Re: determine file-upload's tmp-filename

Michael Fischer wrote:

On Mon, 2008-01-28 at 15:34 -0500, Mr Webber wrote:

On Mon, 2008-01-28 at 12:17 -0600, Richard Lynch wrote:

On Sat, January 26, 2008 5:57 pm, Michael Fischer wrote:

hi there, is there a way to determine the tmp-filename of a file upload while the upload is still in progress? the tmp-file is stored in /tmp and it's name is something like PHPXXXXXXXX. what i would like to do is: i want to upload a file via a html-form and while the upload is in progress make repeatedly ajax-requests to a php-script on the server that replies the size of the tmp file (the amount of data that was already uploaded). So in this script i need to know what the tmp-filename is. or do you think this is a completely useless approach?

Google for PHP upload meter instead. That's probably how it works, more or less... I still think it's STOOPID to round-trip back and forth to the server to get an upload progress meter -- The browser developers should be providing you with some kind of progress notification system locally!

well, i agree with you - the browser should provide some sort of functionality to find out the amount of data already sent. but i don't know any browser that does. all the php upload meters that i found on google either require to patch php or use perl. I have not done this, but am inspired to do it now. My design is to develop my own meter by compare $FILES;'size'] to the actual size on disk as the file uploads; I would use some sort of JavaScript / Ajax service connection so that the upload meter is displayed in the current browsing window or an upload popup. $_FILES['userfile']['name'] $_FILES['userfile']['type'] $_FILES['userfile']['size'] $_FILES['userfile']['tmp_name'] $_FILES['userfile']['error']

two problems: - the $_FILES array becomes available when the file-upload is complete. that's too late.

exactly right!

- how do you determine the actual size on the disk of the uploaded file, while the upload is in progress?

he is probably talking about the amount of data that has been uploaded so far. The size of the tmp file would reflect that amount.

lg, Michi

I would create your own perl/cgi up-loader that runs like the perl version. I would run it from (x)inetd, or something similar. I have created a number of socket servers that run 24/7 on my server with no glitches. Because of the size of data and the connection duration, I would suggest forking the parent for every child connection that comes in. But then have that child process report what it needs to a database, or temp data file. The incoming TCP connection will report the payload size and from that you can then figure the rest out. You would have to handle all the data connections, creating the tmp_file and such, but it would all be possible. I have been thinking about doing this myself and soon. -- Jim Lucas "Some men are born to greatness, some achieve greatness, and some have greatness thrust upon them." Twelfth Night, Act II, Scene V by William Shakespeare