[PATCH] get_maintainer.pl: Add support to match arbitrary text

[PATCH] get_maintainer.pl: Add support to match arbitrary text

[L. Alberto Giménez]

Extend the usage of the K section in the MAINTAINERS file to support matching
regular expressions with the content that precedes the patch (delimited by three
dashes "---").

The change enables the get_maintainer.pl script to get maintainers based on
arbitrary text that may precede the patch itself (for example, the commit
message or mail headers generated by git-format-patch)

Thanks to Jesper Juhl to point me that trivial patches should be CC'd to Jiri.
The idea is that the get_maintainer.pl can detect such situations and add the
additional email addresses and let the MAINTAINERS file handle who should get
copies of what.

Signed-off-by: L. Alberto Giménez <agimenez@sysvalve.es>
---
 MAINTAINERS               |    1 +
 scripts/get_maintainer.pl |   12 +++++++++++-
 2 files changed, 12 insertions(+), 1 deletions(-)

diff --git a/MAINTAINERS b/MAINTAINERS
index cf0f3a5..c1e33ac 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -6145,6 +6145,7 @@ TRIVIAL PATCHES
 M:	Jiri Kosina <trivial@kernel.org>
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial.git
 S:	Maintained
+K:	^Subject:.*\[(?i)trivial\].*
 
 TTY LAYER
 M:	Greg Kroah-Hartman <gregkh@suse.de>
diff --git a/scripts/get_maintainer.pl b/scripts/get_maintainer.pl
index 139e0ff..63d4d90 100755
--- a/scripts/get_maintainer.pl
+++ b/scripts/get_maintainer.pl
@@ -420,6 +420,11 @@ foreach my $file (@ARGV) {
 
 	open(my $patch, "< $file")
 	    or die "$P: Can't open $file: $!\n";
+
+	# We can add arbitrary information before the patch itself (commit message,
+	# mail headers,... This allows us to match arbitrary keywords agains any
+	# part of a git-format-patch generated file (subject tags, etc...)
+	my $in_patch = 0;
 	while (<$patch>) {
 	    my $patch_line = $_;
 	    if (m/^\+\+\+\s+(\S+)/) {
@@ -432,9 +437,14 @@ foreach my $file (@ARGV) {
 		if ($email_git_blame) {
 		    push(@range, "$lastfile:$1:$2");
 		}
+		} elsif (not $in_patch and m/^---/) {
+			# enter "patch area": keywords matched only on changed lines
+			$in_patch = 1;
 	    } elsif ($keywords) {
 		foreach my $line (keys %keyword_hash) {
-		    if ($patch_line =~ m/^[+-].*$keyword_hash{$line}/x) {
+			my $change_hook = $in_patch ? "^[+-].*" : "";
+
+			if ($patch_line =~ m/^${change_hook}$keyword_hash{$line}/x) {
 			push(@keyword_tvi, $line);
 		    }
 		}
-- 
1.7.4.rc0.5.gf2665

Re: [PATCH] get_maintainer.pl: Add support to match arbitrary text

[Joe Perches]

On Tue, 2011-01-25 at 00:32 +0100, L. Alberto Giménez wrote:
> Extend the usage of the K section in the MAINTAINERS file to support matching
> regular expressions with the content that precedes the patch (delimited by three
> dashes "---").
> +		} elsif (not $in_patch and m/^---/) {
> +			# enter "patch area": keywords matched only on changed lines
> +			$in_patch = 1;

Not all patches need or have ---

This would be better as "^---$"

Re: [PATCH] get_maintainer.pl: Add support to match arbitrary text

[L. Alberto Giménez]

On Mon, Jan 24, 2011 at 03:53:46PM -0800, Joe Perches wrote:
> On Tue, 2011-01-25 at 00:32 +0100, L. Alberto Giménez wrote:
> > Extend the usage of the K section in the MAINTAINERS file to support matching
> > regular expressions with the content that precedes the patch (delimited by three
> > dashes "---").
> > +		} elsif (not $in_patch and m/^---/) {
> > +			# enter "patch area": keywords matched only on changed lines
> > +			$in_patch = 1;
> 
> Not all patches need or have ---
> 
> This would be better as "^---$"

Hi Joe,

As I replied to you previously, please feel free to fix the patch. Most of the
patches I've seen use that line as a separator. Could you please give me the
right regexp to detect the beginning of patches?

On the other hand, if it's not properly detected, it will just match the "K"
keywords against all the patch lines (not just deleted or added lines). For a
standard unified diff format that would be 3 context lines that should match
with something like "^Subject: whatever". Again, it's just a proposal waiting
for comments/enhancements.

The idea here was to enable "triggers" in the subject to add the proper CCs. It
just ended up in a "match anything against the file content" change.

Regards,
-- 
L. Alberto Giménez
JabberID agimenez@jabber.sysvalve.es
GnuPG key ID 0x3BAABDE1

[PATCH] get_maintainer.pl: Allow "K:" pattern tests to match non-patch text

[Joe Perches]

Extend the usage of the K section in the MAINTAINERS file to support matching
regular expressions to any arbitrary text that may precede the patch itself.
For example, the commit message or mail headers generated by git-format-patch.

