X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FViewDiff.pm;h=f7422712aa4516818a22183a0630f729bb1b2f0a;hb=3cfc29715f29a33d0f433e66c8e98341d4410c41;hp=57a1b5d66dd9dda27de87c3106da22fb3f385355;hpb=a7c58d0ce022671c9170401054a1131b9ec15bfa;p=public-inbox.git

diff --git a/lib/PublicInbox/ViewDiff.pm b/lib/PublicInbox/ViewDiff.pm
index 57a1b5d6..f7422712 100644
--- a/lib/PublicInbox/ViewDiff.pm
+++ b/lib/PublicInbox/ViewDiff.pm
@@ -25,7 +25,23 @@ my $ANY = qr![^\n]!;
 my $FN = qr!(?:"?[^/\n]+/[^\n]+|/dev/null)!;
 
 # cf. git diff.c :: get_compact_summary
-my $DIFFSTAT_COMMENT = qr/\((?:new|gone|(?:(?:new|mode) [\+\-][lx]))\)/;
+my $DIFFSTAT_COMMENT =
+	qr/(?: *\((?:new|gone|(?:(?:new|mode) [\+\-][lx]))\))? *\z/s;
+my $NULL_TO_BLOB = qr/^(index $OID_NULL\.\.)($OID_BLOB)\b/ms;
+my $BLOB_TO_NULL = qr/^index ($OID_BLOB)(\.\.$OID_NULL)\b/ms;
+my $BLOB_TO_BLOB = qr/^index ($OID_BLOB)\.\.($OID_BLOB)/ms;
+my $EXTRACT_DIFFS = qr/(
+		(?:	# begin header stuff, don't capture filenames, here,
+			# but instead wait for the --- and +++ lines.
+			(?:^diff\x20--git\x20$FN\x20$FN$LF)
+
+			# old mode || new mode || copy|rename|deleted|...
+			(?:^[a-z]$ANY+$LF)*
+		)? # end of optional stuff, everything below is required
+		^index\x20($OID_BLOB)\.\.($OID_BLOB)$ANY*$LF
+		^---\x20($FN)$LF
+		^\+{3}\x20($FN)$LF)/msx;
+my $IS_OID = qr/\A$OID_BLOB\z/s;
 
 # link to line numbers in blobs
 sub diff_hunk ($$$$) {
@@ -62,7 +78,7 @@ sub anchor0 ($$$$) {
 	# So only do best-effort handling of renames for common cases;
 	# which works well in practice. If projects put "=>", or trailing
 	# spaces in filenames, oh well :P
-	$fn =~ s/(?: *$DIFFSTAT_COMMENT)? *\z//so;
+	$fn =~ s/$DIFFSTAT_COMMENT//;
 	$fn =~ s/{(?:.+) => (.+)}/$1/ or $fn =~ s/.* => (.+)/$1/;
 	$fn = git_unquote($fn);
 
@@ -104,8 +120,8 @@ sub anchor1 ($$) {
 	$ok ? "<a\nhref=#i$attr\nid=$attr>diff</a> --git" : undef
 }
 
-sub diff_header ($$$$) {
-	my ($dst, $x, $ctx, $top) = @_;
+sub diff_header ($$$) {
+	my ($x, $ctx, $top) = @_;
 	my (undef, undef, $pa, $pb) = splice(@$top, 0, 4); # ignore oid_{a,b}
 	my $spfx = $ctx->{-spfx};
 	my $dctx = { spfx => $spfx };
@@ -132,27 +148,28 @@ sub diff_header ($$$$) {
 
