[Grub-dev] how do i enter the proxy information in the c# client for windows?

[Bartek Jasicki]

2008-08-12, 20:32:35
ab <spam at abittner.de> wrote:

> about the proxy and cache: arent there some sort of http header tags
> to avoid caching in a proxy (only pass through) or some things like
> that. timestamps or such mechanisms?
> 

If i good find, yes, there is Pragma (HTTP/1.0) and Cache-Control
(HTTP/1.1). At end of mail i write more about history and future of
proxy support in C# client.

> > On Windows if i good think, client use Internet Explorer settings
> > (im not sure, but probably Mozilla Firefox use this same setting?),
> > on Unix
> 
> internetexploder stores stuff as most windows/native apps in the 
> registry. firefox can import during install/setup from the windows 
> registry (ie settings), but stores its own configuration data in
> files such as prefs.js and such. it uses these files during runtime.
> i guess its a pretty complicated problem what settings to use on what
> systems and so on never mind.
> 

Especially if developer don't have access to test it on Windows
systems ;) With this i need help.

> > systems it read from http_proxy environment variable. 
> > To set proxy in client, in HTTP proxy server you must write url to
> > proxy (example: http://proxy.com ) in Port - proxy port.
> 
> there are also applications also for windows that also look for and
> use the environment variables such as http_proxy, ftp_proxy,
> https_proxy and so on. the gnu win32 (wget, curl and many others) are
> such examples.
> 

I based my opinion about proxy setting on Windows systems on .NET
documentation (unfortunately, Mono documentation is still "under
development"). So, i'm not sure, that Mono can use this variables on
systems other than Unix.

> then there is that problem in various platforms for applications on
> how to behave if multiple sources of configuration with contradicting 
> information is present.
> 
> whatever. i will switch to direct access then. i am still wondering
> why it would make sense to upload the arc result file through a proxy
> but to fetch/crawl directly.
> 
> whenever you can crawl directly where is the advantage or the reason
> to use the proxy for pushing the arc resultfile?
> 
> cheers.

Ok, short explanation about status of proxy support in Grub C# client.
I was made this option some time ago mainly for test, thus
custom proxy don't support even proxy login for now - and for this
reason, it was added only for upload. And later (yes, my mistake) i
simply forgot about this ;) But probably i back to work on this option
before version 0.8 (with add this headers (both) to avoid problems
with cache).
For current development version (0.7.6) main goals are drop a memory
usage by client and (if this been possible) drop Mono dependence
(include necessary libraries with client) so probably back to work on
proxy support can take some time.

Bartek