lei rediff: add --drq and --dequote-only

More switches which can be useful for users who pipe from text
editors.  --drq can be helpful while writing patch review email
replies, and perhaps --dequote-only, too.

4 files changed, 113 insertions, 3 deletions

diff --git a/Documentation/lei-rediff.pod b/Documentation/lei-rediff.pod
index f34e946f..e711de97 100644
--- a/Documentation/lei-rediff.pod
+++ b/Documentation/lei-rediff.pod
@@ -38,6 +38,33 @@ to C<lei rediff> before piping it to L<git-am(1)>.  The output
 of C<lei rediff> is compatible with C<git am> if its input was a
 patch email.
 
+=item --drq[=COUNT]
+
+De-Re-Quote.  De-quote the input and re-quotes (the output).
+Removes COUNT levels of C<E<gt> > email reply prefixes and
+re-adds them upon regenerating the diff.
+
+This switch is intended as a convenience for running inside a
+pipe-capable text editor when writing replies to a patch email.
+Note: this may over-add C<E<gt> > prefixes if some input lines
+are missing C<E<gt> > prefixes.
+
+COUNT is 1 if unspecified; in other words, C<--drq=1> and
+C<--drq> are equivalent.
+
+It implies L</--quiet> unless L</--verbose> is specified
+since text editors tend to combine stderr with stdout.
+
+=item --dequote-only[=COUNT]
+
+Like L</--drq>, but does not re-add quote prefixes to the output.
+
+This can be useful for feeding a hunk to L<git-apply(1)>
+or L<patch(1)> while writing a reply or further processing
+by another diff viewer.
+
+Unlike L</--drq>, it does NOT implies L</--quiet>.
+
 =item --git-dir=DIR
 
 Specify an additional .git/ directory to scan.  This option may be
diff --git a/lib/PublicInbox/LEI.pm b/lib/PublicInbox/LEI.pm
index 8faf74a2..b8159cba 100644
--- a/lib/PublicInbox/LEI.pm
+++ b/lib/PublicInbox/LEI.pm
@@ -201,7 +201,7 @@ our %CMD = ( # sorted in order of importance/use:
 'rediff' => [ '--stdin|LOCATION...',
                 'regenerate a diff with different options',
         'stdin|', # /|\z/ must be first for lone dash
-        qw(git-dir=s@ cwd! verbose|v+ color:s no-color),
+        qw(git-dir=s@ cwd! verbose|v+ color:s no-color drq:1 dequote-only:1),
         @diff_opt, @lxs_opt, @net_opt, @c_opt ],
 
 'add-external' => [ 'LOCATION',
diff --git a/lib/PublicInbox/LeiRediff.pm b/lib/PublicInbox/LeiRediff.pm
index f6960560..1e95e55a 100644
--- a/lib/PublicInbox/LeiRediff.pm
+++ b/lib/PublicInbox/LeiRediff.pm
@@ -134,6 +134,27 @@ EOM
         $pid = spawn(\@cmd, $lei->{env}, { 2 => $lei->{2}, 1 => $lei->{1} });
         waitpid($pid, 0);
         $lei->child_error($?) if $?; # for git diff --exit-code
+        undef;
+}
+
+sub wait_requote ($$$) { # OnDestroy callback
+        my ($lei, $pid, $old_1) = @_;
+        $lei->{1} = $old_1; # closes stdin of `perl -pE 's/^/> /'`
+        waitpid($pid, 0) == $pid or die "BUG(?) waitpid: \$!=$! \$?=$?";
+        $lei->child_error($?) if $?;
+}
+
+sub requote ($$) {
+        my ($lei, $pfx) = @_;
+        pipe(my($r, $w)) or die "pipe: $!";
+        my $rdr = { 0 => $r, 1 => $lei->{1}, 2 => $lei->{2} };
+        # $^X (perl) is overkill, but maybe there's a weird system w/o sed
+        my $pid = spawn([$^X, '-pE', "s/^/$pfx/"], $lei->{env}, $rdr);
+        my $old_1 = $lei->{1};
+        $w->autoflush(1);
+        binmode $w, ':utf8';
+        $lei->{1} = $w;
+        PublicInbox::OnDestroy->new(\&wait_requote, $lei, $pid, $old_1);
 }
 
 sub extract_oids { # Eml each_part callback
@@ -142,6 +163,10 @@ sub extract_oids { # Eml each_part callback
         $self->{lei}->out($p->header_obj->as_string, "\n");
         my ($s, undef) = msg_part_text($p, $p->content_type || 'text/plain');
         defined $s or return;
+        my $rq;
+        if ($self->{dqre} && $s =~ s/$self->{dqre}//g) { # '> ' prefix(es)
+                $rq = requote($self->{lei}, $1) if $self->{lei}->{opt}->{drq};
+        }
         my @top = split($PublicInbox::ViewDiff::EXTRACT_DIFFS, $s);
         undef $s;
         my $blobs = $self->{blobs}; # blobs to resolve
@@ -191,6 +216,19 @@ sub extract_oids { # Eml each_part callback
         $ctxq = diff_ctxq($self, $ctxq);
 }
 
