[PATCH v3 00/11] Add new git-cc-cmd helper to contrib

[PATCH v3 00/11] Add new git-cc-cmd helper to contrib

[Felipe Contreras]

Hi,

Here goes another try, I made some changes based on comments, and should be
simpler now. Plus a few improvements. The interdiff will be on the patches, but
only the first one changed substantially.

This script allows you to get a list of relevant persons to Cc when sending a
patch series.

  % git cc-cmd v1.8.1.6^^1..v1.8.1.6^^2
  "Henrik Grubbström" <grubba@grubba.org> (author: 7%)
  junio (signer: 84%, author: 15%)
  "Nguyễn Thái Ngọc Duy" <pclouds@gmail.com> (author: 30%, signer: 7%)
  "Jean-Noël AVILA" <avila.jn@gmail.com> (author: 7%)
  Jean-Noel Avila <jn.avila@free.fr> (signer: 7%)
  Duy Nguyen <pclouds@gmail.com> (author: 7%)
  Michael Haggerty <mhagger@alum.mit.edu> (author: 15%)
  Clemens Buchacher <drizzd@aon.at> (author: 7%)
  Joshua Jensen <jjensen@workspacewhiz.com> (author: 7%)
  Johannes Sixt <j6t@kdbg.org> (signer: 7%)

It find people that might be interesting in a patch, by going back through the
history for each single hunk modified, and finding people that reviewed,
acknowledge, signed, or authored the code the patch is modifying.

It does this by running 'git blame' incrementally on each hunk, and then
parsing the commit message. After gathering all the relevant people, it groups
them to show what exactly was their role when the participated in the
development of the relevant commit, and on how many relevant commits they
participated. They are only displayed if they pass a minimum threshold of
participation.

The code finds the changes in each commit in the list, runs 'git blame'
to see which other commits are relevant to those lines, and then adds
the author and signer to the list.

Finally, it calculates what percentage of the total relevant commits
each person was involved in, and if it passes the threshold, it goes in.

You can also choose to show the commits themselves:

  % git cc-cmd --commits v1.8.1.6^^1..v1.8.1.6^^2
  9db9eec attr: avoid calling find_basename() twice per path
  94bc671 Add directory pattern matching to attributes
  82dce99 attr: more matching optimizations from .gitignore
  593cb88 exclude: split basename matching code into a separate function
  b559263 exclude: split pathname matching code into a separate function
  4742d13 attr: avoid searching for basename on every match
  f950eb9 rename pathspec_prefix() to common_prefix() and move to dir.[ch]
  4a085b1 consolidate pathspec_prefix and common_prefix
  d932f4e Rename git_checkattr() to git_check_attr()
  2d72174 Extract a function collect_all_attrs()
  8cf2a84 Add string comparison functions that respect the ignore_case variable.
  407a963 Merge branch 'rr/remote-helper-doc'
  ec775c4 attr: Expand macros immediately when encountered.

But wait, there's more: you can also specify a list of patch files, which means
this can be used for git send-emails --cc-cmd option.

Felipe Contreras (11):
  Add new git-cc-cmd helper to contrib
  contrib: cc-cmd: add option parsing
  contrib: cc-cmd: add support for multiple patches
  contrib: cc-cmd: add option to show commits
  contrib: cc-cmd: add option to parse from committish
  contrib: cc-cmd: parse committish like format-patch
  contrib: cc-cmd: fix parsing of rev-list args
  contrib: cc-cmd: add option to fetch aliases
  contrib: cc-cmd: support multiple roles
  contrib: cc-cmd: sort by participation
  contrib: cc-cmd: ignore chunks with no original lines

 contrib/cc-cmd/git-cc-cmd | 258 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 258 insertions(+)
 create mode 100755 contrib/cc-cmd/git-cc-cmd

-- 
1.8.2.1.790.g4588561

[PATCH v3 01/11] Add new git-cc-cmd helper to contrib

[Felipe Contreras]

This script find people that might be interesting in a patch, by going
back through the history for each single hunk modified, and finding
people that reviewed, acknowledge, signed, or authored the code the
patch is modifying.

