-# Copyright (C) 2019 all contributors <meta@public-inbox.org>
+# Copyright (C) 2019-2020 all contributors <meta@public-inbox.org>
# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
#
# used by PublicInbox::View
# (or reconstruct) blobs.
package PublicInbox::ViewDiff;
+use 5.010_001;
use strict;
use warnings;
use base qw(Exporter);
our @EXPORT_OK = qw(flush_diff);
use URI::Escape qw(uri_escape_utf8);
-use PublicInbox::Hval qw(ascii_html to_attr from_attr);
+use PublicInbox::Hval qw(ascii_html to_attr);
use PublicInbox::Git qw(git_unquote);
-sub DSTATE_INIT () { 0 }
-sub DSTATE_STAT () { 1 }
-sub DSTATE_HEAD () { 2 } # /^diff --git /, /^index /, /^--- /, /^\+\+\+ /
-sub DSTATE_CTX () { 3 } # /^ /
-sub DSTATE_ADD () { 4 } # /^\+/
-sub DSTATE_DEL () { 5 } # /^\-/
-my @state2class = (
- '', # init
- '', # stat
- 'head',
- '', # ctx
- 'add',
- 'del'
-);
-
sub UNSAFE () { "^A-Za-z0-9\-\._~/" }
my $OID_NULL = '0{7,40}';
my $OID_BLOB = '[a-f0-9]{7,40}';
-my $PATH_A = '"?a/.+|/dev/null';
-my $PATH_B = '"?b/.+|/dev/null';
-
-sub to_html ($$) {
- $_[0]->linkify_1($_[1]);
- $_[0]->linkify_2(ascii_html($_[1]));
-}
+my $LF = qr!\n!;
+my $ANY = qr![^\n]!;
+my $FN = qr!(?:"?[^/\n]+/[^\n]+|/dev/null)!;
+
+# cf. git diff.c :: get_compact_summary
+my $DIFFSTAT_COMMENT =
+ qr/(?: *\((?:new|gone|(?:(?:new|mode) [\+\-][lx]))\))? *\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 ($$$$) {
- my ($dctx, $spfx, $ca, $cb) = @_;
- my $oid_a = $dctx->{oid_a};
- my $oid_b = $dctx->{oid_b};
-
- (defined($spfx) && defined($oid_a) && defined($oid_b)) or
- return "@@ $ca $cb @@";
+ my ($dst, $dctx, $ca, $cb) = @_;
+ my ($oid_a, $oid_b, $spfx) = @$dctx{qw(oid_a oid_b spfx)};
- my ($n) = ($ca =~ /^-(\d+)/);
- $n = defined($n) ? do { ++$n; "#n$n" } : '';
+ if (defined($spfx) && defined($oid_a) && defined($oid_b)) {
+ my ($n) = ($ca =~ /^-([0-9]+)/);
+ $n = defined($n) ? do { ++$n; "#n$n" } : '';
- my $rv = qq(@@ <a\nhref="$spfx$oid_a/s/$dctx->{Q}$n">$ca</a>);
+ $$dst .= qq(@@ <a\nhref="$spfx$oid_a/s/$dctx->{Q}$n">$ca</a>);
- ($n) = ($cb =~ /^\+(\d+)/);
- $n = defined($n) ? do { ++$n; "#n$n" } : '';
-
- $rv .= qq( <a\nhref="$spfx$oid_b/s/$dctx->{Q}$n">$cb</a> @@);
+ ($n) = ($cb =~ /^\+([0-9]+)/);
+ $n = defined($n) ? do { ++$n; "#n$n" } : '';
+ $$dst .= qq( <a\nhref="$spfx$oid_b/s/$dctx->{Q}$n">$cb</a> @@);
+ } else {
+ $$dst .= "@@ $ca $cb @@";
+ }
}
sub oid ($$$) {
defined($spfx) ? qq(<a\nhref="$spfx$oid/s/$dctx->{Q}">$oid</a>) : $oid;
}
-sub to_state ($$$) {
- my ($dst, $state, $new_state) = @_;
- $$dst .= '</span>' if $state2class[$state];
- $_[1] = $new_state;
- my $class = $state2class[$new_state] or return;
- $$dst .= qq(<span\nclass="$class">);
-}
+# returns true if diffstat anchor written, false otherwise
+sub anchor0 ($$$$) {
+ my ($dst, $ctx, $fn, $rest) = @_;
+
+ my $orig = $fn;
+
+ # normal git diffstat output is impossible to parse reliably
+ # without --numstat, and that isn't the default for format-patch.
