Regular expressions being dodgy

[ljm]

I'm not an expert when it comes to these beauties, and I've got a little bit of a problem with one in particular.

What I intended to do was use [minicode]preg_replace()[/minicode] to go through a database entry, and replace all the images (which are screenshots in the context of the site) with a link to their full size versions. I'm using TinyMCE for the text editor, in which I sacrificed strict XHTML output for admin panel user friendliness. Anyway, for images, it creates this:

<img src='img.jpg' alt='text' width='x' height='y' align='left' />

I want it to turn into this:

<a href='img-full.jpg' title='alt text'><img src='img.jpg' alt='text' width='x' height='y' align='left' /></a>

My regular expression is this:

'<img src=\"(.*?[^\.])\.(.*?)\" alt=\"(.*?)\" />'is

And the intended output is this:

<a href=\"\\1-full.\\2\" rel=\"lightbox\" title=\"\\3\"><img src=\"\\1.\\2\" alt=\"\\3\" /></a>

It works on some images, but not on others. It seems to ignore those that have alt text, but leaving the attribute empty works fine. The image tag is left untouched by the function, also, when really it should match that code above!

I think my regular expression is wrong, but I'm not sure how else I could change it (seeing as I want to remove the size attributes from the image, too).

Thanks in advance for any help!

[Shadowfiend]

Try making that alt="([^"]*)". I'm not sure why the question marks are there?

[filburt1]

Question marks = ungreedy.

[Shadowfiend]

Yeah, I know what the question marks do, I'm asking why use them here :-P I'm relatively sure they'll just slow things down unnecessarily, and may be having other unintended consequences.

You're also not matching the full image tag, since you aren't putting the width and height parameters in your regular expression, and your actual tags may have it in them; is that intentional?

[ljm]

I noticed that I wasn't matching them fully, so I changed it. Then everything stopped working.

I'm using [minicode]wordwrap()[/minicode] on the main body of text, with "\n" as the linebreak (to make it look nice in the source). Do you think that might have an adverse effect? The regexp doesn't work if I remove the single line modifier.

I put the question mark in the expressions because I used a book for reference, and that was in it.

I'll have a bit more of a fiddle, anyway. I want to do it this way because I'm trying to automate as many mucky bits of the site as I can. I don't want my system's intended users to be faffing about with links to make lightbox work on the images!

[Shadowfiend]

Gotcha. Why not word wrap after doing this replacement?

[rosland]

Just a regular expression suggestion:

$str = "<img src='image_name.jpg' alt='description' width='x' height='y' align='left' />";

$pattern = "/[\w|\s+]+='(.*)\.[a-z]{3}'[\w|\s+]+='(\w*)'.*/";

preg_match($pattern, $str, $res);

echo "<a href='$res[1]-full.jpg' title='$res[2]'>$str</a>"; 

[ljm]

Gotcha. Why not word wrap after doing this replacement?

I have no idea! :ermm:

I'll give that a go and what not and see if it all works. I've not tried anything yet.

Oh, and I like your suggestion Rosland, but I don't know how I'd do that when my 'string' is a huge article, not just the image tag.

Thanks to everyone for the help though!

[Shadowfiend]

You should be able to use [phpfunction]preg_replace[/phpfunction] with back-references (use \1 and \2 instead of $res[1] and $res[2]) to do it on a full string, I believe.

[rosland]

Oh, and I like your suggestion Rosland, but I don't know how I'd do that when my 'string' is a huge article, not just the image tag.

I don't know if this is of any help, but I embedded some image tags in a text to see if that's closer to what you're after.
The final 'print' command will output the original string, but with all the image tags altered to what you've described above.

$str = "<img src='image_name.jpg' alt='description' width='x' height='y' align='left' />
Lorem ipsum dolor sit amet, consectetuer adipiscing elit. Etiam fermentum semper nulla.
<img src='image_name2.jpg' alt='description2' width='x' height='y' align='left' /> 
Sed tincidunt pharetra metus.<img src='image_name3.jpg' alt='description3' width='x' 
height='y' align='left' />
Nam at turpis. Aliquam lacinia erat vel quam.
<img src='image_name4.jpg' alt='description4' width='x' height='y' align='left' />
Vivamus iaculis odio sed pede. Etiam urna.
<img src='image_name5.jpg' alt='description5' width='x' height='y' align='left' />
Curabitur enim eros, egestas eget, rhoncus ut, blandit eget, ipsum. Fusce id justo. 
Nullam semper felis sit amet urna.
Vivamus laoreet. Suspendisse potenti. Donec aliquam rhoncus lectus. Class aptent taciti sociosqu ad 
litora torquent per conubia nostra, per inceptos hymenaeos. <img src='image_name6.jpg' 
alt='description7' width='x' height='y' align='left' /> 
Quisque arcu est, dictum eget, 
sollicitudin adipiscing, adipiscing eget, arcu.";


$pattern = "/<img(.*)\/>/";


preg_match_all($pattern, $str, $res);

$mod=$res[0];


$pattern = "/[\w|\s+]+='(.*)\.[a-z]{3}'[\w|\s+]+='(\w*)'.*/";
$replace="a href='$1-full.jpg' title='$2'>".$res[0][$i]."</a>";


for($i=0; $i < count($mod); $i++)
{
    $alter[$i]=preg_replace($pattern, "a href='$1-full.jpg' title='$2'>".$res[0][$i]."</a>", $mod[$i]);
}

print str_replace($res[0], $alter, $str);