Re: [xml] problem with gzip decoding in nanohttp

[Daniel Veillard <veillard redhat com>]

On Thu, Jan 26, 2006 at 04:53:40PM +0100, Liron wrote:
> Hello,

  Hi. Could you get your mail agent to format lines ?

> I'm using nanohttp for simple http requests and I've noticed that
> there're several bugs in the code concerning the use of zlib.

  What bugs ?

> It seems
that the only function that supports inflating gzip is xmlNanoHTTPRead
but the problem is that it's *not the only function that reads bytes
from the network*.

  it's the only function which read the HTTP payload as far as I understand !

> Also, when libxml2 is compiled with zlib support it'll send the gzip headers by default, there's no way to 
> control it which means that I can't really avoid this bug unless I change my code to use xmlNanoHTTPRead 
> instead of my current method (xmlNanoHTTPFetch).

  I take patches then

> I also wanted to ask why there're so many different functions to do the same thing (such as reading from 
> the network). Wouldn't it be better to use one generic function for each step which'll make it easier to 
> control such bugs in the future?

  Explain please, what 'so many functions' ?
Removing publicly availble function is not possible for API and ABI
compatibility, but code cleanup may make sense.

Daniel


-- 
Daniel Veillard      | Red Hat http://redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/