checkpatch: Introduce a check for the nesting of brackets
RFC 1/1 checkpatch: Introduce a check for the nesting of brackets
I have a local script that checks files for the right nesting of brackets. With the help of the script I managed to do a couple patches [1]. Unfortunately there are (still) a lot of false positives [2]. I do not filter false positives (yet).
To simplify testing the new check is enabled and warns(WARN()). The final version would be disabled by default and would trigger CHK().
Would this addition, if cleaned up, get into checkpatch.pl?
Thanks
Manuel
[1] Examples https://lore.kernel.org/all/?q=f%3Amanuelebner%40mailbox.org+AND+d%3A2026-06-11+AND+%28s%3A%22bracket%22+OR+s%3A%22brackets%22+%09OR+s%3A%22cleanup%22+OR+s%3A%22sysfs-firmware-dmi%22
[2] Smileys: ;-) :), :-( :(,
Enums: 1) 2), i) ii), a) b),
Arrows: ->, <-
Intervall: (0, 100], [0, 100)
Code snippets in Docs: Unclosed '{' or '('
Scissors: 8<--- cut here ------ scripts/checkpatch.pl | 52 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 52 insertions(+)
scripts/checkpatch.pl +52 -0
@@ -75,6 +75,17 @@ my $git_command ='export LANGUAGE=en_US.UTF-8; git'; my $tabsize = 8; my ${CONFIG_} = "CONFIG_"; +my %brackets = ( + ')' => '(', + '}' => '{', + ']' => '[', +); +my @stack_brackets; +my $line_num = 0; +my $has_error = 0; +my $last_open; +my %open_brackets = reverse %brackets; + my %maybe_linker_symbol; # for externs in c exceptions, when seen in *vmlinux.lds.h sub help {
@@ -3252,6 +3263,37 @@ sub process { $needs_fixes_tag = $1; } +# Check brackets for matching pairs + while ($line =~ /([\(\)\{\}\[\]])/g) { + my $char = $1; + # If it's an opening bracket, push onto stack + if (grep { $_ eq $char } values %brackets) { + push @stack_brackets, { char => $char, line => $line_num }; + } + # If it's a closing bracket, validate + elsif (exists $brackets{$char}) { + my $expected = $brackets{$char}; + if (!@stack_brackets) { + WARN("OPENING_BRACKET_MISSING", + "Missing opening bracket\n" . + "Found closing '$char' without an opening '$expected'\n" . + $herecurr); + $has_error = 1; + last; + } + $last_open = pop @stack_brackets; + if ($last_open->{char} ne $expected) { + WARN("MISMATCH_OF_BRACKETS", + "Mismatched brackets error\n" . + "Expected '$open_brackets{$last_open->{char}}' but found '$char' on" . "\n" . + $herecurr); + $has_error = 1; + last; + } + } + } + $line_num++; + # Check Fixes: styles is correct if (!$in_header_lines && $line =~ /^\s*(fixes:?)\s*(?:commit\s*)?([0-9a-f]{5,40})(?:\s*($balanced_parens))?/i) {
@@ -7806,6 +7848,16 @@ sub process { } } +# Check if there are any unclosed brackets remaining + if (!$has_error && @stack_brackets) { + foreach my $unclosed (reverse @stack_brackets) { + print "\n"; + WARN("CLOSING_BRACKET_MISSING", + "closing bracket missing\n" . + "Unclosed '$unclosed->{char}' started on line '$unclosed->{line}'\n"); + } + } + # If we have no input at all, then there is nothing to report on # so just keep quiet. if ($#rawlines == -1) {
Signature · 1 line
2.54.0
Joe Perches
maintainerQuoted 9 lines · Manuel Ebner
I have a local script that checks files for the right nesting of brackets. With the help of the script I managed to do a couple patches [1]. Unfortunately there are (still) a lot of false positives [2]. I do not filter false positives (yet).
To simplify testing the new check is enabled and warns(WARN()). The final version would be disabled by default and would trigger
CHK().Would this addition, if cleaned up, get into checkpatch.pl?
probably not as it's not suitable for patches.