Original-patch-by: L. Alberto Giménez <agimenez@sysvalve.es>
Signed-off-by: Joe Perches <joe@perches.com>

---

> On Tue, 2011-01-25 at 01:25 +0100, L. Alberto Giménez wrote:
> feel free to fix the patch.

 scripts/get_maintainer.pl |   11 ++++++++++-
 1 files changed, 10 insertions(+), 1 deletions(-)

diff --git a/scripts/get_maintainer.pl b/scripts/get_maintainer.pl
index 139e0ff..d29a8d7 100755
--- a/scripts/get_maintainer.pl
+++ b/scripts/get_maintainer.pl
@@ -420,6 +420,14 @@ foreach my $file (@ARGV) {
 
 	open(my $patch, "< $file")
 	    or die "$P: Can't open $file: $!\n";
+
+	# We can check arbitrary information before the patch
+	# like the commit message, mail headers, etc...
+	# This allows us to match arbitrary keywords against any part
+	# of a git format-patch generated file (subject tags, etc...)
+
+	my $patch_prefix = "";			#Parsing the intro
+
 	while (<$patch>) {
 	    my $patch_line = $_;
 	    if (m/^\+\+\+\s+(\S+)/) {
@@ -428,13 +436,14 @@ foreach my $file (@ARGV) {
 		$filename =~ s@\n@@;
 		$lastfile = $filename;
 		push(@files, $filename);
+		$patch_prefix = "^[+-].*";	#Now parsing the actual patch
 	    } elsif (m/^\@\@ -(\d+),(\d+)/) {
 		if ($email_git_blame) {
 		    push(@range, "$lastfile:$1:$2");
 		}
 	    } elsif ($keywords) {
 		foreach my $line (keys %keyword_hash) {
-		    if ($patch_line =~ m/^[+-].*$keyword_hash{$line}/x) {
+		    if ($patch_line =~ m/${patch_prefix}$keyword_hash{$line}/x) {
 			push(@keyword_tvi, $line);
 		    }
 		}

Re: [PATCH] get_maintainer.pl: Allow "K:" pattern tests to match non-patch text

[L. Alberto Giménez]

On Tue, Jan 25, 2011 at 06:08:49AM -0800, Joe Perches wrote:
> Extend the usage of the K section in the MAINTAINERS file to support matching
> regular expressions to any arbitrary text that may precede the patch itself.
> For example, the commit message or mail headers generated by git-format-patch.
> 
> Original-patch-by: L. Alberto Giménez <agimenez@sysvalve.es>
> Signed-off-by: Joe Perches <joe@perches.com>

Hey, am I supposed to Ack it or something like that?
-- 
L. Alberto Giménez
JabberID agimenez@jabber.sysvalve.es
GnuPG key ID 0x3BAABDE1

Re: [PATCH] get_maintainer.pl: Allow "K:" pattern tests to match non-patch text

[Joe Perches]

On Tue, 2011-01-25 at 19:44 +0100, L. Alberto Giménez wrote:
> On Tue, Jan 25, 2011 at 06:08:49AM -0800, Joe Perches wrote:
> > Extend the usage of the K section in the MAINTAINERS file to support matching
> > regular expressions to any arbitrary text that may precede the patch itself.
> > For example, the commit message or mail headers generated by git-format-patch.
> > Original-patch-by: L. Alberto Giménez <agimenez@sysvalve.es>
> > Signed-off-by: Joe Perches <joe@perches.com>
> Hey, am I supposed to Ack it or something like that?

You can if you want, you don't have to.

Re: [PATCH] get_maintainer.pl: Allow "K:" pattern tests to match non-patch text

["L. Alberto Giménez"]

On 25/01/2011 19:58, Joe Perches wrote:
> On Tue, 2011-01-25 at 19:44 +0100, L. Alberto Giménez wrote:
>> On Tue, Jan 25, 2011 at 06:08:49AM -0800, Joe Perches wrote:
>>> Extend the usage of the K section in the MAINTAINERS file to support matching
>>> regular expressions to any arbitrary text that may precede the patch itself.
>>> For example, the commit message or mail headers generated by git-format-patch.
>>> Original-patch-by: L. Alberto Giménez<agimenez@sysvalve.es>
>>> Signed-off-by: Joe Perches<joe@perches.com>
>> Hey, am I supposed to Ack it or something like that?
> You can if you want, you don't have to.

If it's of any use:

Acked-by: L. Alberto Giménez <agimenez@sysvalve.es>

Re: [PATCH] get_maintainer.pl: Add support to match arbitrary text

[Joe Perches]

On Tue, 2011-01-25 at 00:32 +0100, L. Alberto Giménez wrote:
> Signed-off-by: L. Alberto Giménez <agimenez@sysvalve.es>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index cf0f3a5..c1e33ac 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -6145,6 +6145,7 @@ TRIVIAL PATCHES
>  M:	Jiri Kosina <trivial@kernel.org>
>  T:	git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial.git
>  S:	Maintained
> +K:	^Subject:.*\[(?i)trivial\].*

The trailing .* isn't necessary.

This matches just case insensitive "[trivial]".

Looking at the trivial patches posted:
https://patchwork.kernel.org/project/LKML/list/?q=trivial
that's not a very good matching pattern.

Maybe this is better:

K:	^(?i:subject):.*(?i:trivial)\s*[:,\]]