It does this by running 'git blame' incrementally on each hunk, and then
parsing the commit message. After gathering all the relevant people, it
groups them to show what exactly was their role when the participated in
the development of the relevant commit, and on how many relevant commits
they participated. They are only displayed if they pass a minimum
threshold of participation.

For example:

  % git cc-cmd 0001-remote-hg-trivial-cleanups.patch
  Felipe Contreras <felipe.contreras@gmail.com> (author: 100%)
  Jeff King <peff@peff.net> (signer: 83%)
  Max Horn <max@quendi.de> (signer: 16%)
  Junio C Hamano <gitster@pobox.com> (signer: 16%)

Thus it can be used for 'git send-email' as a cc-cmd.

Comments-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 131 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 131 insertions(+)
 create mode 100755 contrib/cc-cmd/git-cc-cmd

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
new file mode 100755
index 0000000..aa83a1a
--- /dev/null
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -0,0 +1,131 @@
+#!/usr/bin/env ruby
+
+$since = '3-years-ago'
+$min_percent = 5
+
+class Commit
+
+  attr_reader :id, :roles
+
+  def initialize(id)
+    @id = id
+    @roles = []
+  end
+
+  def parse(data)
+    author = msg = nil
+    roles = {}
+    data.each_line do |line|
+      if not msg
+        case line
+        when /^author ([^<>]+) <(\S+)>$/
+          author = $1, $2
+          roles[author] = :author
+        when /^$/
+          msg = true
+        end
+      else
+        if line =~ /^(Signed-off|Reviewed|Acked)-by: ([^<>]+) <(\S+?)>$/
+          person = $2, $3
+          roles[person] = :signer if person != author
+        end
+      end
+    end
+    @roles = roles.map do |person, role|
+      [person, role]
+    end
+  end
+
+end
+
+class Commits
+
+  attr_reader :items
+
+  def initialize()
+    @items = {}
+  end
+
+  def size
+    @items.size
+  end
+
+  def import
+    return if @items.empty?
+    File.popen(%w[git cat-file --batch], 'r+') do |p|
+      p.write(@items.keys.join("\n"))
+      p.close_write
+      p.each do |l|
+        if l =~ /^(\h{40}) commit (\d+)/
+          id, len = $1, $2
+          data = p.read($2.to_i)
+          @items[id].parse(data)
+        end
+      end
+    end
+  end
+
+  def get_blame(source, start, offset, from)
+    return unless source
+    File.popen(['git', 'blame', '--incremental', '-C',
+               '-L', '%u,+%u' % [start, offset],
+               '--since', $since, from + '^',
+               '--', source]) do |p|
+      p.each do |line|
+        if line =~ /^(\h{40})/
+          id = $1
+          @items[id] = Commit.new(id)
+        end
+      end
+    end
+  end
+
+  def from_patch(file)
+    source = nil
+    from = nil
+    File.open(file) do |f|
+      f.each do |line|
+        case line
+        when /^From (\h+) (.+)$/
+          from = $1
+        when /^---\s+(\S+)/
+          source = $1 != '/dev/null' ? $1[2..-1] : nil
+        when /^@@\s-(\d+),(\d+)/
+          get_blame(source, $1, $2, from)
+        end
+      end
+    end
+  end
+
+end
+
+exit 1 if ARGV.size != 1
+
+commits = Commits.new
+commits.from_patch(ARGV[0])
+commits.import
+
+# hash of hashes
+persons = Hash.new { |hash, key| hash[key] = {} }
+
+commits.items.values.each do |commit|
+  commit.roles.each do |person, role|
+    persons[person][role] ||= 0
+    persons[person][role] += 1
+  end
+end
+
+persons.each do |person, roles|
+  roles = roles.map do |role, count|
+    percent = count.to_f * 100 / commits.size
+    next if percent < $min_percent
+    '%s: %u%%' % [role, percent]
+  end.compact
+  next if roles.empty?
+
+  name, email = person
+  # must quote chars?
+  name = '"%s"' % name if name =~ /[^\w \-]/i
+  person = name ? '%s <%s>' % [name, email] : email
+  puts '%s (%s)' % [person, roles.join(', ')]
+end
-- 
1.8.2.1.790.g4588561