+ # So only do best-effort handling of renames for common cases;
+ # which works well in practice. If projects put "=>", or trailing
+ # spaces in filenames, oh well :P
+ $fn =~ s/$DIFFSTAT_COMMENT//;
+ $fn =~ s/{(?:.+) => (.+)}/$1/ or $fn =~ s/.* => (.+)/$1/;
+ $fn = git_unquote($fn);
+
+ # long filenames will require us to walk backwards in anchor1
+ if ($fn =~ s!\A\.\.\./?!!) {
+ $ctx->{-long_path}->{$fn} = qr/\Q$fn\E\z/s;
+ }
-sub anchor0 ($$$$$) {
- my ($dst, $ctx, $linkify, $fn, $rest) = @_;
if (my $attr = to_attr($ctx->{-apfx}.$fn)) {
$ctx->{-anchors}->{$attr} = 1;
+ my $spaces = ($orig =~ s/( +)\z//) ? $1 : '';
$$dst .= " <a\nid=i$attr\nhref=#$attr>" .
- ascii_html($fn) . '</a>'.
- to_html($linkify, $rest);
+ ascii_html($orig) . '</a>' . $spaces .
+ $ctx->{-linkify}->to_html($rest);
return 1;
}
undef;
}
-sub anchor1 ($$$$$) {
- my ($dst, $ctx, $linkify, $pb, $s) = @_;
+# returns "diff --git" anchor destination, undef otherwise
+sub anchor1 ($$) {
+ my ($ctx, $pb) = @_;
my $attr = to_attr($ctx->{-apfx}.$pb) or return;
- my $line = to_html($linkify, $s);
- if (delete $ctx->{-anchors}->{$attr} && $line =~ s/^diff //) {
- $$dst .= "<a\nhref=#i$attr\nid=$attr>diff</a> ".$line;
- return 1;
+ my $ok = delete $ctx->{-anchors}->{$attr};
+
+ # unlikely, check the end of all long path names we captured:
+ unless ($ok) {
+ my $lp = $ctx->{-long_path} or return;
+ foreach my $fn (keys %$lp) {
+ $pb =~ $lp->{$fn} or next;
+
+ delete $lp->{$fn};
+ $attr = to_attr($ctx->{-apfx}.$fn) or return;
+ $ok = delete $ctx->{-anchors}->{$attr} or return;
+ last;
+ }
}
- undef
+ $ok ? "<a\nhref=#i$attr\nid=$attr>diff</a> --git" : undef
}
-sub flush_diff ($$$) {
- my ($dst, $ctx, $linkify) = @_;
- my $diff = $ctx->{-diff};
+sub diff_header ($$$$) {
+ my ($dst, $x, $ctx, $top) = @_;
+ my (undef, undef, $pa, $pb) = splice(@$top, 0, 4); # ignore oid_{a,b}
my $spfx = $ctx->{-spfx};
- my $state = DSTATE_INIT;
- my $dctx = { Q => '' }; # {}, keys: oid_a, oid_b, path_a, path_b
-
- foreach my $s (@$diff) {
- if ($s =~ /^---$/) {
- to_state($dst, $state, DSTATE_STAT);
- $$dst .= $s;
- } elsif ($s =~ /^ /) {
- # works for common cases, but not weird/long filenames
- if ($state == DSTATE_STAT &&
- $s =~ /^ (\S+)(\s+\|.*\z)/s) {
- anchor0($dst, $ctx, $linkify, $1, $2) and next;
- } elsif ($state2class[$state]) {
- to_state($dst, $state, DSTATE_CTX);
- }
- $$dst .= to_html($linkify, $s);
- } elsif ($s =~ /^-- $/) { # email signature begins
- $state == DSTATE_INIT or