+# ensure dequoted parts are available for rebuilding patches:
+sub dequote_add { # Eml each_part callback
+        my ($ary, $self) = @_;
+        my ($p, undef, $idx) = @$ary;
+        my ($s, undef) = msg_part_text($p, $p->content_type || 'text/plain');
+        defined $s or return;
+        if ($s =~ s/$self->{dqre}//g) { # remove '> ' prefix(es)
+                substr($s, 0, 0, "part-dequoted: $idx\n\n");
+                utf8::encode($s);
+                $self->{tmp_sto}->add_eml(PublicInbox::Eml->new(\$s));
+        }
+}
+
 sub input_eml_cb { # callback for all emails
         my ($self, $eml) = @_;
         {
@@ -200,6 +238,7 @@ sub input_eml_cb { # callback for all emails
                         warn @_;
                 };
                 $self->{tmp_sto}->add_eml($eml);
+                $eml->each_part(\&dequote_add, $self) if $self->{dqre};
                 $self->{tmp_sto}->done;
         }
         $eml->each_part(\&extract_oids, $self, 1);
@@ -207,6 +246,10 @@ sub input_eml_cb { # callback for all emails
 
 sub lei_rediff {
         my ($lei, @inputs) = @_;
+        ($lei->{opt}->{drq} && $lei->{opt}->{'dequote-only'}) and return
+                $lei->fail('--drq and --dequote-only are mutually exclusive');
+        ($lei->{opt}->{drq} && !$lei->{opt}->{verbose}) and
+                $lei->{opt}->{quiet} //= 1;
         $lei->_lei_store(1)->write_prepare($lei);
         $lei->{opt}->{'in-format'} //= 'eml';
         # maybe it's a non-email (code) blob from a coderepo
@@ -257,6 +300,10 @@ sub ipc_atfork_child {
                 } @{$self->{lei}->{opt}->{'git-dir'}} ];
         $lei->{env}->{'psgi.errors'} = $lei->{2}; # ugh...
         $lei->{env}->{TMPDIR} = $self->{rdtmp}->dirname;
+        if (my $nr = ($lei->{opt}->{drq} || $lei->{opt}->{'dequote-only'})) {
+                my $re = '\s*> ' x $nr;
+                $self->{dqre} = qr/^($re)/ms;
+        }
         undef;
 }
 
diff --git a/t/solver_git.t b/t/solver_git.t
index 3c0b7f65..cf450e24 100644
--- a/t/solver_git.t
+++ b/t/solver_git.t
@@ -80,12 +80,48 @@ test_lei({tmpdir => "$tmpdir/rediff"}, sub {
         lei_ok(qw(rediff -q -U9 t/solve/0001-simple-mod.patch));
         like($lei_out, qr!^\Q+++\E b/TODO\n@@ -103,9 \+103,11 @@!sm,
                 'got more context with -U9');
-        lei_ok(qw(rediff -q -U9 t/solve/bare.patch));
+
+        my (undef, $re) = split(/\n\n/, $lei_out, 2);
+        $re =~ s/^/> /sgm;
+        substr($re, 0, 0, <<EOM);
+From: me\@example.com
+Subject: Re: awesome advice
+
+WEB DESIGN EXPERT wrote:
+EOM
+        lei_ok([qw(rediff --abbrev=40 -U16 --drq)], undef,
+                { 0 => \$re, %$lei_opt });
         my $exp = <<'EOM';
+From: me@example.com
+Subject: Re: awesome advice
+
+EOM
+        like($lei_out, qr/\Q$exp\E/, '--drq preserved header');
+
+        # n.b. --drq can requote the attribution line ("So-and-so wrote:"),
+        # but it's probably not worth preventing...
+
+        $exp = <<'EOM';
+> ---
+>  TODO | 2 ++
+>  Ω    | 5 --
+>  1 file changed, 2 insertions(+)
+>
+> diff --git a/TODO b/TODO
+> index 605013e4904baabecd4a0a55997aebd8e8477a8f..69df7d565d49fbaaeb0a067910f03dc22cd52bd0 100644
+> --- a/TODO
+> +++ b/TODO
+> @@ -96,16 +96,18 @@ all need to be considered for everything we introduce)
+EOM
+        $exp =~ s/^>$/> /sgm; # re-add trailing white space
+        like($lei_out, qr/\Q$exp\E/, '--drq diffstat + context');
+
+        lei_ok(qw(rediff -q --abbrev=40 -U9 t/solve/bare.patch));
+        $exp = <<'EOM';
 diff --git a/script/public-inbox-extindex b/script/public-inbox-extindex
 old mode 100644
 new mode 100755
-index 15ac20eb..771486c4
+index 15ac20eb871bf47697377e58a27db23102a38fca..771486c425b315bae70fd8a82d62ab0331e0a827
 --- a/script/public-inbox-extindex
 +++ b/script/public-inbox-extindex
 @@ -1,13 +1,12 @@