[PATCH v3 02/11] contrib: cc-cmd: add option parsing

[Felipe Contreras]

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index aa83a1a..d78759d 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -1,8 +1,25 @@
 #!/usr/bin/env ruby
 
+require 'optparse'
+
 $since = '3-years-ago'
 $min_percent = 5
 
+begin
+  OptionParser.new do |opts|
+    opts.program_name = 'git cc-cmd'
+    opts.banner = 'usage: git cc-cmd [options] <file>'
+
+    opts.on('-p', '--min-percent N', Integer, 'Minium percentage of role participation') do |v|
+      $min_percent = v
+    end
+    opts.on('-d', '--since DATE', 'How far back to search for relevant commits') do |v|
+      $since = v
+    end
+  end.parse!
+rescue OptionParser::InvalidOption
+end
+
 class Commit
 
   attr_reader :id, :roles
-- 
1.8.2.1.790.g4588561

[PATCH v3 03/11] contrib: cc-cmd: add support for multiple patches

[Felipe Contreras]

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 34 ++++++++++++++++++----------------
 1 file changed, 18 insertions(+), 16 deletions(-)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index d78759d..3a9c70e 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -8,7 +8,7 @@ $min_percent = 5
 begin
   OptionParser.new do |opts|
     opts.program_name = 'git cc-cmd'
-    opts.banner = 'usage: git cc-cmd [options] <file>'
+    opts.banner = 'usage: git cc-cmd [options] <files>'
 
     opts.on('-p', '--min-percent N', Integer, 'Minium percentage of role participation') do |v|
       $min_percent = v
@@ -61,6 +61,7 @@ class Commits
 
   def initialize()
     @items = {}
+    @main_commits = {}
   end
 
   def size
@@ -91,24 +92,27 @@ class Commits
       p.each do |line|
         if line =~ /^(\h{40})/
           id = $1
-          @items[id] = Commit.new(id)
+          @items[id] = Commit.new(id) if not @main_commits.include?(id)
         end
       end
     end
   end
 
-  def from_patch(file)
+  def from_patches(files)
     source = nil
-    from = nil
-    File.open(file) do |f|
-      f.each do |line|
-        case line
-        when /^From (\h+) (.+)$/
-          from = $1
-        when /^---\s+(\S+)/
-          source = $1 != '/dev/null' ? $1[2..-1] : nil
-        when /^@@\s-(\d+),(\d+)/
-          get_blame(source, $1, $2, from)
+    files.each do |file|
+      from = nil
+      File.open(file) do |f|
+        f.each do |line|
+          case line
+          when /^From (\h+) (.+)$/
+            from = $1
+            @main_commits[from] = true
+          when /^---\s+(\S+)/
+            source = $1 != '/dev/null' ? $1[2..-1] : nil
+          when /^@@\s-(\d+),(\d+)/
+            get_blame(source, $1, $2, from)
+          end
         end
       end
     end
@@ -116,10 +120,8 @@ class Commits
 
 end
 
-exit 1 if ARGV.size != 1
-
 commits = Commits.new
-commits.from_patch(ARGV[0])
+commits.from_patches(ARGV)
 commits.import
 
 # hash of hashes
-- 
1.8.2.1.790.g4588561

[PATCH v3 04/11] contrib: cc-cmd: add option to show commits

[Felipe Contreras]

Instead of showing the authors and signers, show the commits themselves.

