I was just reading over on cf.Objective() about how it's important to use Content-Length when streaming files to the client. Without it, the browser knows how much it has downloaded so far, but not necessarily how much it will have to download in total (I guess it goes till some sort of End Of File marker?). I can understand the need for this fix - there is perhaps nothing more frustrating than having no idea how much file you have left to download.
The problem is that while I sometimes stream actual files, it is more common that I stream binary data using ColdFusion CFContent's Variable attribute. Jared only covered the File attribute, so I wasn't sure what to do in this case. But after a bit of testing, it looks like the browser is happy know the number of bytes in the byte array that represents the binary object:
- <!--- Kill extra output. --->
- <!--- Read in the image binary. --->
- file="#ExpandPath( './test_girl.jpg' )#"
- Set the content disposition and give a suggested
- file name for the client.
- value="attachment; filename=test_girl.jpg"
- Set the content length. This is the length of the
- byte array that represenets the binary image object.
- value="#ArrayLen( binImage )#"
- Set the content type and then stream the binary
- image to the client.
This downloads the file nicely, and when I check the headers in FireBug, the content length is set correctly. Thanks Jared, good tip.
Also, as a funny little thing, if I increase the content-length by doing something like this:
- #(ArrayLen( binImage ) + 1)#
... FireFox will actually never stop trying to download the image! I guess it just won't give up hope on that last byte :) So, I guess the moral of the story is that if you are going to provide a content length to the browser, make sure that it is correct otherwise you will do more damage than good.
Looking For A New Job?
Cool tip! I'll definitely remember this when I start working on an app in a few weeks that will be uploading/downloading files.
Cool. I will be using that at work tomorrow.
This is good, but not perfect. Docs say:
"Note: This action reads the file into a variable in the local Variables scope. It is not intended for use with large files, such as logs, because they can bring down the server."
What if you allow users/customers to download files routinely in the megs.
I haven't found a good solution to be able to give the file size for large files. Any ideas?
To be honest, I don't deal with a lot of file security most of the time. Therefore, I usually just link to files directly on the public server which bypasses all the ColdFusion stuff and just lets IIS deal with the file transfer.
But I agree with the documentation - this is not a good idea for large files. If you need file security, I guess the best thing to do is copy the file to a temp directory and then forward the person to it.
This would be a good test:
Compare CF's native approach to using the java.io class, across multiple file sizes.
This was a huge help for me! Thank you so much for posting this!