rc filetype diff: extract diff parsing from diff-jump

Most diff consumers we've written only care about the "final" state
after parsing through a diff. Let's extract the diff parsing part,
for reuse in several new commands.

In future we should try to use this (or better, a diff-parsing library)
for patch-range.pl. We'd add a callback argument that is invoked once
perl hunk (or line). Unfortunately I haven't found that library for
Perl yet.

2 files changed, 179 insertions, 102 deletions

diff --git a/rc/filetype/diff-parse.pl b/rc/filetype/diff-parse.pl
new file mode 100755
index 00000000..f7858eb7
--- /dev/null
+++ b/rc/filetype/diff-parse.pl
@@ -0,0 +1,127 @@
+#!/usr/bin/env perl
+
+use warnings;
+
+sub quote {
+    my $token = shift;
+    $token =~ s/'/''/g;
+    return "'$token'";
+}
+sub fail {
+    my $reason = shift;
+    print "set-register e fail " . quote("diff-parse.pl: $reason");
+    exit;
+}
+
+my $begin;
+my $end;
+
+while (defined $ARGV[0]) {
+    if ($ARGV[0] eq "--") {
+        shift;
+        last;
+    }
+    if ($ARGV[0] =~ m{^(BEGIN|END)$}) {
+        if (not defined $ARGV[1]) {
+            fail "missing argument to $ARGV[0]";
+        }
+        if ($ARGV[0] eq "BEGIN") {
+            $begin = $ARGV[1];
+        } else {
+            $end = $ARGV[1];
+        }
+        shift, shift;
+        next;
+    }
+    fail "unknown argument: $ARGV[0]";
+}
+
+# Inputs
+our $directory = $ENV{PWD};
+our $strip;
+our $version = "+";
+
+eval $begin if defined $begin;
+
+# Outputs
+our $file;
+our $file_line;
+our $diff_line_text;
+
+my $other_version;
+if ($version eq "+") {
+    $other_version = "-";
+} else {
+    $other_version = "+";
+}
+my $is_recursive_diff = 0;
+my $state = "header";
+my $fallback_file;
+my $other_file;
+my $other_file_line;
+
+sub strip {
+    my $is_recursive_diff = shift;
+    my $f = shift;
+
+    my $effective_strip;
+    if (defined $strip) {
+        $effective_strip = $strip;
+    } else {
+        # A "diff -r" or "git diff" adds "diff" lines to
+        # the output.  If no such line is present, we have
+        # a plain diff between files (not directories), so
+        # there should be no need to strip the directory.
+        $effective_strip = $is_recursive_diff ? 1 : 0;
+    }
+
+    if ($f !~ m{^/}) {
+        $f =~ s,^([^/]+/+){$effective_strip},, or fail "directory prefix underflow";
+        $f = "$directory/$f";
+    }
+    return $f;
+}
+
+while (<STDIN>) {
+    s/^(> )*//g;
+    $diff_line_text = $_;
+    if (m{^diff\b}) {
+        $state = "header";
+        $is_recursive_diff = 1;
+        if (m{^diff -\S* (\S+) (\S+)$}) {
+            $fallback_file = strip $is_recursive_diff, ($version eq "+" ? $2 : $1);
+        }
+        next;
+    }
+    if ($state eq "header") {
+        if (m{^[$version]{3} ([^\t\n]+)}) {
+            $file = strip $is_recursive_diff, $1;
+            next;
+        }
+        if (m{^[$other_version]{3} ([^\t\n]+)}) {
+            $other_file = strip $is_recursive_diff, $1;
+            next;
+        }
+    }
+    if (m{^@@ -(\d+)(?:,\d+)? \+(\d+)(?:,\d+)? @@}) {
+        $state = "contents";
+        $file_line = ($version eq "+" ? $2 : $1) - 1;
+        $other_file_line = ($version eq "+" ? $1 : $2) - 1;
+    } else {
+        my $iscontext = m{^[ ]};
+        if (m{^[ $version]}) {
+           $file_line++ if defined $file_line;
+        }
+        if (m{^[ $other_version]}) {
+           $other_file_line++ if defined $other_file_line;
+        }
+    }
+}
+if (not defined $file) {
+    $file = ($fallback_file or $other_file);
+}
+if (not defined $file) {
+    fail "missing diff header";
+}
+
+eval $end if defined $end;
diff --git a/rc/filetype/diff.kak b/rc/filetype/diff.kak
index 6d7ba695..530bf308 100644
--- a/rc/filetype/diff.kak
+++ b/rc/filetype/diff.kak
@@ -29,7 +29,7 @@ define-command diff-jump -params .. -docstring %{
