将所有类型的BBCODE图像链接替换为HTML图像链接

Question

I have this script for replacing BBCODE image tags to HTML tags

$text = "[img]https://somelink/2/3/pic.jpg[/img]
[img]https://somelink/2/3/pic.jpg[/img]
[img]https://somelink/2/3/pic.jpg[/img]";

echo preg_replace(
'~\[img\](https?://.*?\.(?:jpg|jpeg|gif|png|bmp))\[/img\]~s',
'<img src="$1" alt="" />',
$text);

This forks for links, which ends up whit some of given extensions: .jpg .jpeg ...

But if link are like: [img]https://somelink/2/3/pic.jpg ?w=bla [/img]

Then this regex pattern fails.

What would be appropriate pattern for all type of image links? I tried:

~\\[img\\](https?://.*?\\.(?:jpg|jpeg|gif|png|bmp).*)\\[/img\\]~Us

But this also not works

Answer 1

You can use

\[img\](https?://.*?\.(?:jpe?g|gif|png|bmp).*?)\[/img\]

• \\[img\\] - Matches [img]
• (https?://.*?\\.(?:jpe?g|gif|png|bmp).*?) - Matches URL containing any of above extensions
• \\[/img\\] - Matches [/img]

Demo

You can visualize regex here

Note:- as you're using U flag you can safely remove ? after quantifiers, ie

\[img\](https?://.*\.(?:jpe?g|gif|png|bmp).*)\[/img\]

Answer 2

If it is possible that the query string parameters can also contain [ or ] you could match the question mark \\? followed by matching until you encounter [/img] and make use of a possessive quantifier to prevent unnecessary backtracking

Explanation

\[img\](https?://.*?\.(?:jpe?g|gif|png|bmp)(?:\?(?:[^[]++|\[(?!/img\]))*+)?)\[/img]

• \\[img\\] Match [img]
• ( Capturing group
  • https?://.*? Match http with optional s, :// and 0+ times any char non greedy
  • \\.(?:jpe?g|gif|png|bmp) Match a dot and any of the listed options
  • (?: Non capturing group
    • \\? Match ?
    • (?:[^[]++|\\[(?!/img\\]))*+ Repeat 0+ times matching not [ or match [ when what is directly on the right is not /img]
  • )? Close non capturing group and make it optional
• ) Close capturing group
• \\[/img] Match [/img]

Regex demo