$bond_strength_to_go[$i] = $strength;
+ # Fix for case c001: be sure NO_BREAK's are enforced by later
+ # routines, except at a '?' because '?' as quote delimiter is
+ # deprecated.
+ if ( $strength >= NO_BREAK && $next_nonblank_type ne '?' ) {
+ $nobreak_to_go[$i] ||= 1;
+ }
+
DEBUG_BOND && do {
my $str = substr( $token, 0, 15 );
$str .= ' ' x ( 16 - length($str) );
=over 4
+=item B<Fix a bad line break choice at a slash>
+
+Random testing produced case c001 in which the following snipppet
+
+ ok $mi/(@mtime-1) >= 0.75 && $ai/(@atime-1) >= 0.75 &&
+ $ss/(@mtime+@atime) >= 0.2;
+
+when processed with these parameters
+
+ --maximum-line-length=20
+ --nowant-right-space=' / '
+ --want-break-before='* /'
+
+produced the following result
+
+ ok $mi
+ /( @mtime - 1 ) >=
+ 0.75
+ && $ai
+ /( @atime - 1 )
+ >= 0.75
+ && $ss
+ /( @mtime +
+ @atime ) >=
+ 0.2;
+
+using 'perl -cw' on this snippet gives a syntax error
+
+ syntax error at /tmp/issues.t line 5, near "/( "
+ (Might be a runaway multi-line // string starting on line 2)
+
+The error is due to perl's weird parsing rules near a possible indrect object.
+This is a situation where perltidy must ignore a user spacing and line break
+request. This should have been done but in this case a flag to prevent this
+was not being propagated to later stages of formatting. This has been fixed.
+
+20 Apr 2021.
+
=item B<Fix rare problem with -lp -wn>
Random testing produced case b1114 which gave unstable formatting with these
is(length(pack("j", 0)),
$Config{ivsize});
-Fixed 19 Apr 2021.
+Fixed 19 Apr 2021, eeeaf09.
=item B<Fix issue git#63>
my $fragment = $parser-> #parse_html_string
parse_balanced_chunk($I);
-Fixed 18 Apr 2021.
+Fixed 18 Apr 2021, c2030cf.
=item B<Avoid welding at sort/map/grep paren calls>