- to_state($dst, $state, DSTATE_INIT);
- $$dst .= $s;
- } elsif ($s =~ m!^diff --git ($PATH_A) ($PATH_B)$!) {
- if ($state != DSTATE_HEAD) {
- my ($pa, $pb) = ($1, $2);
- to_state($dst, $state, DSTATE_HEAD);
- $pa = (split('/', git_unquote($pa), 2))[1];
- $pb = (split('/', git_unquote($pb), 2))[1];
- $dctx = {
- Q => "?b=".uri_escape_utf8($pb, UNSAFE),
- };
- if ($pa ne $pb) {
- $dctx->{Q} .=
- "&a=".uri_escape_utf8($pa, UNSAFE);
+ my $dctx = { spfx => $spfx };
+
+ # get rid of leading "a/" or "b/" (or whatever --{src,dst}-prefix are)
+ $pa = (split('/', git_unquote($pa), 2))[1] if $pa ne '/dev/null';
+ $pb = (split('/', git_unquote($pb), 2))[1] if $pb ne '/dev/null';
+ if ($pa eq $pb && $pb ne '/dev/null') {
+ $dctx->{Q} = "?b=".uri_escape_utf8($pb, UNSAFE);
+ } else {
+ my @q;
+ if ($pb ne '/dev/null') {
+ push @q, 'b='.uri_escape_utf8($pb, UNSAFE);
+ }
+ if ($pa ne '/dev/null') {
+ push @q, 'a='.uri_escape_utf8($pa, UNSAFE);
+ }
+ $dctx->{Q} = '?'.join('&', @q);
+ }
+
+ # linkify early and all at once, since we know the following
+ # subst ops on $$x won't need further escaping:
+ $$x = $ctx->{-linkify}->to_html($$x);
+
+ # no need to capture oid_a and oid_b on add/delete,
+ # we just linkify OIDs directly via s///e in conditional
+ if (($$x =~ s/$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'!ems;
+ $$dst .= qq(<span\nclass="head">);
+ $$dst .= $$x;
+ $$dst .= '</span>';
+ $dctx;
+}
+
+sub diff_before_or_after ($$$) {
+ my ($dst, $ctx, $x) = @_;
+ my $linkify = $ctx->{-linkify};
+ for my $y (split(/(^---\n)/sm, $$x)) {
+ if ($y =~ /\A---\n\z/s) {
+ $$dst .= "---\n"; # all HTML is "\r\n" => "\n"
+ } elsif ($y =~ /^ [0-9]+ files? changed, /sm) {
+ # ok, looks like a diffstat, go line-by-line:
+ for my $l (split(/^/m, $y)) {
+ if ($l =~ /^ (.+)( +\| .*\z)/s) {
+ anchor0($dst, $ctx, $1, $2) and next;
}
- anchor1($dst, $ctx, $linkify, $pb, $s) and next;
+ $$dst .= $linkify->to_html($l);
}
- $$dst .= to_html($linkify, $s);
- } elsif ($s =~ s/^(index $OID_NULL\.\.)($OID_BLOB)\b//o) {
- $$dst .= $1 . oid($dctx, $spfx, $2);
- $dctx = { Q => '' };
- $$dst .= to_html($linkify, $s) ;
- } elsif ($s =~ s/^index ($OID_BLOB)(\.\.$OID_NULL)\b//o) {
- $$dst .= 'index ' . oid($dctx, $spfx, $1) . $2;
- $dctx = { Q => '' };
- $$dst .= to_html($linkify, $s);
- } elsif ($s =~ /^index ($OID_BLOB)\.\.($OID_BLOB)/o) {
- $dctx->{oid_a} = $1;
- $dctx->{oid_b} = $2;
- $$dst .= to_html($linkify, $s);
- } elsif ($s =~ s/^@@ (\S+) (\S+) @@//) {
