----- Original Message ----- From: "Jason M. Kusar" <[EMAIL PROTECTED]> To: "Louis Lau" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> Sent: Thursday, February 06, 2003 11:55 AM Subject: Re: [squid-users] Squid modification help
> First of all, have you looked at the programmers guide on the website? It's > a liitle out of date I believe, but it shows the process a request takes. > Yep. But i am still digesting it and i am pretty fresh in linux programming. I have looked at the part about process a request takes. It seems that a request is being answered by the storemanager. When i follow the code, there are just too many data structure and process to trace, i get lost by looking back and forth of the code in different file. It would be more helpful in the programming guide if it can tells what tools we can used to manupulate the structure used in each module and functions. > Second, unless you're using a 14.4K modem as your connection, squid can > probably have the content by the time the user sees any message telling them > that the content is being fetched. Also, unless your talking large files, I > don't think there would be much point to reporting the status of a fetch to > the client. Or are you talking about just using squid to cache large files > say from an ftp site? > Yep, close guess. As we want to put the squid server in china, the network in china is very slow when the requested content is in foreign country. The speed can be as slow as a 14.4K modem. (Downloading a red hat disc can take 1 day...) Of course we are not trying to cache large files like the image of red hat. But i want to tell the user to surf other page first while we are getting page from a foreign country and then we can share the cache to other user that may hit this foreign page later. > Perhaps a little more info would help. > > Cheers, > --Jason > > ----- Original Message ----- > From: "Louis Lau" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Wednesday, February 05, 2003 10:07 PM > Subject: [squid-users] Squid modification help > > > > Hi all, > > > > I wonder if anyone can give me hints in completing my assignment. > > > > In my assignment i need to accomplish the following: > > > > 1. I want to configure/modify squid such that when there is a cache = > > miss, the squid server should reply to the client that we are trying to = > > get the content for the user now. The squid server will then try to get = > > the content for the user in background so that next time (after some = > > minute i suppose), the content should be found in the cache. > > > > 2. If the user desperately need to get the cache-missed content = > > immediately, he/she can 'clt refresh' in the broswer and request the = > > most updated content from the cache. The squid will then check the http = > > request header 'if-modified-since' header and if it request the most = > > update, the cache will by-pass the scenario 1 mentioned above and = > > retrieve the content for client immediately. Even better is when i can = > > show the status of the retrieve if possible while the user is waiting = > > for the content.. (eg. such as telling the client that we have already = > > retrieved 50% of the content.) > > > > I have read some of the source code of squid now.. i am thinking of = > > modifiying the forward.c module in squid 2.3. As it seems that squid 2.3 = > > seems to be simplier to modify. And i have touched some code in = > > client_side.c too.. yet, it seems not working very well at the moment, = > > are there any module that i should have changed in order to finish my = > > assignment?=20 > > I want to know the internal architecture used in squid. It is quite = > > trouble to trace the function call in squid to see how squid works in = > > the code level. > > > > Thanks for any input that maybe suggested by you guys/gals.=20 > > > > Best regards, > > > > Louis Lau > > > > >
