- Does anyone know of apache support for transfer-encoding? - Does anyone know of linux web applications that use TE:?
On Tue, Mar 25, 2003 at 05:44:21AM -0500, David Carter wrote: > If you're interested, I have some working perl code that does deflate > decompression. It does it at the application level, and needs to be moved > down into Net::HTTP and/or LWP. David, Yes, I would like to see that code. Last night I thought that the only thing I need to do is _not_ "_removeGzipHeader" on the deflate data streams. I can't be sure... until I see it working. > No need to teach mod_gzip deflate for testing - just find a site on the > internet that already emits "Content-encoding: deflate" & test with it. > Such as http://www.homedepot.com Very good point... thanks for the URL, I'll try that first. ;) > All commercial "http accelerators" I have looked at use content-encoding > rather than transfer-encoding. I think it has something to do with what > Internet Explorer supports, or perhaps even how well it supports one vs. the > other. It's been a couple of years since I worked with this extensively, so > the details are a little foggy. Before I looked at the headers from Net::HTTP I'd never seen 'TE:' (transfer-encoding support) advertised before... I don't see any apache modules that enable this transfer mode. I still don't understand "chunked" transfer mode issues, but it seems to be a way to keep-alive the TCP channel and do streamed encoding. Normal content-encoding needs to know the content-length of the compressed stream to send in the header or the connection needs to be closed for each request. If deflate doesn't have a header like gzip (containing the uncompressed stream length and CRC), deflate would be better to use for streamed content... since you won't need to have the complete stream compressed before you can start sending the first byte. If you want keep-alive to work you would either need to use chunked mode or compress and buffer all the data to use in Content-Length. I'm mystified by why the mod_gzip authors wrote support only for gzip and not deflate, it also appears that they did not use zlib, but have their own compression code. I wonder if they are working on TE support? I'm going to have to check out their mailing list too... Later, Mike Simons