 	# no need to capture oid_a and oid_b on add/delete,
 	# we just linkify OIDs directly via s///e in conditional
-	if (($$x =~ s/^(index $OID_NULL\.\.)($OID_BLOB)\b/
-			$1 . oid($dctx, $spfx, $2)/emos) ||
-		($$x =~ s/^index ($OID_BLOB)(\.\.$OID_NULL)\b/
-			'index ' . oid($dctx, $spfx, $1) . $2/emos)) {
-	} elsif ($$x =~ /^index ($OID_BLOB)\.\.($OID_BLOB)/mos) {
+	if (($$x =~ s/$NULL_TO_BLOB/$1 . oid($dctx, $spfx, $2)/e) ||
+		($$x =~ s/$BLOB_TO_NULL/
+			'index ' . oid($dctx, $spfx, $1) . $2/e)) {
+	} elsif ($$x =~ $BLOB_TO_BLOB) {
 		# modification-only, not add/delete:
 		# linkify hunk headers later using oid_a and oid_b
 		@$dctx{qw(oid_a oid_b)} = ($1, $2);
 	} else {
 		warn "BUG? <$$x> had no ^index line";
 	}
-	$$x =~ s!^diff --git!anchor1($ctx, $pb) // 'diff --git'!emos;
+	$$x =~ s!^diff --git!anchor1($ctx, $pb) // 'diff --git'!ems;
+	my $dst = $ctx->{obuf};
 	$$dst .= qq(<span\nclass="head">);
 	$$dst .= $$x;
 	$$dst .= '</span>';
 	$dctx;
 }
 
-sub diff_before_or_after ($$$) {
-	my ($dst, $ctx, $x) = @_;
+sub diff_before_or_after ($$) {
+	my ($ctx, $x) = @_;
 	my $linkify = $ctx->{-linkify};
+	my $dst = $ctx->{obuf};
 	for my $y (split(/(^---\n)/sm, $$x)) {
 		if ($y =~ /\A---\n\z/s) {
 			$$dst .= "---\n"; # all HTML is "\r\n" => "\n"
@@ -171,32 +188,34 @@ sub diff_before_or_after ($$$) {
 }
 
 # callers must do CRLF => LF conversion before calling this
-sub flush_diff ($$$) {
-	my ($dst, $ctx, $cur) = @_;
-
-	my @top = split(/(
-		(?:	# begin header stuff, don't capture filenames, here,
-			# but instead wait for the --- and +++ lines.
-			(?:^diff\x20--git\x20$FN\x20$FN$LF)
+sub flush_diff ($$) {
+	my ($ctx, $cur) = @_;
 
-			# old mode || new mode || copy|rename|deleted|...
-			(?:^[a-z]$ANY+$LF)*
-		)? # end of optional stuff, everything below is required
-		^index\x20($OID_BLOB)\.\.($OID_BLOB)$ANY*$LF
-		^---\x20($FN)$LF
-		^\+{3}\x20($FN)$LF)/smxo, $$cur);
+	my @top = split($EXTRACT_DIFFS, $$cur);
 	$$cur = undef;
 
 	my $linkify = $ctx->{-linkify};
+	my $dst = $ctx->{obuf};
 	my $dctx; # {}, keys: Q, oid_a, oid_b
 
 	while (defined(my $x = shift @top)) {
 		if (scalar(@top) >= 4 &&
-				$top[1] =~ /\A$OID_BLOB\z/os &&
-				$top[0] =~ /\A$OID_BLOB\z/os) {
-			$dctx = diff_header($dst, \$x, $ctx, \@top);
+				$top[1] =~ $IS_OID &&
+				$top[0] =~ $IS_OID) {
+			$dctx = diff_header(\$x, $ctx, \@top);
 		} elsif ($dctx) {
 			my $after = '';
+
+			# Quiet "Complex regular subexpression recursion limit"
+			# warning.  Perl will truncate matches upon hitting
+			# that limit, giving us more (and shorter) scalars than
+			# would be ideal, but otherwise it's harmless.
+			#
+			# We could replace the `+' metacharacter with `{1,100}'
+			# to limit the matches ourselves to 100, but we can
+			# let Perl do it for us, quietly.
+			no warnings 'regexp';
+
 			for my $s (split(/((?:(?:^\+[^\n]*\n)+)|
 					(?:(?:^-[^\n]*\n)+)|
 					(?:^@@ [^\n]+\n))/xsm, $x)) {
@@ -222,9 +241,9 @@ sub flush_diff ($$$) {
 					$$dst .= $linkify->to_html($s);
 				}
 			}
-			diff_before_or_after($dst, $ctx, \$after) unless $dctx;
+			diff_before_or_after($ctx, \$after) unless $dctx;
 		} else {
-			diff_before_or_after($dst, $ctx, \$x);
+			diff_before_or_after($ctx, \$x);
 		}
 	}
 }