regex101: PCRE (Legacy) Match Workspace

Non-Capturing Group(?:(?<=\n)|(?<=\A))

1st Alternative(?<=\n)

Positive Lookbehind(?<=\n)

\na line-feed (newline) character (ASCII 10)

2nd Alternative(?<=\A)

Positive Lookbehind(?<=\A)

\Astart of string

Comment: Necessarily at the begining of a new line or start of string

Group leading_space(?<leading_space>[ ]{0,3})

Character Class[ ]{0,3}

{0,3}matches between 0 and 3 characters from the set below:

the literal (32₁₀ / 40₈ / 20₁₆)

Group tag_all(?<tag_all> (?: (?<div_open> <(?<tag_name>\S+) (?<tag_attributes>(?&tag_attr))* [^\>]* >\n* ) (?<content>.+?) \n* (?<div_close> (?(<leading_space>) # If leading spaces were found (?: (?<=\n)\g{leading_space} # Either there is a symmetry in leading space for open and closing tag | (?:(?<=\S)[[:blank:]\h]*) # or the closing tag is on the same line with preceding data ) | (?=<\/\g{tag_name}>) # No leading space, so we don't expect anything before the closing tag other than what has already been caught in the 'content' ) <\/\g{tag_name}> ) [[:blank:]\h]*\n ) | (?:  ) | (?: < [[:blank:]\h\v]* (?<tag_name>[a-zA-Z0-9][\w\-]+) (?<tag_attributes>(?&tag_attr))* [[:blank:]\h\v]* \/? [[:blank:]\h\v]* > ) )

1st Alternative (?: (?<div_open> <(?<tag_name>\S+) (?<tag_attributes>(?&tag_attr))* [^\>]* >\n* ) (?<content>.+?) \n* (?<div_close> (?(<leading_space>) # If leading spaces were found (?: (?<=\n)\g{leading_space} # Either there is a symmetry in leading space for open and closing tag | (?:(?<=\S)[[:blank:]\h]*) # or the closing tag is on the same line with preceding data ) | (?=<\/\g{tag_name}>) # No leading space, so we don't expect anything before the closing tag other than what has already been caught in the 'content' ) <\/\g{tag_name}> ) [[:blank:]\h]*\n )

Non-Capturing Group(?: (?<div_open> <(?<tag_name>\S+) (?<tag_attributes>(?&tag_attr))* [^\>]* >\n* ) (?<content>.+?) \n* (?<div_close> (?(<leading_space>) # If leading spaces were found (?: (?<=\n)\g{leading_space} # Either there is a symmetry in leading space for open and closing tag | (?:(?<=\S)[[:blank:]\h]*) # or the closing tag is on the same line with preceding data ) | (?=<\/\g{tag_name}>) # No leading space, so we don't expect anything before the closing tag other than what has already been caught in the 'content' ) <\/\g{tag_name}> ) [[:blank:]\h]*\n )

Group div_open(?<div_open> <(?<tag_name>\S+) (?<tag_attributes>(?&tag_attr))* [^\>]* >\n* )

<the literal text <

Group tag_name(?<tag_name>\S+)

\S+any non-whitespace character, repeated at least once

Group tag_attributes(?<tag_attributes>(?&tag_attr))*

*repeats the group contents below any number of times:

(?&tag_attr)calls group tag_attr as a subroutine

Negated Character Class[^\>]*

*matches any number of characters outside the set below:

\>the literal > (62₁₀ / 76₈ / 3E₁₆)

>the literal text >

\n*a line-feed (newline) character (ASCII 10), repeated any number of times

Group content(?<content>.+?)

.+?any character, repeated at least once

\n*a line-feed (newline) character (ASCII 10), repeated any number of times

Group div_close(?<div_close> (?(<leading_space>) # If leading spaces were found (?: (?<=\n)\g{leading_space} # Either there is a symmetry in leading space for open and closing tag | (?:(?<=\S)[[:blank:]\h]*) # or the closing tag is on the same line with preceding data ) | (?=<\/\g{tag_name}>) # No leading space, so we don't expect anything before the closing tag other than what has already been caught in the 'content' ) <\/\g{tag_name}> )

Condition(?(<leading_space>) # If leading spaces were found (?: (?<=\n)\g{leading_space} # Either there is a symmetry in leading space for open and closing tag | (?:(?<=\S)[[:blank:]\h]*) # or the closing tag is on the same line with preceding data ) | (?=<\/\g{tag_name}>) # No leading space, so we don't expect anything before the closing tag other than what has already been caught in the 'content' )

Uses the first branch if group leading_space matched

If the condition is true, match this branch: # If leading spaces were found (?: (?<=\n)\g{leading_space} # Either there is a symmetry in leading space for open and closing tag | (?:(?<=\S)[[:blank:]\h]*) # or the closing tag is on the same line with preceding data )

Comment: If leading spaces were found

Non-Capturing Group(?: (?<=\n)\g{leading_space} # Either there is a symmetry in leading space for open and closing tag | (?:(?<=\S)[[:blank:]\h]*) # or the closing tag is on the same line with preceding data )

