48275 - Bypassing cache does not remove/update cache entry

Status: VERIFIED FIXED

(Core :: Networking: Cache, defect, P2)

Description

[Bill Law]

In the course of testing my fix for bug 21137 I observed the following, which I
believe to be a problem:

1. Load a page.
2. Reload the page, bypassing the cache (and proxy) by using Shift-Reload (note,
you can't really do this without applying my fix for bug 21137, plus, applying
the tweak to navigator.xul needed to work around bug 48243).  This will fetch a
new page from the server.
3. Then, reload again by pressing the Reload button *without* pressing the shift
key.

Result: The old (original) page is redisplayed, not the one fetched at step 2.

Expected result: The page displayed at step 2 should be redisplayed after a Reload.

The observed behavior might be partially due to the way I conducted my test (for
my fix to bug 21137).  I created a .html file on my NT box (using the simple
"Personal Web Server" that comes/came with NT).  Anyway, I then did "touch -t
08091234.56 foobar.html" to fix the timestamp on the file.  I then viewed the
file in Mozilla.  Next, I editted the source, and did the "touch" again.  Reload
didn't show the changes, but Shift-Reload did (that's the good news, since it
means my fix for bug 21137 seemed to be working).  I then pressed Reload and the
*original* version of the file reappeared.

Anyway, maybe it was some kind of fluke due to encountering files with the same
timestamp (i.e., the cache wasn't updated since the newly encountered page
wasn't newer).  That scenario isn't likely to be encountered in real life, I
shouldn't think.

Comment 1

[neeti]

Bill,

Any timeline of when the fix for 21137 will be checked in , so I can investigate 
this bug?

Thanks,
Neeti

Than

Comment 2

[Bill Law]

The fix for bug 21137 should be checked in later today.

Comment 3

[Suresh Duddi (gone)]

beta3 as this is cache correctness. 

Comment 4

[Suresh Duddi (gone)]

Should be fixe dnow.

Comment 5

[Tom Everingham]

verified:
WinNT 2000091108