problems with updated version

[frequent]

Happy to see there's a Coldfusion port for AI. I set it it on localhost and it's not really working (I'm using CF8).

I have fixed the CF errors I found (missing a few semicolons, declaring var on top etc), but these seem to be the easy fixes.. If I ever get to work I will do a pull request (If not, I'm back to trying to fiddle in the php version).

[cfjedimaster]

To be clear though - I don't think CF8 compat issues are bugs. I wrote it for CF9 and I think that backporting to 8 would be a mistake (imo ;).

[frequent]

I understand. I'm still stuck on CF8, so I'll be backporting away.

[frequent]

@cfjedimaster - ok, got it to work. You are correct about backporting. It's lame... :-)

I found the following things to be tricky (not sure if CF8/CF9 related, or supposed to be like this, my first month of CF...)

1. request_uri sets the requested image path to "adaptive-images.cfm"

   I'm defaulting to fallback on Apache and use CGI.redirect_url

2. if no width/height parameters were passed structKeyExists(url,"width") seems to fail (not sure)

   I'm doing all the ?parameters in a if( !StructIsEmpty(url) ) { scale/crop } else{ regular }

3. on my localhost I'm ending up with a\lot\of\forward/backward/slashed/path/names.jpg

   I'm setting document root like so:

root = Replace( expandPath("/"), "\", "/", "all" );
    if ( Right(root, 1) EQ "/" )
        document_root = Left(root, Len(root)-1 );
    else
        document_root = root;

I would have a few questions, if you could spare some thoughts:

• Initial load times take time. Is this CF re-processing the images or my lame localhost? Can this be sped up?
• I'm trying this in a 10 image slider. If I delete the cache and refresh the page, the first couple of image fail, the rest works. Seems like a "traffic jam" inside the function (too many images requested). Any idea how to avoid this?
• Would this be better off in a cfc?
• If I wanted to do this with large image volumes, do you see any downside? I like that images are cached the first time a user requests them, but I'm trying to think this through for possible caveats.

A BIG THANKS for writing this up (and sparing me my first php session). I'll be playing around with your Resize options to see how to get the best quality & fastest load time)

[cfjedimaster]

"if no width/height parameters were passed structKeyExists(url,"width") seems to fail (not sure)" That makes no sense - if you can find out exactly what it was - let me know.

"on my localhost I'm ending up with a\lot\of\forward/backward/slashed/path/names.jpg"

But did they break? I always use / since it works with Windows or Unix. So on Windows, you may see a path with both \ and / in it, but if it works, it shouldn't matter.

"Initial load times take time. Is this CF re-processing the images or my lame localhost? Can this be sped up?" If you tweak the quality setting, yes. :) I see you saw that (last part of your comment)

"I'm trying this in a 10 image slider. If I delete the cache and refresh the page, the first couple of image fail, the rest works. Seems like a "traffic jam" inside the function (too many images requested). Any idea how to avoid this?"

Got a zip I can try?

"Would this be better off in a cfc?" I used a CFM just to make the URL a bit simpler, if I remember right anyway.

"If I wanted to do this with large image volumes, do you see any downside? I like that images are cached the first time a user requests them, but I'm trying to think this through for possible caveats."

You would possibly want to consider creating a script that would request the images in all possible sizes. Ie, a way to 'force' the creation so that images are - basically - precached.

[frequent]

Thanks for the infos. I'm trying to pull the slider page out in a zip file. I'll post the link here.

[frequent]

Sorry for getting back late.

While I managed to get it to work on my dev/localhost, production was a different story. I wanted to test AI in the slider and mainly use it for a product search. Slider worked ok, but the product search was a pain, because I needed to pull in and render images from user specified sources and with users potentially not sticking to image size recommendations.

I tested a little - both adaptive images and my own script pulling in and resizing just took too long, so the search was terribly slow to return results (waiting for image processing).

I eventually ditched the whole approach and am now offering users to auto-upload images to S3 when they import product data. Since this is not time critical, I can generate S,M,L,XL versions of each image. I added a log entry for generated images in the database, so in my product search I can just include this info in my search routine and request the S3 image if log enty has been made for the respective item.

Anyway: I only wanted to post these two links (picturefill, Git ). Maybe a nice alternative to AI. I'm playing around with this now vs. using the AI leftover cookie.