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index 3a9c70e..c49225f 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -4,6 +4,7 @@ require 'optparse'
 
 $since = '3-years-ago'
 $min_percent = 5
+$show_commits = false
 
 begin
   OptionParser.new do |opts|
@@ -16,6 +17,9 @@ begin
     opts.on('-d', '--since DATE', 'How far back to search for relevant commits') do |v|
       $since = v
     end
+    opts.on('-c', '--commits[=FORMAT]', [:raw, :full], 'List commits instead of persons') do |v|
+      $show_commits = v || true
+    end
   end.parse!
 rescue OptionParser::InvalidOption
 end
@@ -124,6 +128,20 @@ commits = Commits.new
 commits.from_patches(ARGV)
 commits.import
 
+if $show_commits
+  cmd = nil
+  case $show_commits
+  when :raw
+    puts commits.items.keys
+  when :full
+    cmd = %w[git log --patch --no-walk]
+  else
+    cmd = %w[git log --oneline --no-walk]
+  end
+  system(*cmd + commits.items.keys) if cmd
+  exit 0
+end
+
 # hash of hashes
 persons = Hash.new { |hash, key| hash[key] = {} }
 
-- 
1.8.2.1.790.g4588561

[PATCH v3 05/11] contrib: cc-cmd: add option to parse from committish

[Felipe Contreras]

For example master..feature-a.

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 36 ++++++++++++++++++++++++++++++++++--
 1 file changed, 34 insertions(+), 2 deletions(-)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index c49225f..aecfa45 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -5,11 +5,13 @@ require 'optparse'
 $since = '3-years-ago'
 $min_percent = 5
 $show_commits = false
+$files = []
+$rev_args = []
 
 begin
   OptionParser.new do |opts|
     opts.program_name = 'git cc-cmd'
-    opts.banner = 'usage: git cc-cmd [options] <files>'
+    opts.banner = 'usage: git cc-cmd [options] <files | rev-list options>'
 
     opts.on('-p', '--min-percent N', Integer, 'Minium percentage of role participation') do |v|
       $min_percent = v
@@ -122,10 +124,40 @@ class Commits
     end
   end
 
+  def from_rev_args(args)
+    return if args.empty?
+    source = nil
+    File.popen(%w[git rev-list --reverse] + args) do |p|
+      p.each do |e|
+        id = e.chomp
+        @main_commits[id] = true
+        File.popen(%w[git show -C --oneline] + [id]) do |p|
+          p.each do |e|
+            case e
+            when /^---\s+(\S+)/
+              source = $1 != '/dev/null' ? $1[2..-1] : nil
+            when /^@@\s-(\d+),(\d+)/
+              get_blame(source, $1, $2, id)
+            end
+          end
+        end
+      end
+    end
+  end
+
+end
+
+ARGV.each do |e|
+  if File.exists?(e)
+    $files << e
+  else
+    $rev_args << e
+  end
 end
 
 commits = Commits.new
-commits.from_patches(ARGV)
+commits.from_patches($files)
+commits.from_rev_args($rev_args)
 commits.import
 
 if $show_commits
-- 
1.8.2.1.790.g4588561

[PATCH v3 06/11] contrib: cc-cmd: parse committish like format-patch

[Felipe Contreras]

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index aecfa45..4bca7f1 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -126,6 +126,20 @@ class Commits
 
   def from_rev_args(args)
     return if args.empty?
+
+    revs = []
+
+    File.popen(%w[git rev-parse --revs-only --default=HEAD --symbolic] + args).each do |rev|
+      revs << rev.chomp
+    end
+
+    case revs.size
+    when 1
+      committish = [ '%s..HEAD' % revs[0] ]
+    else
+      committish = revs
+    end
+
     source = nil
     File.popen(%w[git rev-list --reverse] + args) do |p|
       p.each do |e|
-- 
1.8.2.1.790.g4588561

[PATCH v3 07/11] contrib: cc-cmd: fix parsing of rev-list args

[Felipe Contreras]

For example '-1'.

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index 4bca7f1..4fad030 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -23,7 +23,8 @@ begin
       $show_commits = v || true
     end
   end.parse!
-rescue OptionParser::InvalidOption
+rescue OptionParser::InvalidOption => e
+  $rev_args += e.args
 end
 
 class Commit
@@ -135,9 +136,11 @@ class Commits
 
     case revs.size
     when 1
-      committish = [ '%s..HEAD' % revs[0] ]
+      r = revs[0]
+      r = '^' + r if r[0] != '-'
+      args = [ r, 'HEAD' ]
     else
-      committish = revs
+      args = revs
     end
 
     source = nil
-- 
1.8.2.1.790.g4588561

[PATCH v3 08/11] contrib: cc-cmd: add option to fetch aliases

[Felipe Contreras]

Only the mutt format is supported for now.

