Disable hard line breaks using two spaces in the default syntax.

Changes

@@ -1,5 +1,10 @@
 Revision history for pmarkdown and the Markdown::Perl module.
 
+1.03 - ??
+
+ - Disable hard line breaks using two spaces in our default syntax because they
+   are invisible in the source file.
+
 1.02 - 2024-04-06
 
  - Improvement to the support of the original markdown syntax. This is probably

lib/Markdown/Perl.pm

@@ -73,16 +73,30 @@ sub convert {  ## no critic (RequireArgUnpacking)
   my $md = \(shift @_);  # Taking a reference to avoid copying the input. is it useful?
   $this->SUPER::set_options(local_options => @_);
 
-  my $parser = Markdown::Perl::BlockParser->new($this, $md);
-
   # TODO: introduce an HtmlRenderer object that carries the $linkrefs states
   # around (instead of having to pass it in all the calls).
-  my ($linkrefs, $blocks) = $parser->process();
+  my ($blocks, $linkrefs) = $this->_parse($md);
   my $out = $this->_emit_html(0, 'root', $linkrefs, @{$blocks});
   $this->{local_options} = {};
   return $out;
 }
 
+# This is an internal call for now because the structure of the parse tree is
+# not defined.
+# Note that while convert() takes care not to copy the md argument, this is not
+# the case of this method, however, it can receive a scalar ref instead of a
+# scalar, to avoid the copy.
+# TODO: create a BlockTree class and document it, then make this be public.
+sub _parse {
+  my ($this, $md_or_ref) = &_get_this_and_args;  ## no critic (ProhibitAmpersandSigils)
+  my $md = ref($md_or_ref) ? $md_or_ref : \$md_or_ref;
+
+  my $parser = Markdown::Perl::BlockParser->new($this, $md);
+  my ($linkrefs, $blocks) = $parser->process();
+  return ($blocks, $linkrefs) if wantarray;
+  return $blocks;
+}
+
 sub _render_inlines {
   my ($this, $linkrefs, @lines) = @_;
   return Markdown::Perl::Inlines::render($this, $linkrefs, @lines);

lib/Markdown/Perl/BlockParser.pm

@@ -516,6 +516,8 @@ sub _do_fenced_code_block {
 sub _do_html_block {
   my ($this) = @_;
   # HTML blocks can interrupt a paragraph.
+  # TODO: add an option so that they don’t interrupt a paragraph (make it be
+  # the default?).
   # TODO: PERF: test that $l =~ m/^ {0,3}</ to short circuit all these regex.
   my $html_end_condition;
   if ($l =~ m/ ^\ {0,3} < (?:pre|script|style|textarea) (?:\ |\t|>|$) /x) {

lib/Markdown/Perl/Inlines.pm

@@ -472,7 +472,13 @@ sub process_whitespaces {
     next unless $n->{type} eq 'text';
     # TODO: add tests for the fact that we don’t want hard break at the end of a
     # paragraph.
-    my @hard_breaks = split(/(?: {2,}|\\)\n(?=.) */s, $n->{content}, -1);
+    my $re;
+    if ($that->get_two_spaces_hard_line_breaks) {
+      $re = qr/(?: {2,}|\\)\n(?=.) */s;
+    } else {
+      $re = qr/\\\n(?=.) */s;
+    }
+    my @hard_breaks = split($re, $n->{content}, -1);
     for (my $j = 0; $j < @hard_breaks; $j++) {
       # $hard_breaks[$j] = '' unless defined($hard_breaks[$j]);
       $hard_breaks[$j] =~ s/^ +// if !$not_root && $i == 0 && $j == 0;

lib/Markdown/Perl/Options.pm

@@ -601,6 +601,19 @@ _make_option(
 
 =pod
 
+=head3 B<two_spaces_hard_line_breaks> I<(boolean, default: false)>
+
+Controls if a hard line break can be generated by ending a line with two spaces.
+
+=cut
+
+_make_option(
+  two_spaces_hard_line_breaks => 0,
+  _boolean,
+  (markdown => 1, cmark => 1, github => 1));
+
+=pod
+
 =head2 Options controlling the rendering of inline elements
 
 =head3 B<html_escaped_characters> I<(character_class)>

t/202-basic-inline-rendering.t

@@ -14,7 +14,7 @@ is(run("abc\n"), "<p>abc</p>\n", 'line2');
 is(run(" abc "), "<p>abc</p>\n", 'line3');
 
 is(run("abc\ndef\n"), "<p>abc\ndef</p>\n", 'soft_break');
-is(run("abc  \ndef\n"), "<p>abc<br />\ndef</p>\n", 'hard_break1');
+is(run("abc  \ndef\n", two_spaces_hard_line_breaks => 1), "<p>abc<br />\ndef</p>\n", 'hard_break1');
 is(run("abc\ndef  \n"), "<p>abc\ndef</p>\n", 'hard_break2');
 is(run("abc\\\ndef"), "<p>abc<br />\ndef</p>\n", 'hard_break3');
 is(run("abc\\\\\ndef"), "<p>abc\\\ndef</p>\n", 'hard_break4');

t/400-opt-two_spaces_hard_line_breaks.t

@@ -0,0 +1,15 @@
+use strict;
+use warnings;
+use utf8;
+
+use Markdown::Perl 'convert';
+use Test2::V0;
+
+is(convert("foo\nbar", two_spaces_hard_line_breaks => 0), "<p>foo\nbar</p>\n", 'soft_opt_off');
+is(convert("foo\nbar", two_spaces_hard_line_breaks => 1), "<p>foo\nbar</p>\n", 'soft_opt_on');
+is(convert("foo\\\nbar", two_spaces_hard_line_breaks => 0), "<p>foo<br />\nbar</p>\n", 'hard_with_backslash_opt_off');
+is(convert("foo\\\nbar", two_spaces_hard_line_breaks => 1), "<p>foo<br />\nbar</p>\n", 'hard_with_backslash_opt_on');
+is(convert("foo  \nbar", two_spaces_hard_line_breaks => 0), "<p>foo\nbar</p>\n", 'hard_with_spaces_opt_off');
+is(convert("foo  \nbar", two_spaces_hard_line_breaks => 1), "<p>foo<br />\nbar</p>\n", 'hard_with_spaces_opt_on');
+
+done_testing;