Adapting gzip content

There is a gzip eCAP adapter at http://code.google.com/p/squid-ecap-gzip/

You can uncompress and compress gzip content by chunks. The above eCAP adapter appears to do that, and we have done that elsewhere (e.g., Web Polygraph benchmark and various Traffic Spicer ICAP server adapters). I believe our approach did not involve so many low-level buffer manipulations as found in the above eCAP adapter (we used public gzip APIs instead), but there may be advantages to this low-level manipulation. Both the above eCAP adapter and Polygraph licenses allow free reuse of the code.

Except for header/footer manipulation which I find rather annoying, the [de]compression code itself is not complex, but keep in mind that you cannot, in general, decompress the entire chunk you have just received from the host application -- you have to "stream" incoming data through the decompressing code which may need more data to finish decompressing a chunk than is currently available. On the other hand, accumulating the entire response body is usually not necessary and can be harmful if the body is large.

Finally, please keep in mind that there are other encodings besides identity and gzip, some of which are not standard.

Thanks Alex. I looked at the Web Polygraph code available online and I couldn't find where any gzip manipulation was happening, could you point me in the right direction?

Thanks Alex Rousskov, that solved my question.

 http://www.web-polygraph.org/docs/userman/compression.html is only Compression . is there any way to decompress chunks ?

Yes, there is a way to decompress chunks. Public Polygraph does not contain such examples (because Polygraph does not need to decompress) but you can find them elsewhere.

when I try to uncompress the gzip chunks I face segmentation fault . Would you please help me to find an example ?

Sorry, I cannot help at that level.