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index 4fad030..200da0d 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -7,6 +7,8 @@ $min_percent = 5
 $show_commits = false
 $files = []
 $rev_args = []
+$get_aliases = false
+$aliases = {}
 
 begin
   OptionParser.new do |opts|
@@ -22,11 +24,32 @@ begin
     opts.on('-c', '--commits[=FORMAT]', [:raw, :full], 'List commits instead of persons') do |v|
       $show_commits = v || true
     end
+    opts.on('-a', '--aliases', 'Use aliases') do |v|
+      $get_aliases = v
+    end
   end.parse!
 rescue OptionParser::InvalidOption => e
   $rev_args += e.args
 end
 
+def get_aliases
+  type = %x[git config sendemail.aliasfiletype].chomp
+  return if type != 'mutt'
+  file = %x[git config sendemail.aliasesfile].chomp
+  File.open(File.expand_path(file)) do |f|
+    f.each do |line|
+      if line =~ /^\s*alias\s+(?:-group\s+\S+\s+)*(\S+)\s+(.*)$/
+        key, addresses = $1, $2.split(', ')
+        addresses.each do |address|
+          $aliases[address] = key
+        end
+      end
+    end
+  end
+end
+
+get_aliases if $get_aliases
+
 class Commit
 
   attr_reader :id, :roles
@@ -56,6 +79,8 @@ class Commit
       end
     end
     @roles = roles.map do |person, role|
+      address = "%s <%s>" % person
+      person = nil, $aliases[address] if $aliases.include?(address)
       [person, role]
     end
   end
-- 
1.8.2.1.790.g4588561

[PATCH v3 09/11] contrib: cc-cmd: support multiple roles

[Felipe Contreras]

Currently only the roles of 'author' and 'signer' and handler, but now
there's also 'reviewer' and 'acker'.

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 29 ++++++++++++++++++++---------
 1 file changed, 20 insertions(+), 9 deletions(-)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index 200da0d..67a276d 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -50,6 +50,12 @@ end
 
 get_aliases if $get_aliases
 
+KNOWN_ROLES = {
+  'Signed-off' => :signer,
+  'Reviewed' => :reviewer,
+  'Acked' => :acker,
+}
+
 class Commit
 
   attr_reader :id, :roles
@@ -61,27 +67,30 @@ class Commit
 
   def parse(data)
     author = msg = nil
-    roles = {}
+    # hash of arrays
+    roles = Hash.new { |hash, key| hash[key] = [] }
     data.each_line do |line|
       if not msg
         case line
         when /^author ([^<>]+) <(\S+)>$/
           author = $1, $2
-          roles[author] = :author
+          roles[author] << :author
         when /^$/
           msg = true
         end
       else
-        if line =~ /^(Signed-off|Reviewed|Acked)-by: ([^<>]+) <(\S+?)>$/
+        role_regex = KNOWN_ROLES.keys.join('|')
+        if line =~ /^(#{role_regex})-by: ([^<>]+) <(\S+?)>$/
           person = $2, $3
-          roles[person] = :signer if person != author
+          role = KNOWN_ROLES[$1]
+          roles[person] << role if person != author
         end
       end
     end
-    @roles = roles.map do |person, role|
+    @roles = roles.map do |person, roles|
       address = "%s <%s>" % person
       person = nil, $aliases[address] if $aliases.include?(address)
-      [person, role]
+      [person, roles]
     end
   end
 
@@ -220,9 +229,11 @@ end
 persons = Hash.new { |hash, key| hash[key] = {} }
 
 commits.items.values.each do |commit|
-  commit.roles.each do |person, role|
-    persons[person][role] ||= 0
-    persons[person][role] += 1
+  commit.roles.each do |person, roles|
+    roles.each do |role|
+      persons[person][role] ||= 0
+      persons[person][role] += 1
+    end
   end
 end
 