             -       jump to the old file instead of the new file
             -<num> strip <num> leading directory components, like -p<num> in patch(1). Defaults to 1 if there is a 'diff' line (as printed by 'diff -r'), or 0 otherwise.
     } %{
-    evaluate-commands -draft -save-regs ac| %{
+    evaluate-commands -draft -save-regs c| %{
         # Save the column because we will move the cursor.
         set-register c %val{cursor_column}
         # If there is a "diff" line, we don't need to look further back.
@@ -41,120 +41,70 @@ define-command diff-jump -params .. -docstring %{
             # or content.
             execute-keys Gk
         }
-        set-register a %arg{@}
-        set-register | %{
-            [ -n "$kak_reg_a" ] && eval set -- "$kak_quoted_reg_a"
-            cmd=$(column=$kak_reg_c perl -we '
-                sub quote {
-                    $SQ = "'\''";
-                    $token = shift;
-                    $token =~ s/$SQ/$SQ$SQ/g;
-                    return "$SQ$token$SQ";
-                }
-                sub fail {
-                    $reason = shift;
-                    print "fail " . quote("diff-jump: $reason");
-                    exit;
-                }
-                $version = "+", $other_version = "-";
-                $strip = undef;
-                $directory = $ENV{PWD};
-                $seen_ddash = 0;
-                foreach (@ARGV) {
-                    if ($seen_ddash or !m{^-}) {
-                        $directory = $_;
-                    } elsif ($_ eq "-") {
-                        $version = "-", $other_version = "+";
-                    } elsif (m{^-(\d+)$}) {
-                        $strip = $1;
-                    } elsif ($_ eq "--") {
-                        $seen_ddash = 1;
-                    } else {
-                        fail "unknown option: $_";
-                    }
+        diff-parse BEGIN %{
+            my $seen_ddash = 0;
+            foreach (@ARGV) {
+                if ($seen_ddash or !m{^-}) {
+                    $directory = $_;
+                } elsif ($_ eq "-") {
+                    $version = "-", $other_version = "+";
+                } elsif (m{^-(\d+)$}) {
+                    $strip = $1;
+                } elsif ($_ eq "--") {
+                    $seen_ddash = 1;
+                } else {
+                    fail "unknown option: $_";
                 }
-                $have_diff_line = 0;
-                $state = "header";
-                while (<STDIN>) {
-                    s/^(> )*//g;
-                    $last_line = $_;
-                    if (m{^diff\b}) {
-                        $state = "header";
-                        $have_diff_line = 1;
-                        if (m{^diff -\S* (\S+) (\S+)$}) {
-                            $fallback_file = $version eq "+" ? $2 : $1;
-                        }
-                        next;
-                    }
-                    if ($state eq "header") {
-                        if (m{^[$version]{3} ([^\t\n]+)}) {
-                            $file = $1;
-                            next;
-                        }
-                        if (m{^[$other_version]{3} ([^\t\n]+)}) {
-                            $fallback_file = $1;
+            }
+        } END %exp{
+            my $file_column;
+            if (not defined $file_line) {
+                $file_line = "";
+                $file_column = "";
+            } else {
+                my $diff_column = %reg{c};
+                $file_column = $diff_column - 1; # Account for [ +-] diff prefix.
+                # If the cursor was on a hunk header, go to the section header if possible.
+                if ($diff_line_text =~ m{^(@@ -\d+(?:,\d+)? \+\d+(?:,\d+) @@ )([^\n]*)}) {
+                    my $hunk_header_prefix = $1;
+                    my $hunk_header_from_userdiff = $2;
