump $fullname failed $!\n"; Looks it just want return, are you sure have to change it to die?

Quoted 1 line

+>                my $comment;> -               while (<OBJDUMPDATA>) {> +               while (<$objdumpdata>) {>                        chomp();>                        if (/^In archive/) {>                                # Archives are always conglomerate> @@ -190,18 +189,18 @@ sub do_nm>                        next if (! /^[ 0-9a-f]{5,} /);>                        $comment .= substr($_, 43);>                }> -               close(OBJDUMPDATA);> +               close($objdumpdata);> +>                if (!defined($comment) || $comment !~ /GCC\:.*GCC\:/m) {>                        printf STDERR "No source file found for $fullname\n";>                }>                return;>        }> -       if (! open(NMDATA, "$nm $basename|")) {> -               printf STDERR "$nm $fullname failed $!\n";> -               return;> -       }> +       open (my $nmdata, "$nm $basename|")> +           or die "$nm $fullname failed $!\n";> +>        my @nmdata;> -       while (<NMDATA>) {> +       while (<$nmdata>) {>                chop;>                ($type, $name) = (split(/ +/, $_, 3))[1..2];>                # Expected types> @@ -268,7 +267,8 @@ sub do_nm>                        }>                }>        }> -       close(NMDATA);> +       close($nmdata);> +>        if ($#nmdata < 0) {>                if (>                        $fullname ne "lib/brlock.o"> @@ -316,8 +316,7 @@ sub drop_def>>  sub list_multiply_defined>  {> -       my ($name, $module);> -       foreach $name (keys(%def)) {> +       foreach my $name (keys(%def)) {>                if ($#{$def{$name}} > 0) {>                        # Special case for cond_syscall>                        if ($#{$def{$name}} == 1 && $name =~ /^sys_/ &&> @@ -333,8 +332,9 @@ sub list_multiply_defined>                                &drop_def("arch/x86/kernel/vsyscall-sysenter_32.o", $name);>                                next;>                        }> +>                        printf "$name is multiply defined in :-\n";> -                       foreach $module (@{$def{$name}}) {> +                       foreach my $module (@{$def{$name}}) {>                                printf "\t$module\n";>                        }>                }> @@ -343,12 +343,13 @@ sub list_multiply_defined>>  sub resolve_external_references>  {> -       my ($object, $type, $name, $i, $j, $kstrtab, $ksymtab, $export);> +       my ($kstrtab, $ksymtab, $export);> +>        printf "\n";> -       foreach $object (keys(%nmdata)) {> +       foreach my $object (keys(%nmdata)) {>                my $nmdata = $nmdata{$object};> -               for ($i = 0; $i <= $#{$nmdata}; ++$i) {> -                       ($type, $name) = split(' ', $nmdata->[$i], 2);> +               for (my $i = 0; $i <= $#{$nmdata}; ++$i) {> +                       my ($type, $name) = split(' ', $nmdata->[$i], 2);>                        if ($type eq "U" || $type eq "w") {>                                if (exists($def{$name}) || exists($ksymtab{$name})) {>                                        # add the owning object to the nmdata> @@ -357,7 +358,7 @@ sub resolve_external_references>                                        $kstrtab = "R __kstrtab_$name";>                                        $ksymtab = "R __ksymtab_$name";>                                        $export = 0;> -                                       for ($j = 0; $j <= $#{$nmdata}; ++$j) {> +                                       for (my $j = 0; $j <= $#{$nmdata}; ++$j) {>                                                if ($nmdata->[$j] eq $kstrtab ||>                                                    $nmdata->[$j] eq $ksymtab) {>                                                        $export = 1;> @@ -424,11 +425,11 @@ sub resolve_external_references>  sub list_extra_externals>  {>        my %noref = ();> -       my ($name, @module, $module, $export);> -       foreach $name (keys(%def)) {> +> +       foreach my $name (keys(%def)) {>                if (! exists($ref{$name})) {> -                       @module = @{$def{$name}};> -                       foreach $module (@module) {> +                       my @module = @{$def{$name}};> +                       foreach my $module (@module) {>                                if (! exists($noref{$module})) {>                                        $noref{$module} = [];>                                }> @@ -438,16 +439,16 @@ sub list_extra_externals>        }>        if (%noref) {>                printf "\nExternally defined symbols with no external references\n";> -               foreach $module (sort(keys(%noref))) {> +               foreach my $module (sort(keys(%noref))) {>                        printf "  $module\n";>                        foreach (sort(@{$noref{$module}})) {> -                               if (exists($export{$_})) {> -                                       $export = " (export only)";> -                               }> -                               else {> -                                       $export = "";> -                               }> -                               printf "    $_$export\n";> +                           my $export;> +                           if (exists($export{$_})) {> +                               $export = " (export only)";> +                           } else {> +                               $export = "";> +                           }> +                           printf "    $_$export\n";>                        }>                }>        }>> -->>ÿôèº{.nÇ+‰·Ÿ®‰­†+%ŠËÿ±éݶ¥Šwÿº{.nÇ+‰·¥Š{±þG«�éÿŠ{ayºʇڙë,j­¢f£¢·hš�ï�êÿ‘êçz_è®(­éšŽŠÝ¢j"�ú¶m§ÿÿ¾«þG«�éÿ¢¸?™¨è­Ú&£ø§~�á¶iO•æ¬z·švØ^¶m§ÿÿà ÿ¶ìÿ¢¸?–I¥

#21 lore raw

Re: [PATCH 04/10] namespace: perlcritic warnings

Quoted 7 lines · On Sat, 20 Feb 2010 16:26:24 +0800 Hui Zhu <[email protected]> wrote:
Quoted 6 lines · Stephen Hemminger

Use local file handle not global.
Make loop and other variables local in scope.

