allow processing of some tags (e.g. `<br>`, `<span>`) inside `<pre>`/`<code>` blocks

[asdfzdfj]

Version(s) affected

5.1.1

Description

basically, when converting (possibly rendered) html <pre> or <code> block that contains some tags inside them, these extra tags will be included in the converted markdown output verbatim, which could results in the converted markdown that looks quite different from what the source markdown was or what the input html would renders to, and from what I could make out from <pre> and <code> docs, these tags does allow its content to contain some other tags.

some cases where this could happen:

• converting output from whatever html renderer that really, really likes using <br> for newline even in these blocks
• converting a code block that used <span> inside the block for the purpose of syntax highlighting

How to reproduce

note: all these examples were run with strip_tags option enabled, but otherwise using default converters and options

first case: <br> usage for newlines inside <pre><code>

input

(yes it's all in one line)

<pre><code># When I first wrote this regex I thought it was slick.<br># I still think that, but 2y after doing it the first time<br># it just hurt to look at.  So, /x modifier we go!<br><br>my @Set = map { [ split( m/\s*:\s*/, $_, 2 ) ] } $args =~ m/<br>    \s*         # ignore preceeding whitespace<br>    (           # begin capturing<br>     (?:        # grab characters we want<br>         \\.    # skip over escapes<br>         | <br>         [^;]   # or anything but a ; <br>     )+?        # ? greedyness hack lets the \s* actually match<br>    )           # end capturing<br>    \s*         # ignore whitespace between value and ; or end of line<br>    (?:         # stop anchor at ...<br>      ;         # semicolon<br>      |         # or<br>      $         # end of line<br>    ) <br>    \s*/gx;<br></code></pre>

expected

(no actual markdown source to compare so this is an approximation of what I expect to see)

When I first wrote this regex I thought it was slick.

I still think that, but 2y after doing it the first time

it just hurt to look at. So, /x modifier we go!

my @Set = map { [ split( m/\s:\s/, $_, 2 ) ] } $args =~ m/ \s # ignore preceeding whitespace ( # begin capturing (?: # grab characters we want \. # skip over escapes | [^;] # or anything but a ; )+? # ? greedyness hack lets the \s actually match ) # end capturing \s # ignore whitespace between value and ; or end of line (?: # stop anchor at ... ; # semicolon | # or $ # end of line ) \s/gx;

html-to-markdown output

When I first wrote this regex I thought it was slick.

# I still think that, but 2y after doing it the first time

# it just hurt to look at. So, /x modifier we go!



my @Set = map { [ split( m/\s:\s/, $_, 2 ) ] } $args =~ m/

\s # ignore preceeding whitespace

( # begin capturing

(?: # grab characters we want

. # skip over escapes

|

[^;] # or anything but a ;

)+? # ? greedyness hack lets the \s actually match

) # end capturing

\s # ignore whitespace between value and ; or end of line

(?: # stop anchor at ...

; # semicolon

| # or

$ # end of line

)

\s/gx;

second case: mixing <span> inside <pre>/<code>

input

<pre style="background-color:#ffffff;">
<span style="color:#323232;">while (number &gt; 1) {
</span><span style="color:#323232;">  number -= 2;
</span><span style="color:#323232;">}
</span><span style="color:#323232;">return number;
</span></pre>

expected

while (number > 1) { number -= 2; } return number;

html-to-markdown output

while (number > 1) { number -= 2; } return number;