- $$dst .= '</span>' if $state2class[$state];
- $$dst .= qq(<span\nclass="hunk">);
- $$dst .= diff_hunk($dctx, $spfx, $1, $2);
- $$dst .= '</span>';
- $state = DSTATE_CTX;
- $$dst .= to_html($linkify, $s);
- } elsif ($s =~ m!^--- $PATH_A! ||
- $s =~ m!^\+{3} $PATH_B!) {
- # color only (no oid link)
- $state <= DSTATE_STAT and
- to_state($dst, $state, DSTATE_HEAD);
- $$dst .= to_html($linkify, $s);
- } elsif ($s =~ /^\+/) {
- if ($state != DSTATE_ADD && $state > DSTATE_STAT) {
- to_state($dst, $state, DSTATE_ADD);
- }
- $$dst .= to_html($linkify, $s);
- } elsif ($s =~ /^-/) {
- if ($state != DSTATE_DEL && $state > DSTATE_STAT) {
- to_state($dst, $state, DSTATE_DEL);
+ } else { # commit message, notes, etc
+ $$dst .= $linkify->to_html($y);
+ }
+ }
+}
+
+# callers must do CRLF => LF conversion before calling this
+sub flush_diff ($$$) {
+ my ($dst, $ctx, $cur) = @_;
+
+ my @top = split($EXTRACT_DIFFS, $$cur);
+ $$cur = undef;
+
+ my $linkify = $ctx->{-linkify};
+ my $dctx; # {}, keys: Q, oid_a, oid_b
+
+ while (defined(my $x = shift @top)) {
+ if (scalar(@top) >= 4 &&
+ $top[1] =~ $IS_OID &&
+ $top[0] =~ $IS_OID) {
+ $dctx = diff_header($dst, \$x, $ctx, \@top);
+ } elsif ($dctx) {
+ my $after = '';
+ for my $s (split(/((?:(?:^\+[^\n]*\n)+)|
+ (?:(?:^-[^\n]*\n)+)|
+ (?:^@@ [^\n]+\n))/xsm, $x)) {
+ if (!defined($dctx)) {
+ $after .= $s;
+ } elsif ($s =~ s/\A@@ (\S+) (\S+) @@//) {
+ $$dst .= qq(<span\nclass="hunk">);
+ diff_hunk($dst, $dctx, $1, $2);
+ $$dst .= $linkify->to_html($s);
+ $$dst .= '</span>';
+ } elsif ($s =~ /\A\+/) {
+ $$dst .= qq(<span\nclass="add">);
+ $$dst .= $linkify->to_html($s);
+ $$dst .= '</span>';
+ } elsif ($s =~ /\A-- $/sm) { # email sig starts
+ $dctx = undef;
+ $after .= $s;
+ } elsif ($s =~ /\A-/) {
+ $$dst .= qq(<span\nclass="del">);
+ $$dst .= $linkify->to_html($s);
+ $$dst .= '</span>';
+ } else {
+ $$dst .= $linkify->to_html($s);
+ }
}
- $$dst .= to_html($linkify, $s);
- # ignore the following lines in headers:
- } elsif ($s =~ /^(?:dis)similarity index/ ||
- $s =~ /^(?:old|new) mode/ ||
- $s =~ /^(?:deleted|new) file mode/ ||
- $s =~ /^(?:copy|rename) (?:from|to) / ||
- $s =~ /^(?:dis)?similarity index /) {
- $$dst .= to_html($linkify, $s);
+ diff_before_or_after($dst, $ctx, \$after) unless $dctx;
} else {
- $state <= DSTATE_STAT or
- to_state($dst, $state, DSTATE_INIT);
- $$dst .= to_html($linkify, $s);
+ diff_before_or_after($dst, $ctx, \$x);
}
}
- @$diff = ();
- $$dst .= '</span>' if $state2class[$state];
- undef;
}
1;