commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sale, Doug" <ds...@us.britannica.com>
Subject RE: HTTP Client: HttpMethod/HttpMethodBase patches
Date Mon, 14 Jan 2002 16:12:32 GMT
+1

> -----Original Message-----
> From: Waldhoff, Rodney [mailto:rwaldhof@us.britannica.com]
> Sent: Sunday, January 13, 2002 4:32 PM
> To: 'Otis Gospodnetic '; 'commons-dev@jakarta.apache.org '
> Subject: RE: HTTP Client: HttpMethod/HttpMethodBase patches
> 
> 
> By your own description--"for setting headers such as 
> 'User-Agent', which
> most often stay fixed during the whole life of an application 
> and do not
> change from request to request", I'd suggest that the "default header"
> aspects should live in HttpClient or some other user-agent 
> (browser-like)
> class.  That way the agent could apply the headers to every 
> HttpMethod it
> executes, rather than having to set up each method 
> independently with the
> set of defaults.
> 
> -----Original Message-----
> From: Otis Gospodnetic
> To: commons-dev@jakarta.apache.org
> Sent: 1/13/02 2:33 PM
> Subject: HTTP Client: HttpMethod/HttpMethodBase patches
> 
> Hello,
> 
> I added the two methods for setting/adding default request 
> headers for 
> setting headers such as 'User-Agent', which most often stay 
> fixed during
> 
> the whole life of an application and do not change from request to
> request.
> 
> I'm including the patch to HttpMethod (interface) and HttpMethodBase 
> (impl of that interface).
> 
> Originally I wanted to add the ability to set the default request 
> headers for 1) efficiency, and 2) simpler code.
> 
> However, now after I added them it seems to me that this is 
> actually not
> 
> the most efficient way of doing it (checking 2 HashMaps vs. 1 
> HashMap in
> 
>   5 of those add*RequestHeader methods).  It is probably more 
> efficient 
> to just call setRequestHeader(String, String) over and over, which is 
> what I wanted to avoid in the first place, thinking that it's
> inefficient.
> 
> On the other hand, the call that uses HTTP Client can be a 
> bit simpler 
> and cleaner because it won't have to call setRequestHeader(String, 
> String) over and over for the non-variable headers, between method 
> recycling .
> 
> Therefore, a question for all: what do you think, do these patches 
> improve things or not?
> Please let me know if they can or cannot be applied to the 
> versions in 
> the repository.  At this point I'm fine either way :)
> 
> Thanks,
> Otis
>  <<HttpMethod.patch>>  <<HttpMethodBase.patch>>  <<ATT98116.txt>>

> 

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message