+                    open FILE, "<", $file or fail "failed to open file: $!: $file";
+                    my @lines = <FILE>;
+                    for (my $i = $file_line - 1; $i >= 0 and $i < scalar @lines; $i--) {
+                        if ($lines[$i] !~ m{\Q$hunk_header_from_userdiff}) {
                             next;
                         }
-                    }
-                    if (m{^@@ -(\d+)(?:,\d+)? \+(\d+)(?:,\d+)? @@}) {
-                        $state = "contents";
-                        $line = ($version eq "+" ? $2 : $1) - 1;
-                    } elsif (m{^[ $version]}) {
-                        $line++ if defined $line;
-                    }
-                }
-                if (not defined $file) {
-                    $file = $fallback_file;
-                }
-                if (not defined $file) {
-                    fail "missing diff header";
-                }
-                if (not defined $strip) {
-                    # A "diff -r" or "git diff" adds "diff" lines to
-                    # the output.  If no such line is present, we have
-                    # a plain diff between files (not directories), so
-                    # there should be no need to strip the directory.
-                    $strip = $have_diff_line ? 1 : 0;
-                }
-                if ($file !~ m{^/}) {
-                    $file =~ s,^([^/]+/+){$strip},, or fail "directory prefix underflow";
-                    $file = "$directory/$file";
-                }
-
-                if (defined $line) {
-                    $column = $ENV{column} - 1; # Account for [ +-] diff prefix.
-                    # If the cursor was on a hunk header, go to the section header if possible.
-                    if ($last_line =~ m{^(@@ -\d+(?:,\d+)? \+\d+(?:,\d+) @@ )([^\n]*)}) {
-                        $hunk_header_prefix = $1;
-                        $hunk_header_from_userdiff = $2;
-                        open FILE, "<", $file or fail "failed to open file: $!: $file";
-                        @lines = <FILE>;
-                        for (my $i = $line - 1; $i >= 0 && $i < scalar @lines; $i--) {
-                            if ($lines[$i] !~ m{\Q$hunk_header_from_userdiff}) {
-                                next;
-                            }
-                            $line = $i + 1;
-                            # Re-add 1 because the @@ line does not have a [ +-] diff prefix.
-                            $column = $column + 1 - length $hunk_header_prefix;
-                            last;
-                        }
+                        $file_line = $i + 1;
+                        # Re-add 1 because the @@ line does not have a [ +-] diff prefix.
+                        $file_column = $diff_column + 1 - length $hunk_header_prefix;
+                        last;
                     }
                 }
-
-                printf "edit -existing -- %s $line $column", quote($file);
-            ' -- "$@")
-            echo "set-register c $cmd" >"$kak_command_fifo"
-        }
-        execute-keys <a-|><ret>
+            }
+            printf "set-register c %%s $file_line $file_column\n", quote($file);
+        } -- %arg{@}
         evaluate-commands -client %val{client} %{
             evaluate-commands -try-client %opt{jumpclient} %{
-                %reg{c}
+                edit -existing -- %reg{c}
             }
         }
     }
 }
 complete-command diff-jump file
 
+define-command -hidden diff-parse -params 2.. %{
+    evaluate-commands -save-regs ae %{
+        set-register a %arg{@}
+        set-register e nop
+        set-register | %{
+            eval set -- "$kak_quoted_reg_a"
+            perl "${kak_runtime}/rc/filetype/diff-parse.pl" "$@" >"$kak_command_fifo"
+        }
+        execute-keys <a-|><ret>
+        %reg{e}
+    }
+}
+
 §
 
 define-command \