Re: mod_proxy/2277: PATCH: cache completion of partially loaded documents

[Brian Behlendorf <br...@hyperreal.org>]

The following reply was made to PR mod_proxy/2277; it has been noted by GNATS.

From: Brian Behlendorf <br...@hyperreal.org>
To: glenebob@nwlink.com, apbugs@hyperreal.org
Cc:  Subject: Re: mod_proxy/2277: PATCH: cache completion of partially
  loaded documents
Date: Thu, 28 May 1998 18:01:14 -0700

 At 08:24 AM 5/25/98 -0000, Glen Parker wrote:
 >This patch will cause the proxy module to finish loading a file to the
 >cache in the event that the client cancels the transfer, provided that
 >the configured percentage of the file has already been transfered.
 >It works for http transfers only.
 
 Great, since this is a bug (the proxy module will cache incomplete
 transfers currently as if it were the whole file) I do want to see this
 resolved in the 1.3.0 timeframe.  This is the correct way to fix it too.
 
 >I have compiled this on WinNT 4.0 and Redhat Linux 5.0 (2.0.32 kernel).
 >I tested *briefly* on NT, and I'm using it on 3 different Linux servers
 >with good results.
 
 The more NT testing the better, we've found this to be our most troublesome
 platform.
 
 In looking at your patches I saw something which seemed incorrect:
 
 >=== source patch ========================================================
 >
 >diff -C3 apache_1.3b6/src/modules/proxy.orig/proxy_cache.c
 apache_1.3b6/src/modules/proxy/proxy_cache.c
 >*** apache_1.3b6/src/modules/proxy.orig/proxy_cache.c	Wed Apr 15 10:09:29
 1998
 >--- apache_1.3b6/src/modules/proxy/proxy_cache.c	Sat May 23 17:36:55 1998
 >***************
 >*** 551,557 ****
 >--- 551,560 ----
 >  	}
 >      }
 >      if (cachefp == NULL)
 >+ /* fixed?  in this case, we want to get the headers from the remote server
 >+    it will be handled later if we don't do this (I hope ;-)
 >  	c->hdrs = ap_make_array(r->pool, 2, sizeof(struct hdr_entry));
 >+ */
 >      /* FIXME: Shouldn't we check the URL somewhere? */
 >      now = time(NULL);
 >  /* Ok, have we got some un-expired data? */
 
 Shouldn't the "if(cachefp == NULL)" be commented out as well?  Otherwise
 "now" might not get initialized.
 
 Other than that I didn't see anything wrong, but this isn't a part of the
 code I understand fully.  I'm going to give it some testing, and if it
 looks good commit it.
 
 	Brian
 
 
 
 --=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
 pure chewing satisfaction                                  brian@apache.org
                                                         brian@hyperreal.org