-- 
1.8.2.1.790.g4588561

[PATCH v3 10/11] contrib: cc-cmd: sort by participation

[Felipe Contreras]

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index 67a276d..9a2d2fd 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -237,7 +237,12 @@ commits.items.values.each do |commit|
   end
 end
 
-persons.each do |person, roles|
+# sort by number of participations
+count_sort = lambda do |a, b|
+  b[1].values.reduce(:+) <=> a[1].values.reduce(:+)
+end
+
+persons.sort(&count_sort).each do |person, roles|
   roles = roles.map do |role, count|
     percent = count.to_f * 100 / commits.size
     next if percent < $min_percent
-- 
1.8.2.1.790.g4588561

[PATCH v3 11/11] contrib: cc-cmd: ignore chunks with no original lines

[Felipe Contreras]

Suggested-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 contrib/cc-cmd/git-cc-cmd | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index 9a2d2fd..02841c4 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -125,7 +125,7 @@ class Commits
   end
 
   def get_blame(source, start, offset, from)
-    return unless source
+    return unless source and offset
     File.popen(['git', 'blame', '--incremental', '-C',
                '-L', '%u,+%u' % [start, offset],
                '--since', $since, from + '^',
-- 
1.8.2.1.790.g4588561

Re: [PATCH v3 01/11] Add new git-cc-cmd helper to contrib

[Felipe Contreras]

On Fri, Apr 19, 2013 at 2:30 PM, Felipe Contreras
<felipe.contreras@gmail.com> wrote:
> This script find people that might be interesting in a patch, by going
> back through the history for each single hunk modified, and finding
> people that reviewed, acknowledge, signed, or authored the code the
> patch is modifying.
>
> It does this by running 'git blame' incrementally on each hunk, and then
> parsing the commit message. After gathering all the relevant people, it
> groups them to show what exactly was their role when the participated in
> the development of the relevant commit, and on how many relevant commits
> they participated. They are only displayed if they pass a minimum
> threshold of participation.
>
> For example:
>
>   % git cc-cmd 0001-remote-hg-trivial-cleanups.patch
>   Felipe Contreras <felipe.contreras@gmail.com> (author: 100%)
>   Jeff King <peff@peff.net> (signer: 83%)
>   Max Horn <max@quendi.de> (signer: 16%)
>   Junio C Hamano <gitster@pobox.com> (signer: 16%)
>
> Thus it can be used for 'git send-email' as a cc-cmd.

Here's the interdiff with the previous one:

diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
index c7ecf79..aa83a1a 100755
--- a/contrib/cc-cmd/git-cc-cmd
+++ b/contrib/cc-cmd/git-cc-cmd
@@ -5,40 +5,35 @@ $min_percent = 5

 class Commit

-  attr_reader :id
-  attr_accessor :roles
+  attr_reader :id, :roles

   def initialize(id)
     @id = id
     @roles = []
   end

-  def self.parse(data)
-    id = author = msg = nil
+  def parse(data)
+    author = msg = nil
     roles = {}
     data.each_line do |line|
       if not msg
         case line
-        when /^commit (.+)$/
-          id = $1
         when /^author ([^<>]+) <(\S+)>$/
           author = $1, $2
-          roles[author] = 'author'
+          roles[author] = :author
         when /^$/
           msg = true
         end
       else
         if line =~ /^(Signed-off|Reviewed|Acked)-by: ([^<>]+) <(\S+?)>$/
           person = $2, $3
-          roles[person] = 'signer' if person != author
+          roles[person] = :signer if person != author
         end
       end
     end
-    roles = roles.map do |person, role|
-      address = "%s <%s>" % person
+    @roles = roles.map do |person, role|
       [person, role]
     end
-    [id, roles]
   end