1st Alternative (?<=\n)\g{leading_space} # Either there is a symmetry in leading space for open and closing tag

Positive Lookbehind(?<=\n)

\na line-feed (newline) character (ASCII 10)

\g{leading_space}the text saved by group leading_space

Comment: Either there is a symmetry in leading space for open and closing tag

2nd Alternative (?:(?<=\S)[[:blank:]\h]*) # or the closing tag is on the same line with preceding data

Non-Capturing Group(?:(?<=\S)[[:blank:]\h]*)

Positive Lookbehind(?<=\S)

\Sany non-whitespace character

Character Class[[:blank:]\h]*

*matches any number of characters from the set below:

[:blank:]a whitespace character, excluding line breaks [ \t]

\hany horizontal whitespace character

Comment: or the closing tag is on the same line with preceding data

Otherwise, match this branch: (?=<\/\g{tag_name}>) # No leading space, so we don't expect anything before the closing tag other than what has already been caught in the 'content'

<the literal text <

\/the literal / (47₁₀ / 57₈ / 2F₁₆)

\g{tag_name}the text saved by group tag_name

> the literal text >

Character Class[[:blank:]\h]*

\na line-feed (newline) character (ASCII 10)

2nd Alternative (?:  )

3rd Alternative (?: < [[:blank:]\h\v]* (?<tag_name>[a-zA-Z0-9][\w\-]+) (?<tag_attributes>(?&tag_attr))* [[:blank:]\h\v]* \/? [[:blank:]\h\v]* > )

Definition Group(?(DEFINE) (?<tag_attr> (?: [[:blank:]\h]* [\w\-]+ [[:blank:]\h]* = [^\"\'[:blank]\h]+ [[:blank:]\h]* ) | (?: [[:blank:]\h]* [\w\-]+ [[:blank:]\h]* = [[:blank:]\h]* (?<quote>["']) (.*?) \g{quote} [[:blank:]\h]* ) ) )

Global pattern flags

g modifier: global. Finds all matches instead of stopping after the first

x modifier: extended. Ignores unescaped whitespace and comments outside character classes

s modifier: single line. Allows a dot to match line terminators

m modifier: multiline. Causes ^ and $ to match the start and end of each line, not only the start and end of the string

J modifier: Allow duplicate subpattern names

25 tokens

Regular Expression

(?:(?<=\n)|(?<=\A))                # Necessarily at the begining of a new line or start of string
(?<leading_space>[ ]{0,3})
(?<tag_all>
    (?:
        (?<div_open>
            <(?<tag_name>\S+)
                (?<tag_attributes>(?&tag_attr))*
                [^\>]*
            >\n*
        )
        (?<content>.+?)
        \n*
        (?<div_close>
            (?(<leading_space>)                     # If leading spaces were found
                (?:
                    (?<=\n)\g{leading_space}              # Either there is a symmetry in leading space for open and closing tag
                    |
                    (?:(?<=\S)[[:blank:]\h]*)       # or the closing tag is on the same line with preceding data
                )
                |
                (?=<\/\g{tag_name}>)                # No leading space, so we don't expect anything before the closing tag other than what has already been caught in the 'content'
            )
            <\/\g{tag_name}>
        )
        [[:blank:]\h]*\n
    )
    |
    (?:
        <!--[[:blank:]\h\v]*(?<html_comment>.*?)[[:blank:]\h\v]*-->
    )
    |
    (?:
        <
            [[:blank:]\h\v]*
            (?<tag_name>[a-zA-Z0-9][\w\-]+)
            (?<tag_attributes>(?&tag_attr))*
            [[:blank:]\h\v]*
            \/?
            [[:blank:]\h\v]*
        >
    )
)
(?(DEFINE)
    (?<tag_attr>
        (?:
            [[:blank:]\h]*
            [\w\-]+
            [[:blank:]\h]*
            =
            [^\"\'[:blank]\h]+
            [[:blank:]\h]*
        )
        |
        (?:
            [[:blank:]\h]*
            [\w\-]+
            [[:blank:]\h]*
            =
            [[:blank:]\h]*
            (?<quote>["'])
            (.*?)
            \g{quote}
            [[:blank:]\h]*
        )
    )
)

New

<abbr title="`first backtick!">SB</abbr> 

<abbr title="`second backtick!">SB</abbr>

<table>
<tr><td markdown="block">test _emphasis_ (block)</td></tr>
</table>

## More complicated

<table>
<tr><td markdown="1">
* this is _not_ a list item</td></tr>
<tr><td markdown="span">
* this is _not_ a list item</td></tr>
<tr><td markdown="block">
* this _is_ a list item
</td></tr>
</table>

Extraction

<abbr title="`first backtick!">SB</abbr> 

<abbr title="`second backtick!">SB</abbr>

<table>
<tr><td markdown="block">test _emphasis_ (block)</td></tr>
</table>

## More complicated

<table>
<tr><td markdown="1">
* this is _not_ a list item</td></tr>
<tr><td markdown="span">
* this is _not_ a list item</td></tr>
<tr><td markdown="block">
* this _is_ a list item
</td></tr>
</table>