Question

I need to match links, but only those who don’t begin with the string [image:x] where x is the link itself.

I need this in a preg_replace so that only links that fit will be changed, here is an example:

“a dog once ate my book http://mysite.com/dog.jpg and I had to go to
the store https://www.mysite.com/images/store.png and buy a new iPad
apple.com/iPad/iPadMini.jpg And there it was fun
[image:http://apple.com/iPad/images/iPads.com]”

would result as

“a dog once ate my book [image:http://mysite.com/dog.jpg] and I had to go to
the store [image:https://www.mysite.com/images/store.png] and buy a new iPad
[image:apple.com/iPad/iPadMini.jpg] And there it was fun
[image:http://apple.com/iPad/images/iPads.com]”

Notice the last link, as for now I get it doubled (the link is still picked up by preg_replace). like here (not the desired effect, I need it untouched):

...[image:apple.com/iPad/iPadMini.jpg] And there it was fun
[image:[image:http://apple.com/iPad/images/iPads.com]]“

Here’s my preg_replace as for now:

preg_replace('~(https?://)?[\w-]+(\.[\w-]+)+\.?(:\d+)?(\S*)\.(jpg|png|jpeg|bmp|gif)~','[[image:$0]]',$text);

Answer 1

You can use a negative lookbehind.

We don’t want it to start with ‘[image:’. We also don’t want to start in the middle of a word. this is prevented using \b. If the string starts with http:// it should be included in the match, so don’t match after http:// or https://.

preg_replace('~(?<!\[image:)(?<!http://)(?<!https://)(?<!\w\.)(https?://)?\b[\w-]+(\.[\w-]+)+\.?(:\d+)?(\S*)\.(jpg|png|jpeg|bmp|gif)~','[image:$0]',$text);

Check this PHP fiddle

Note: that also won’t replace a string with the final ] missing, like [image:example.com.