Quoted 14 lines

scripts/namespace.pl      2010-02-19 09:05:52.672644797 -0800 sub do_nm · line 179

Quoted 12 lines
scripts/namespace.pl      2010-02-19 09:05:52.672644797 -0800 renamed from scripts/namespace.pl      2010-02-19 08:59:00.432184740 -0800 +2 -4
@@ -175,12 +175,11 @@ sub do_nm
        }
        if (! -e "$source.c" && ! -e "$source.S") {
                # No obvious source, exclude the object if it is conglomerate
-               if (! open(OBJDUMPDATA, "$objdump $basename|")) {
-                       printf STDERR "$objdump $fullname failed $!\n";
-                       return;
-               }
+               open(my $objdumpdata, "$objdump $basename|")
+                   or die "$objdump $fullname failed $!\n";

Looks it just want return, are you sure have to change it to die?

Read my previous comment

Quoted 2 lines

If objdump program is missing program should die with message to standard error and exit with non zero status, not continue on!

#8 lore raw

Re: [PATCH] scripts: improve checkstack

In reply to Cong Wang
Quoted 21 lines · On Sat, 20 Feb 2010 15:35:06 +0800 Cong Wang <[email protected]> wrote:
Quoted 5 lines · Stephen Hemminger

Turn on strict checking, and get rid of annoying use of prototype. Fix syntax error in declaration

Use efficient sort algorithm by using schwartzian transform.
http://en.wikipedia.org/wiki/Schwartzian_transform

Yeah, the idea is good, this can also make it more perlish. ;)

But...

scripts/checkstack.pl while (my $line = <STDIN>) { · line 173

Quoted 5 lines
scripts/checkstack.pl +4 -1
-print sort bysize @stack;
+# Use Schwartzian transform to sort by last field (size)
+print map { $_->[0] }
+      sort { $b->[1] <=> $a->[1] }
+      map { [$_, /:\t*(\d+)$/] } @stack;

This regex here is not strictly the same as before.

Quoted 1 line

Can we just keep the original regex? If not, what's wrong?

The original one had extra cruft:
1. The expression is anchored on right, so leading .* is meaningless
2. Putting tab directly makes it invisible when reading source (use \t)
3. Want to match on number, not just anything (that is why the \d)

#9 lore raw

Re: [PATCH] scripts: improve checkstack

On 2/22/2010 8:43, Stephen Hemminger wrote:

btw as a general comment on these guys

"make it more perl-ish" is not a good feature for me.

"make it more readable" would be.

these guys are not performance sensitive, lets make them as readable as possible, and not line noise.

#10 lore raw

Re: [PATCH] scripts: improve checkstack

Quoted 11 lines · Arjan van de Ven

On 2/22/2010 8:43, Stephen Hemminger wrote:

btw as a general comment on these guys

"make it more perl-ish" is not a good feature for me.

"make it more readable" would be.

these guys are not performance sensitive, lets make them as readable as possible, and not line noise.

Acked-by Steven Rostedt

That is, I've purposely did not do "perlish" code in the perl scripts that I wrote, and have consistently NACKed changes to make it so.

The perl scripts in the kernel need to be "C-ish" not "perl-ish", that way the other 70% of the kernel programmers that do not do perl can still understand these scripts.

-- Steve

#11 lore raw

Re: [PATCH] scripts: improve checkstack

Cleanup checkstack script:
* Turn on strict checking
* Fix resulting error message because the declaration syntax was incorrect.
* Remove incorrect and misleading use of prototype
- prototype not required for this type of sort function because $a and $b are being used in this contex
- if prototype was being used it should be for both arguments
* Use closure for sort function

scripts/checkstack.pl +6 -10
@@ -21,6 +21,8 @@
 #
 #	TODO :	Port to all architectures (one regex per arch)
 
+use strict;
+
 # check for arch
 #
 # $re is used for two matches:
@@ -104,19 +106,11 @@ my (@stack, $re, $dre, $x, $xs);
 	}
 }
 
-sub bysize($) {
-	my ($asize, $bsize);
-	($asize = $a) =~ s/.*:	*(.*)$/$1/;
-	($bsize = $b) =~ s/.*:	*(.*)$/$1/;
-	$bsize <=> $asize
-}
-
 #
 # main()
 #
 my $funcre = qr/^$x* <(.*)>:$/;
-my $func;
-my $file, $lastslash;
+my ($func, $file, $lastslash);
 
 while (my $line = <STDIN>) {
 	if ($line =~ m/$funcre/) {
@@ -173,4 +167,6 @@ while (my $line = <STDIN>) {
 	}
 }
 
-print sort bysize @stack;
+# Sort output by size (last field)
+print sort { ($b =~ /:\t*(\d+)$/)[0] <=> ($a =~ /:\t*(\d+)$/)[0] } @stack;
+