 end
@@ -57,22 +52,15 @@ class Commits

   def import
     return if @items.empty?
-    format = [ 'commit %H', 'author %an <%ae>', '', '%B' ].join('%n')
-    File.popen(['git', 'show', '-z', '-s', '--format=format:' +
format] + @items.keys) do |p|
-      p.each("\0") do |data|
-        next if data == "\0" # bug in git show?
-        id, roles = Commit.parse(data)
-        commit = @items[id]
-        commit.roles = roles
-      end
-    end
-  end
-
-  def each_person_role
-    commit_roles = @items.values.map { |commit| commit.roles }.flatten(1)
-    commit_roles.group_by { |person, role| person }.each do |person,
commit_roles|
-      commit_roles.group_by { |person, role| role }.each do |role,
commit_roles|
-        yield person, role, commit_roles.size
+    File.popen(%w[git cat-file --batch], 'r+') do |p|
+      p.write(@items.keys.join("\n"))
+      p.close_write
+      p.each do |l|
+        if l =~ /^(\h{40}) commit (\d+)/
+          id, len = $1, $2
+          data = p.read($2.to_i)
+          @items[id].parse(data)
+        end
       end
     end
   end
@@ -107,34 +95,37 @@ class Commits
         end
       end
     end
-    import
   end

 end

 exit 1 if ARGV.size != 1

-commits = Commits.new()
+commits = Commits.new
 commits.from_patch(ARGV[0])
+commits.import

 # hash of hashes
 persons = Hash.new { |hash, key| hash[key] = {} }

-commits.each_person_role do |person, role, count|
-  persons[person][role] = count
+commits.items.values.each do |commit|
+  commit.roles.each do |person, role|
+    persons[person][role] ||= 0
+    persons[person][role] += 1
+  end
 end

 persons.each do |person, roles|
   roles = roles.map do |role, count|
     percent = count.to_f * 100 / commits.size
     next if percent < $min_percent
-    "%s: %u%%" % [role, percent]
+    '%s: %u%%' % [role, percent]
   end.compact
   next if roles.empty?

   name, email = person
   # must quote chars?
   name = '"%s"' % name if name =~ /[^\w \-]/i
-  person = name ? "%s <%s>" % [name, email] : email
-  puts "%s (%s)" % [person, roles.join(', ')]
+  person = name ? '%s <%s>' % [name, email] : email
+  puts '%s (%s)' % [person, roles.join(', ')]
 end

-- 
Felipe Contreras

Re: [PATCH v3 11/11] contrib: cc-cmd: ignore chunks with no original lines

[Junio C Hamano]

Felipe Contreras <felipe.contreras@gmail.com> writes:

> Suggested-by: Junio C Hamano <gitster@pobox.com>
> Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
> ---
>  contrib/cc-cmd/git-cc-cmd | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/contrib/cc-cmd/git-cc-cmd b/contrib/cc-cmd/git-cc-cmd
> index 9a2d2fd..02841c4 100755
> --- a/contrib/cc-cmd/git-cc-cmd
> +++ b/contrib/cc-cmd/git-cc-cmd
> @@ -125,7 +125,7 @@ class Commits
>    end
>  
>    def get_blame(source, start, offset, from)
> -    return unless source
> +    return unless source and offset

I do not think a separate patch at the end is a good idea for a
fix-up like this, even if to give credit to others.

Just squashing it in to the original would help reviewers; that way,
they do not have to wonder "why is this broken" and later notice
"ah, it is fixed at the end".

I do not know if that is an "offset", though.  It is the "offset"
for the end of the block relative to its beginning, but people
usually call that the "size" of the block, I think.

>      File.popen(['git', 'blame', '--incremental', '-C',
>                 '-L', '%u,+%u' % [start, offset],
>                 '--since', $since, from + '^',

Re: [PATCH v3 01/11] Add new git-cc-cmd helper to contrib

[Junio C Hamano]

Felipe Contreras <felipe.contreras@gmail.com> writes:

> This script find people that might be interesting in a patch, by going
> back through the history for each single hunk modified, and finding
> people that reviewed, acknowledge, signed, or authored the code the
> patch is modifying.
>
> It does this by running 'git blame' incrementally on each hunk, and then
> parsing the commit message. After gathering all the relevant people, it
> groups them to show what exactly was their role when the participated in
> the development of the relevant commit, and on how many relevant commits
> they participated. They are only displayed if they pass a minimum
> threshold of participation.
>
> For example:
>
>   % git cc-cmd 0001-remote-hg-trivial-cleanups.patch
>   Felipe Contreras <felipe.contreras@gmail.com> (author: 100%)
>   Jeff King <peff@peff.net> (signer: 83%)
>   Max Horn <max@quendi.de> (signer: 16%)
>   Junio C Hamano <gitster@pobox.com> (signer: 16%)

Should the contribution weight for each role add up to (close to)
100% if there were no minimum cut-off?

In general, it is not all that clear what these numbers mean.  Does
a sign/review on a single commit, no matter what kind of commit it
is or how big it is, count as contribution with the same weight?

I am not saying that the counting criteria needs to be configurable.
It just needs to be explainable to the end users.

> +commits = Commits.new
> +commits.from_patch(ARGV[0])
> +commits.import
> +
> +# hash of hashes
> +persons = Hash.new { |hash, key| hash[key] = {} }
> +
> +commits.items.values.each do |commit|
> +  commit.roles.each do |person, role|
> +    persons[person][role] ||= 0
> +    persons[person][role] += 1
> +  end
> +end
> +
> +persons.each do |person, roles|
> +  roles = roles.map do |role, count|
> +    percent = count.to_f * 100 / commits.size
> +    next if percent < $min_percent
> +    '%s: %u%%' % [role, percent]
> +  end.compact
> +  next if roles.empty?
> +
> +  name, email = person
> +  # must quote chars?
> +  name = '"%s"' % name if name =~ /[^\w \-]/i
> +  person = name ? '%s <%s>' % [name, email] : email
> +  puts '%s (%s)' % [person, roles.join(', ')]
> +end

Re: [PATCH v3 00/11] Add new git-cc-cmd helper to contrib

[Duy Nguyen]

On Sat, Apr 20, 2013 at 5:30 AM, Felipe Contreras
<felipe.contreras@gmail.com> wrote:
> This script allows you to get a list of relevant persons to Cc when sending a
> patch series.
>
>   % git cc-cmd v1.8.1.6^^1..v1.8.1.6^^2
>   "Nguyễn Thái Ngọc Duy" <pclouds@gmail.com> (author: 30%, signer: 7%)
>   Duy Nguyen <pclouds@gmail.com> (author: 7%)

You see, I like to cause confusion. These two are the same person.
Maybe you should check based on email only instead.

>   "Jean-Noël AVILA" <avila.jn@gmail.com> (author: 7%)
>   Jean-Noel Avila <jn.avila@free.fr> (signer: 7%)

And these two are like the same. Perhaps mailmap support will help?
--
Duy

Re: [PATCH v3 00/11] Add new git-cc-cmd helper to contrib

[Felipe Contreras]

On Fri, Apr 19, 2013 at 8:16 PM, Duy Nguyen <pclouds@gmail.com> wrote:
> On Sat, Apr 20, 2013 at 5:30 AM, Felipe Contreras
> <felipe.contreras@gmail.com> wrote:
>> This script allows you to get a list of relevant persons to Cc when sending a
>> patch series.
>>
>>   % git cc-cmd v1.8.1.6^^1..v1.8.1.6^^2
>>   "Nguyễn Thái Ngọc Duy" <pclouds@gmail.com> (author: 30%, signer: 7%)
>>   Duy Nguyen <pclouds@gmail.com> (author: 7%)
>
> You see, I like to cause confusion. These two are the same person.
> Maybe you should check based on email only instead.

Should be easy to implement.

>>   "Jean-Noël AVILA" <avila.jn@gmail.com> (author: 7%)
>>   Jean-Noel Avila <jn.avila@free.fr> (signer: 7%)
>
> And these two are like the same. Perhaps mailmap support will help?

There's already support for user-defined alias
(sendemail.aliasesfile), it shouldn't be too difficult to parse a
.mailmap file too.

Cheers.

-- 
Felipe Contreras